Carbon Dioxide Capture and Acid Gas Injection

Preface

The Sixth International Acid Gas Injection Symposium (AGIS VI) was held in Houston, Texas, in September 2016. As with previous Symposia, the focus of AGIS VI was the injection of acid gas (CO2, H2S, and mixtures of these components) for the purposes of disposal or for enhanced oil and/or gas recovery. This book contains select papers from the Symposium in Houston.

The capture of carbon dioxide from flue gas and its disposal into a subsurface geological formation remains a viable option for the clean use of hydrocarbon fuels. The related technology is acid gas injection. Here the H2S and CO2 are removed from raw natural gas. This volume contains papers directly related to these two topics ranging from the physical properties of the gas mixtures, evaluation of new and existing solvents, and subsurface engineering aspects of the process. Furthermore, contributors came from Canada, Europe, and China, as well as from the host country, the United States. And this is reflected in the papers in this volume.

On a very sad note, Marco Satyro passed away on September 8, 2016, just prior to the Symposium. Marco was a good friend of AGIS being an active member of the Technical Committee for many years. He contributed many papers and encouraged many others to participate. At the first AGIS he presented the paper The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water and this paper appeared in the first volume of the Advances in Natural Gas Engineering. He also was the coauthor of several other contributions to the Series and they are listed below. This volume is dedicated to the memory of Dr. Satyro.

References – papers of M.A. Satyro from the Advances in Natural Gas Engineering series.

M.A. Satyro, and J. van der Lee, The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water, pp. 21–34, Acid Gas Injection and Related Technologies, Y. Wu and J.J. Carroll (eds.), Scrivener Publishing (2011).

H. Motahhari, M.A. Satyro, and H.W. Yarranton, Acid Gas Viscosity Modeling with the Expanded Fluid Viscosity Correlation, pp. 41–52, Carbon Dioxide Sequestration and Related Technologies, (2011), Y. Wu, J.J. Carroll, and Z. Du (eds.), Scrivener Publishing (2011).

J. van der Lee, J.J. Carroll, and M.A. Satyro, A Look at Solid CO2 Formation in Several High CO2 Concentration Depressuring Scenarios, pp. 117–128, Sour Gas and Related Technologies, Y. Wu, J.J. Carroll, and W. Zhu (eds), Scrivener Publishing (2012).

M.A. Satyro, and J.J. Carroll, Phase Equilibrium in the Systems Hydrogen Sulfide + Methanol and Carbon Dioxide + Methanol, pp. 99–109, Gas Injection for Disposal and Enhanced Recovery, Y. Wu, J.J. Carroll, and Q. Li (eds.), Scrivener Publishing (2014).

A.R.J. Arendsen, G.F. Versteeg, J. van der Lee, R. Cota, and M.A. Satyro, Comparison of the Design of CO2-capture Processes using Equilibrium and Rate Based Models, pp. 155–174, Gas Injection for Disposal and Enhanced Recovery, Y. Wu, J.J. Carroll, and Q. Li (eds.), Scrivener Publishing (2014).

M.A. Satyro and H.W. Yarranton, A Simple Model for the Calculation of Electrolyte Mixture Viscosities, pp. 95–104, Acid Gas Extraction for Disposal and Related Topics, Y. Wu, J.J. Carroll, and W. Zhu (eds.), Scrivener Publishing (2016).

Chapter 1

Enthalpies of Carbon Dioxide-Methane and Carbon Dioxide-Nitrogen Mixtures: Comparison with Thermodynamic Models

Erin L. Roberts and John J. Carroll

Gas Liquids Engineering, Calgary, Alberta, Canada

Abstract

The physical properties of acid-gas injection streams are important for use in design considerations of the acid-gas scheme. One such property is the enthalpy of the stream. As carbon dioxide is rarely pure, with methane and nitrogen being common impurities in the stream, the effect of these impurities on the enthalpy is also important to consider.

This study compares experimentally determined excess enthalpies and enthalpy departures from literature to the enthalpy predictions of five different models, Benedict-Webb-Rubin, Lee-Kesler, Soave-Redlich-Kwong, and Peng-Robinson from VMGSim, as well as AQUAlibrium software. The mixtures studied are carbon dioxide-methane, as well as carbon dioxide- nitrogen mixtures at a wide range of compositions.

The Soave-Redlich-Kwong model gave the most accurate predictions for both the excess enthalpies and enthalpy departures, with Lee-Kesler frequently giving the least accurate predictions for the mixtures.

1.1 Introduction

An increase in demand of natural gas has led producers to pursue poorer quality reservoirs. These contain higher levels of carbon dioxide that then must be responsibly disposed. Regulations prevent the flaring of the acid-gas mixtures, therefore requiring an alternate means of disposal. One such method is the injection of acid gas into subsurface reservoirs.

An understanding of the physical properties of the stream is essential in the design of the acid-gas injection scheme. The enthalpy of the stream is required in the design of the compressor for injection. Common impurities in the carbon dioxide include methane and nitrogen; therefore the effect of these impurities on the enthalpy of carbon dioxide is required for design.

