HYDROGEN RECOVERY FROM FUEL GAS CAN HELP MEET REFINERY DEMAND

March 22, 1993
Energy conservation measures in refineries often result in an excess of fuel. Therefore, fuel gas is sometimes flared or more often used in burners. These practices waste not only valuable hydrogen, but also other light hydrocarbons. Recovering sufficient hydrogen from these fuel gas streams is the best way of debottlenecking the refinery and bringing it into hydrogen balance. A report by Harry Isalski, technology and licensing manager for Costain Oil, Gas & Process, Manchester, England,

Energy conservation measures in refineries often result in an excess of fuel. Therefore, fuel gas is sometimes flared or more often used in burners. These practices waste not only valuable hydrogen, but also other light hydrocarbons.

Recovering sufficient hydrogen from these fuel gas streams is the best way of debottlenecking the refinery and bringing it into hydrogen balance. A report by Harry Isalski, technology and licensing manager for Costain Oil, Gas & Process, Manchester, England, reviews the technologies available for hydrogen recovery.

HYDROGEN BALANCE

The modern, integrated refinery produces a wide range of products and intermediates from a variety of crude oil feedstocks. Each refinery has a unique design basis and slate of products, and therefore a unique arrangement of process units.

One common aspect of all refineries that is receiving increased attention is hydrogen balance (see article p. 45). The heavier the processed crude and the higher its sulfur content, the more likely hydrogen consumption will exceed internal hydrogen production.

Changes in crude and product slates from the design basis can make it difficult for even a new refinery to operate efficiently. And for older refineries it is particularly important to optimize hydrogen usage.

The hydrogen shortage will become even more important in the future. Demand for gasoline and middle distillates means crude oil will require greater processing and thus more hydrogen.

Product quality requirements are continually becoming more stringent. These requirements have implications on hydrogen demand and production. For instance, reduction in the maximum gasoline aromatics levels, as required by the U.S. 1990 Clean Air Act Amendments, may force a reduction in the output of aromatics from the catalytic reformer, which also reduces hydrogen production from this unit.

Environmental pressures also mean increased hydrodesulfurization of refinery streams and therefore increased hydrogen demand. And processing heavier crudes means increased use of hydrogenation downstream through the use of such units as Flexicokers and residue catalytic crackers.

As refineries become more complex, hydrogen balance becomes more complex and it is more difficult to match supply with demand. The typical refinery is either now bottlenecked because of lack of hydrogen, or will be in the near future as more severe product specifications come into effect.

HYDROGEN RECOVERY

Fuel gas streams must be treated in a gas processing unit in order to recover hydrogen. A complex refinery can have as many as 50 sources of fuel gas, all supplied into the refinery fuel header. These sources contain varying concentrations of hydrogen, methane, ethane, propane, and higher hydrocarbons, including small concentrations of aromatics and C8+.

It may also be worthwhile to recover other components. The valuable components of fuel gas streams are usually:

  • Hydrogen, for use in hydrotreating, hydrocracking, and other hydrogen-consuming operations

  • Ethane, for use in steam crackers

  • Propane, for sale or for use in crackers

  • Butanes and heavier components, for fractionation and blending into automotive fuels, or for use as feedstock to etherification, alkylation, isomerization, and other units.

The choice of technology for hydrogen recovery depends mainly on feed pressure, product pressure, and purity requirements. A qualitative discussion of these and other aspects of each technology will help refiners choose one technology over another.

A number of alternative technologies can be used for the recovery of hydrogen from refinery and petrochemical plant offgases:

  • Pressure swing adsorption

  • Semipermeable membranes

  • Cryogenics.

Table 1 compares the advantages and disadvantages of these hydrogen-recovery technologies.

PSA

Pressure swing adsorption (PSA) is a means of removing compounds from a light product without applying any heat. The impurities are adsorbed as the gas flows through an adsorbent bed and are then desorbed by reducing the pressure. Hydrogen is adsorbed in only small amounts, thus enabling high-purity hydrogen to be produced.

Simpler PSA units have typical recoveries of only about 75-85%. This recovery can be increased to about 90% by using more beds of molecular sieve, but with greater cost and complexity. Commercial-scale units can have between 4 and 10 adsorbers.

PSA tends to become expensive at higher feed gas rates, but is very competitive at lower rates of about 5,000 normal cu m/hr (about 4.2 MMscfd).

