METHANOL PLANT DESIGN CHOICES AFFECT OPERATIONS, COSTS, OTHER EQUIPMENT

March 29, 1993
With the expected increase in worldwide methanol demand, particularly for methyl tertiary butyl ether (MTBE) feedstock, there will be an increase in the number of new plants and retrofits to existing plants. Several options are available for overall plant design and synthesis-converter design, each having its advantages and disadvantages. And for retrofits, there are a number of ways to increase capacity.

With the expected increase in worldwide methanol demand, particularly for methyl tertiary butyl ether (MTBE) feedstock, there will be an increase in the number of new plants and retrofits to existing plants.

Several options are available for overall plant design and synthesis-converter design, each having its advantages and disadvantages. And for retrofits, there are a number of ways to increase capacity.

PLANT DESIGN

Methanol production is the most profitable means of adding value to natural gas. Potential methanol projects exist wherever there are large reserves of competitively priced natural gas. Consequently, future methanol plants will be located in increasingly varied and remote locations.

These views, expressed in a paper by Mark Linthwaite and Mark Sutton of John Brown, Davy Process Technology, London, were presented at Crocco & Associates Inc.'s 1992 World Methanol Conference, Dec. 8-10, 1992, in Monte-Carlo. Their paper, "Methanol Plant Designs for the Demands of the Nineties," examined the factors influencing the development of worldscale chemical-grade methanol projects.

Linthwaite and Sutton say the most important factor in ensuring profitability of a methanol plant is the manufacturing costs it will be able to achieve. The factors influencing the unit cost are energy expenditure, reliability, environmental impact, location, and capital cost.

But all of these demands cannot be met with a single flowsheet. For example, minimizing environmental impact and energy expenditures increase capital costs. The appropriate compromise must therefore be made on a case-by-case basis to account for location-specific factors and operator preferences.

ENERGY EFFICIENCY

The synthesis gas generation area of the plant requires the largest proportion of the plant energy input, thus the choice of syngas generation route will likely have the greatest influence on plant energy efficiency.

Available generation schemes include:

  • Conventional steam reforming

  • Partial oxidation

  • Combined reforming

  • Parallel reforming

  • Gas-heated reforming.

Of these, the three current contenders for world-scale methanol plants are conventional, combined, and parallel reforming (Figs. 1, 2, and 3). Table 1 shows the energy requirements for a 2,500 metric ton/day (mtd) plant using these three routes. In each case, Linthwaite and Sutton designed the plants on the same basis to give a reasonable balance between capital and energy expenditure.

The two oxygen routes (parallel and combined) at first show significant energy savings. But oxygen production is very energy intensive, which must be accounted for when comparing flowschemes. Table 1 shows that, when comparing like with like, the oxygen-blown routes may offer no energy savings at all or, at best, a savings of about 1.4%.

For the two major schemes for cleaning up crude methanol-two and three-column distillation-the additional capital spending required for a three-column system is unjustified except when low-grade heat is expensive.

ENVIRONMENTAL IMPACT

Environmental legislation worldwide is aiming at reducing combustion emissions-specifically CO2 and NOx. And with the likelihood of these restrictions tightening in the future, environmental impact will clearly influence the selection of process route and equipment.

The major source of carbon dioxide from a methanol plant is the syngas generation unit. Table 2 compares CO2 emissions from the three reforming routes. Again, on first examination, the oxygen-blown routes result in lower CO2 emissions. But, say Linthwaite and Sutton, the environmental impact relating to the extraction and generation of feeds and utilities to the plant must be considered.

The only major difference between the flowsheets in Table 2 is the electrical power requirement for oxygen generation.

Adding the emissions from the electrical generating stations shows that the oxygen routes produce more CO2 than does conventional reforming.

Table 3 compares NOx emissions for the three routes.

The base-case emissions from the conventional reformer are for a unit with high air preheat and no NOx-reduction measures in use. If catalytic reduction technology is applied to all three steam reformers, conventional reforming again produces the lowest total NOx emissions.