This paper investigates the accuracy of five different thermodynamic models for predicting such mixtures. Four different equations of state, Benedict-Webb-Rubin (BWR), Lee-Kesler (LK), Soave-Redlich-Kwong (SRK), Peng-Robinson (1978) were used with VMGSim software, as well as the AQUAlibrium model. BWR and LK are multi-constant equations, and SRK and PR78 are cubic equations of state. The AQUAlibrium model uses a variation of Peng-Robinson.

1.2 Enthalpy

The enthalpy of mixtures can be determined in a number of ways. One method is to use excess enthalpy (enthalpy of mixing). Excess enthalpy is defined as

(1.1)

where: HE – Excess enthalpy

Hm – Enthalpy of mixture

Hi – Enthalpy of component i

xi – mol fraction of component i

Alternatively, the enthalpy of the mixture can be represented as an enthalpy departure, a difference between the enthalpy at a given pressure, and the enthalpy at a reference pressure while keeping the temperature constant.

Enthalpies can be expressed in J/mol, or for greater relevance to acid-gas injection design, can be expressed in HP/MMSCFD. The conversion between units is 1 HP/MMSCFD to 53.86 J/mol.

1.3 Literature Review

A review of literature was performed to compile experimental data for the enthalpy of carbon dioxide-methane mixtures as well as carbon dioxide-nitrogen mixtures. Table 1.1 summarizes the relevant data used in this study.

Table 1.1 Summary of experimental data of enthalpy of carbon dioxide mixtures.

1. Lee & Mather (1972)

2. Barry et al. (1982)

3. Ng & Mather (1976)

4. Peterson & Wilson (1974)

5. Lee & Mather (1970)

6. Hejmadi et al. (1971)

1.3.1 Carbon Dioxide-Methane

The most extensive study performed for enthalpies of carbon dioxide-methane mixtures was performed by Lee & Mather (1972). Their study consisted of mol fractions of 0.1–0.9, taken at intervals of 0.1, for a total of 9 different mol fractions. Measurements of excess enthalpy were reported at 8 different temperatures from 10–80 °C, with ranges of pressure of 1.0–4.4 MPa for 10 °C, 1.0- 5.07 for 20 °C, 1.0–11.1 for 40 °C, and 1.0–10.1 for 32 °C, 50 °C, 60 °C, 70 °C, and 80 °C. In total, 648 data points were reported. Two typographical errors were found in the data set; they are not included in the numerical error analysis but are represented in the figures.

Another smaller study was performed by Barry et al. (1982), for excess enthalpies of carbon dioxide-methane mixtures. Data was taken at three different temperatures, 20 °C, 32 °C, and 40 °C. Seven different pressures were used, ranging from 0.51 MPa to 4.6 MPa, with pressure of over 2 MPa only being measured for 40 °C. The mol fractions measured were not taken in increments, instead were taken at a wide variety of fractions ranging from 0.1 to 0.9.

Two other studies were done using enthalpy departures by Ng & Mather (1976) and Peterson & Wilson (1974). Ng & Mather (1976) used pressures of 3–13.7 MPa, and temperatures of 0–90 °C for mol fractions of 0.145 and 0.423. They used the ideal gas enthalpy as a reference point to measure the enthalpy departure. Peterson & Wilson (1974) only measured equimolar mixtures of carbon dioxide and methane with pressures from 0.7–13.8 MPa and temperatures of 255.4 K–422 K. The reference enthalpy used was measured at a pressure of 0.138 MPa. These two studies were the only ones that measured both liquid and vapor enthalpies, instead of just vapor.

1.3.2 Carbon Dioxide-Nitrogen

Lee & Mather (1970) and Hejmadi et al. (1971) studied the excess enthalpies of carbon dioxide-nitrogen mixtures. Lee & Mather (1970) looked at mole fractions from 0.1–0.9 at intervals of 0.1. Pressures from 1.01 MPa to 12.16 MPa were used, at only a single temperature of 40 °C.

Hejmadi et al. (1971) used only two different temperatures of 31 °C and 40 °C, and two different pressures of 3.5 MPa and 6.5 MPa. They used mole fraction of nitrogen from 0.2–0.7.

1.4 Calculations

The experimental enthalpies were compared to calculated enthalpies using BWR, LK, SRK, and PR78 thermodynamic models from VMGSim software, as well as using AQUAlibrium software.

The six different mixtures (four with methane, two with nitrogen) as summarized in Table 1.1 were evaluated. Four error functions for both the excess enthalpies and the enthalpy departures were used to analyze the accuracy of the prediction of each method.