Where very high-purity hydrogen is required, PSA is very competitive. High-purity hydrogen, however, is not normally a refinery requirement; the quantity of hydrogen recovered is more important.

MEMBRANES

Membrane technology relies on high-pressure feed gas to the membrane unit. This high pressure is necessary because separation is achieved by virtue of the different relative permeabilities that different components have through gas-permeable membranes, and the differences in the partial pressures of each component across the membrane.

Membrane technology is suitable for refinery hydrogen recovery if the feed gas is available at high pressure and the desired hydrogen product (the permeate) is required at low pressure. This is not a common situation.

It is difficult to recover substantial amounts of hydrogen at reasonable purity without high compression costs. As a result, membrane technology is generally used for bulk separation to give relatively low to medium-purity hydrogen. Both purity of the hydrogen product and hydrogen-recovery levels tend to be about 90%.

For low gas feed rates where high purity and recovery are not important and low pressure product is acceptable, membrane units are an attractive option. At increased feed gas rates, the high power penalty they incur for compression makes them less economical than cryogenic processing units. Membranes are not practical for high-purity applications, which remain in the domain of PSA units.

CRYOGENICS

The separation of a gas mixture using low temperatures relies on the difference in relative volatilities of the components to achieve separation. As hydrogen has a high relative volatility compared to hydrocarbons, the recovery of medium to high-purity hydrogen can be high. And in contrast to membranes, product is also available at feed gas pressure.

Cryogenics falls between PSA and membranes, offering moderately high purities and high recoveries (Table 1).

SUITABLE FEED GASES

Typical refinery offgases with relatively high hydrogen content include:

  • Hydrodealkylation unit offgas

  • Catalytic reformer offgas

  • Hydrodesulfurization unit offgas

  • Catalytic cracker offgas

  • Hydrotreater offgas.

The hydrogen content and operating pressure of the offgas have a large influence on both the choice of process route and the capital investment for the recovery unit. For both membrane and cryogenic processes, the higher the feed gas pressure, the better the performance for a given capital cost.

Table 2 shows three groups of gases that are the sources of the net refinery fuel header gas. Group 1 gases usually include hydrodealkylation, hydrotreater, and perhaps hydrodesulfurization offgases. Group 2 usually includes reformer offgas. And Group 3 gases usually include fluid catalytic cracker offgas and some of the gas plant offgases.

Hydrogen recovery is appropriate for Groups 1 and 2, and all three technologies can give product purities of more than 90 vol %. Cryogenics can recover more than 90% of the hydrogen in the feed, but for Group 2 gases, this may require external refrigeration and increased cost.

Group 1 gases would normally be expected to be processable via cryogenics without the need for external refrigeration.

Membrane processes are appropriate for gases in Group 1, and especially Group 2, as the high hydrogen concentration in Group 2 feeds enhances membrane performance. The viability of the membrane route, however, is highly dependent on a large pressure difference between feed and hydrogen product.

PSA is appropriate for Group 2 gases; Group 1 gases often contain too many impurities for PSA to be viable.

Cryogenics is viable for Group 3 gases, but generally, recovery of C2 components and LPG-not hydrogen alone-is necessary to justify a project. Membranes and PSA are not suitable for Group 3 gases.

Steam methane reforming, incidentally, is often appropriate for Group 3 gases because it converts hydrocarbons to hydrogen and carbon oxides and recovers all of the hydrogen in the stream. A good example of this is Texaco Inc.'s HYTEX process, which uses oxygen as the oxidant.

Sometimes a combination of recovery technologies is the most appropriate choice. Cryogenics in conjunction with PSA is probably the best combination to get high recovery and high purity of hydrogen at high pressure.

Because of the wide variety of refinery fuel gases available for processing, it is important that each option be studied carefully and that the full economic and operating consequences be understood. The potential to recover other valuable components should also be considered.

When grouping refinery streams, care should be taken to ensure that the best advantages of pressure and composition are taken when providing an integrated solution.

Even in refineries with a hydrogen-generation unit, recovery of hydrogen from fuel gases can be justified to minimize capital and operating costs for the steam methane reformer. Cryogenics is often a good economic choice for hydrogen recovery and is a well-proven and flexible solution to the problem of hydrogen deficiency.

Copyright 1993 Oil & Gas Journal. All Rights Reserved.