Varying degrees of NOx-reduction can be achieved by:

  • Reducing combustion air preheat

  • Burning/purge gas

  • Injecting steam

  • Using low-NOx/staged combustion burners

  • Catalytically reducing NOx to nitrogen and water.

Davy Process Technology is working with Balke-Durr of the Deutsche-Babcock Group to adapt its catalytic reduction technology for use on large-scale reformers.

PLANT RELIABILITY

As major methanol projects rely more heavily on commercial lenders for funding, reliability is of prime importance in ensuring plant viability. Two additional days of downtime can result in a loss of $440,000/year for a 2,500 mtd plant.

This is equivalent to a 1.5% drop in plant efficiency, according to the authors. The savings from preventing this loss is greater than any energy savings that the combined or parallel flowsheets can potentially achieve over the more-reliable steam reformer.

Conventional steam reforming plants are now designed for 355 days/year operation with 2-4 years between shutdowns, as dictated by catalyst changeout requirements. This high availability is achievable because of the relative simplicity of the plant, its infrequent maintenance requirements, and considerable operating experience.

By contrast, combined and parallel plants contain additional equipment and are thus more complex, operate at more-severe conditions, handle and burn pure oxygen, and require regular oxygen burner replacement. These factors, plus the resulting more-complex control and protection systems, increase the likelihood of plant outages, and hence serious losses of revenue.

CAPITAL COST

Major factors affecting capital costs include:

  • Location-related factor such as construction, site access, and labor costs

  • Site-preparation costs

  • Project-procurement policy

  • Local legislation on the environment, health, and safety

  • Imposed design codes and practices

  • Site infrastructure requirements and utility-accessing costs

  • Land costs.

To achieve profitable production, plant design must balance capital cost against energy expenditure. Balancing synthesis-loop catalyst and compressor costs against loop pressure and circulation is one such example. And unnecessary expenditures on oversized equipment can be reduced by limiting built-in design margins.

It is also important to prevent costs from rising during project execution. Owner/contractor partnering arrangements help minimize project growth during the implementation stage.

REACTOR DESIGN

Another presentation at the Crocco & Associates conference-"Methanol Reactor Design Choices" by P. E. J. Abbott of ICI Katalco, Billingham, U.K.-compared the four major design options for methanol synthesis conversion:

  • Quench-cooled converters

  • Adiabatic-bed converters

  • Tube-cooled converters

  • Steam-raising converters.

QUENCH-COOLED

The ICI quench-cooled converter is the most common methanol-converter design in the world (Fig. 4). The operational history of these plants, some of which have been on line for more than 20 years, has proven their robustness and reliability. Quench converters can be designed and successfully operated for capacities ranging from 50 to 3,000 mtd.

In the warm quench system, the quench is fed to the converter at 100-150 C. An advantage of this system, according to Abbott, is that 75% of the converter effluent is available for heat recovery at high temperatures (150-270 C.).

In comparison with other converter types, the quench converter requires a relatively large catalyst volume because the temperature profile does not follow the maximum rate path. In comparison to a tube-cooled converter, the quench system has more variables to control and is therefore somewhat more complex to optimize.

ADIABATIC BED

Adiabatic-bed converters have a smaller total catalyst volume than do quench systems because all of the gas passes through all of the catalyst (Fig. 5). The maximum rate trajectory path, however, also is not followed closely.

Steam at 30 bara is raised within the converter system, but there can be no heat recovery downstream of the converter because all of the converter effluent is required to heat the feed gas.

A minor advantage of the adiabatic-bed system is that the reactors and steam-raising heat exchangers can all be the same size, which reduces capital costs.

The major drawback of this system is the much larger number of vessels and interconnecting pipework and the much greater loop-interchanger area. Another disadvantage is that the 30-bara steam cannot be introduced directly into the reforming section in a combined-reforming-type plant.

TUBE-COOLED

The tube-cooled converter is a very simple, relatively new design (Fig. 6). There are five such units on-line; two are in ammonia service, and another is under construction. The first methanol plant to employ this design has been operating successfully for more than 3 years.

