Apr-2013

Simulation of the Benfield HiPure process of natural gas sweetening for LNG production

Achieving specifications given by customers, including pipeline-operating companies, LNG storage facilities, and gas-processing plants requires the removal of CO2 and H2S from natural gas. It is also becoming more important to meet environmental regulations set by national and local governments.

R Ochieng, A S Berrouk and C J Peters, Department of Chemical Engineering, Petroleum Institute
Justin Slagle, Lili Lyddon and Peter Krouskop
Bryan Research & Engineering, Inc

Article Summary

This paper summarises a study which compares the Benfield HiPure–LNG Train of Abu Dhabi Gas Liquefaction Company Limited (ADGAS) sweetening plant to other sweetening processes using the modelling software ProMax. Natural gas from the gas reservoirs, containing about 6-7 mole% acid gas, first comes into contact with hot potassium carbonate (30 wt% K2CO3) promoted with diethanolamine solution (3 wt% DEA), and finally with 20 wt% DEA solution. The simulation proved to predict the plant operating data accurately.

Subsequently, additional alternatives to the Benfield HiPure process were investigated as potential options for replacement, including MDEA, MDEA/piperazine, and MDEA/DEA mixtures. The activated MDEA (50 wt% MDEA + 3 wt% PZ) with a two-stage flash is the best alternative, with a 36% decrease in the reboiler duty.

This paper shows the possibility of shutting down the potassium carbonate section of the sweetening train and swapping the DEA solution in the immediate downstream unit for a mixed amine in order to reduce operating costs while continuing to meet the treated gas specifications.

Preliminary results are presented here.
Introduction
There are many processes available for removal of acid gases from natural gas; the selection of these processes is based on economic feasibility and cleanup ability. These processes include chemical solvents, physical solvents, hybrid solvents, adsorption processes, and physical separation (e.g. membrane systems)1. Chemical and physical solvents, or a combination of the two, have been used extensively in many existing LNG facilities. The removal of both H2S and CO2 from natural gas before liquefaction is done primarily to meet the LNG product specifications, prevent corrosion of process equipment, and meet environmental performance standards. The recommended specifications for LNG are typically less than 1 ppmv H2S and 50 ppmv CO2 in the sweet gas^1-4.

Solvent cost, equipment costs, and energy requirements for regeneration are the most important factors to be considered in selecting an appropriate process^{2, 17}.

Depending on the process requirements, several options for alkanolamine based chemical solvents may be proposed^{1, 5}. Apart from alkanolamine based processes, other methods for removal of H2S and CO2 include alkaline salts such as sodium or potassium carbonate (with or without an amine activator) and physical solvents such as DEPG or methanol.

The HiPure process used currently in the ADGAS plant and described by Benson and Parrish⁶ uses two independent but compatible circulating solutions in series, specifically, hot potassium carbonate (K2CO3) promoted with DEA followed by DEA to achieve high purity at high efficiency.

The hot potassium carbonate process introduces major process concerns of corrosion, erosion, and column instability which affect the capital and maintenance costs in the form of design and operation⁵. If CO2 is not present, it becomes difficult to regenerate potassium bisulfide; therefore, potassium carbonate is not a suitable option for H2S only cases⁵. A drawback of diethanolamine is the possible need for vacuum distillation while reclaiming contaminated solutions. DEA also undergoes numerous irreversible reactions with CO2 forming corrosive degradation products. For that reason, DEA may not be the optimum choice for treating high CO2 content gases¹.

The use of blended amines in gas treating can bring about a significant improvement in the absorption capacity, absorption rate, and also savings in solvent regeneration energy requirements^7-10. This approach could dramatically reduce capital and operating costs while providing more flexibility in achieving specific purity requirements. For a given economic analysis, choosing a process with low initial installation cost might not be the best option since the operating cost may be high, making the breakeven point unattainable.

In most cases, amine mixtures contain MDEA as the base amine with the addition of one or two more reactive amines such as MEA, Piperazine, or DEA. These amine mixtures are also known as formulated amines, activated MDEA, promoted MDEA, and MDEA based amines.

Piperazine activated MDEA has a higher energy requirement than the physical solvent processes, but has lower hydrocarbon solubility.

Compared to MDEA with other activators, the Piperazine activated MDEA has a low energy requirement due to its ability to liberate the bulk acid gases in a simple flash. For the two-stage absorption process, only a portion of the semi-lean solution is regenerated. This reduces capital cost and energy requirements.

MDEA based processes have commercial advantages over the current Benfield HiPure Process in that MDEA is less corrosive to carbon steel, the solution is stable, and it is not as susceptible to degradation. Since MDEA is not very corrosive, higher concentrations of up to 50% can be used without any significant effects on the process equipment².

Process industries widely recognise process simulators as an essential predictive tool. Providing predictive models provide many benefits: studying process alternatives, assessing feasibility, performing preliminary economics, interpreting pilot-plant data, optimising process design hardware, estimating equipment, calculating operating costs, investigating feedstock flexibility, and optimising plant operations to reduce energy use, increase yield and improve pollution control².