For the excess enthalpies, the absolute average difference (AAD) was defined as;

(1.2)

where: NP – number of points

HEexp – experimental excess enthalpy

HEcalc – calculated excess enthalpy

and the average difference (AD) was defined as:

(1.3)

The absolute average error (AAE) in excess enthalpies was defined as:

(1.4)

and the average error (AE) was defined as:

(1.5)

For enthalpy departures, the absolute average difference

(1.6)

where H° – enthalpy of mixture at reference pressure

H – enthalpy of mixture at measured pressure

and the average difference was defined as:

(1.7)

The absolute average error for enthalpy departure was defined as:

(1.8)

and the average error was defined as:

(1.9)

1.4.1 Benedict-Webb-Rubin

For the Lee & Mather (1972) methane data of excess enthalpies, the AAD was 78.1 J/mol and the AD was 2.6 J/mol. The AAE was 19.0% and the AE was –14.6%. The maximum difference was 2113.2 J/mol occurring at 8.11 MPa and a mole fraction of 0.2. The maximum error was 131.7% at the same conditions as the maximum difference. At lower pressures, the enthalpies were overestimated, and at the higher pressures they were underestimated. The greatest deviations occurred when there was a rapid change in enthalpy with pressure. This occurred at around 7–10 MPa for the 32 °C and 40 °C temperatures. There was also a very large difference between the calculated and experimental enthalpy for the 10.13 MPa isobar at 50 °C. Figures 1.1 through 1.8 show the experimental and calculated enthalpies for the different temperatures.

Figure 1.1 Experimental and calculated enthalpies at 10 °C using BWR (Lee & Mather, 1972).

Figure 1.2 Experimental and calculated enthalpies at 20 °C using BWR (Lee & Mather, 1972).

Figure 1.3 Experimental and calculated enthalpies at 32 °C using BWR (Lee & Mather, 1972).

Figure 1.4 Experimental and calculated enthalpies at 40 °C using BWR (Lee & Mather, 1972).

Figure 1.5 Experimental and calculated enthalpies at 50 °C using BWR (Lee & Mather, 1972).

Figure 1.6 Experimental and calculated enthalpies at 60 °C using BWR (Lee & Mather, 1972).

Figure 1.7 Experimental and calculated enthalpies at 70 °C using BWR (Lee & Mather, 1972).

Figure 1.8 Experimental and calculated enthalpies at 80 °C using BWR (Lee & Mather, 1972).

The Barry et al. (1982) methane data of excess enthalpies had an AAD of 9.1 J/mol, an AD of –8.3 J/mol, an AAE of 14.2% and an AE of –11.0%. The maximum difference was 46.5 J/mol at 4.6 MPa, 40 °C and 0.351 mole fraction methane. The maximum error was 42.5% at 0.53 MPa, 32 °C and 0.63 mole fraction methane The deviations are smaller due to the lower pressure range of the data.

The Lee & Mather (1970) nitrogen data of excess enthalpies taken at 40 °C had similar results as the Lee & Mather (1972) methane data for the 40 °C data, with the greatest difference occurring at 9.12 MPa. The AAD was 151.1 J/mol, the AD was 58.7 J/mol, the AAE was 15.0% and the AE was –0.7%. The maximum difference was 969.8 J/mol at 9.1 MPa, and 0.1 mole fraction nitrogen. The maximum error was 70% at the same conditions as the maximum difference. Figure 1.9 shows the calculated and experimental enthalpies for the BWR model at 40 °C.

Figure 1.9 Experimental and calculated enthalpies at 40 °C using BWR (Lee & Mather, 1970).

The Hejmadi et al. (1971) nitrogen of excess enthalpies data had an AAD of 26.1 J/mol, and AD of –11.0 J/mol, an AAE of 9.5% and an AE of –7.9%. The maximum difference was 90.8 J/mol at 6.5 MPa, 31 °C, and 0.239 mole fraction nitrogen. The maximum error was 14.1% at 3.4 MPa, 40 °C and 0.67 mole fraction nitrogen. As with the Barry et al. (1982) methane data, the lower deviations are likely due to the lower pressure range used in the measurements, as the highest pressure used was 6.5 MPa and the greatest deviations typically occurred around 7–10 MPa for temperatures in the 30–40 °C range.

For the Peterson & Wilson (1974) methane data for enthalpy departures, the AAD was 56.4 J/mol, the AD was 26.2 J/mol, the AAE was 3.7% and the AE was 1.4%. Two points were omitted from the error calculations due to BWR predicting a vapor/liquid mix. The Ng & Mather (1976) methane data for enthalpy departures had an AAD of 192.3 J/mol, an AD of 182.2 J/mol, an AAE of 3.8% and an AE of 3.0%

1.4.2 Lee-Kesler

The Lee & Mather (1972) methane data for excess enthalpies had an AAD of 46.7 J/mol, an AD of –43.2 J/mol, an AAE of 20.1%, and an AE of –19.7%. Figures 1.10 through 1.17 show the experimental and calculated enthalpies for the 8 different temperatures. The greatest differences typically occurred at the highest pressure and at low methane mole fractions for all temperatures. The maximum difference was 505.5 J/mol occurring at 50 °C, 10.1 MPa and 0.1 mol fraction methane. The greatest errors always occurred at a mole fraction of 0.1 and a pressure of 1.01 MPa. The maximum error was 98.0% occurring at 80 °C. For almost all data points, LK overestimated the enthalpies. The only conditions where they were underestimated was at high methane mole fraction and high pressures.

Figure 1.10 Experimental and calculated enthalpies at 10 °C using LK (Lee & Mather, 1972).

Figure 1.11 Experimental and calculated enthalpies at 20 °C using LK (Lee & Mather, 1972).