The main advantage of a tube-cooled converter is a relatively low catalyst volume. This leads to a small converter volume because, in this system, the reaction path closely follows the maximum rate path. Furthermore, because catalyst volume and heat-exchange area are combined, there are fewer equipment items needed than for a quench-converter system. This, in turn, reduces capital costs.

Another benefit of the tube-cooled design is that increased heat recovery, as compared to a quench converter, is possible. The mechanical design of the converter is simple and the converter can be fabricated without a tube sheet. And finally, a tube-cooled converter is very easy to control because all temperatures within the converter are set by the inlet temperature.

The only drawback, according to Abbott, is that the converter is relatively heavy, as are all tubular converters. For a world-scale plant, this may produce transportation problems.

STEAM RAISING

A variety of tubular steam-raising converters are available, each of which has unique features. For example, the converters can be either axial flow or radial flow, which will have a much lower pressure drop. Designs can feature catalyst on either the shell side or the tube side.

Fundamentally, the choice of steam-raising converter design does not alter the flowsheet (Fig. 7). The difference between steam-raising and tube-cooled converters is that the loop interchanger will be larger for the steam-raising design because all the gas must be heated to 225-240 C., as compared with 150 C. The steam-raising converter may also require more than one shell.

The main advantage of a steam-raising converter is the small catalyst volume required as a result of good catalyst stability. This is true because the converter operates almost isothermally, says Abbott, and at low catalyst peak temperatures (250-260 C.).

A further advantage of these low temperatures is that by-product make is low. Because steam can be produced at 40 bara-a much higher pressure than for an adiabatic system-it can be injected directly into the reforming section for use as process steam.

A drawback to having tube-side catalyst is the large number of tubes required to accommodate the catalyst. This is costly. On the other hand, similar near-isothermal performance can be achieved with a lower-heat-transfer area using catalyst on the shell side of the converter.

GRASSROOTS COMPARISON

ICI compared theoretical 2,000 mtd, 80-bara plants employing the four converter types discussed (Table 4). The plants had similar heat recoveries and circulation rates.

Table 4 shows that the steam-raising converter has the smallest relative catalyst volume of the four types, and overall, the quench converter has the largest vessel.

The quench and tube-cooled designs require much smaller loop-interchanger size. This illustrates the distinct advantage in not having to heat all the process gas up to the converter inlet temperature. In terms of total heat-exchanger area, the adiabatic-bed converter requires far more than the other three types.

A simplified estimation of the equipment procurement and installation costs includes only the converter configurations, loop, and associated heat-recovery equipment (steam-raising or saturator water systems).

The large number of tubes in steam-raising reactors and the associated fabrication costs make it the most expensive option. Conversely, the simplicity and ease of construction of the spherical, adiabatic-bed reactors make them the cheapest configuration, especially if the vessels are identical.

Total equipment costs for steam-raising and adiabatic-bed designs are greater because of the large contribution of the loop interchanger. And finally, the installed costs are largest for the series adiabatic converters because the large number of items requires more interconnecting pipework and plot area. (Installed costs include civil work, structures, pipework, etc.)

As plant size increases, there is a point at which it becomes cheaper to construct two parallel reactors, says Abbott. The exact limit depends on process details such as the allowable pressure drop, the gas composition in the synthesis loop, and the practical considerations of transporting large, heavy vessels.

Table 4 shows the maximum production from a single converter for three of the four designs. It should be noted that a production rate of greater than 3,000 mtd has already been achieved in a quench converter.

At capacities even greater than 3,500 mtd, the duties of many of the main process-plant items would have to be split. Single-stream methanol synthesis would only be possible with series adiabatic vessels or a radial-flow vessel such as the Toyo MRF-Z reactor.

In the Toyo reactor, the syngas flows through alternate, concentric adiabatic and steam-raising beds. The design allows a large catalyst volume to be contained within normal vessel diameter limits while achieving low pressure drop.

RETROFIT OPTIONS

Abbott evaluated three schemes for retrofitting existing methanol plants to increase capacity. Typical incremental capacity increases are given, but benefits for specific plants should be evaluated on a case-by-case basis.

COMPRESSOR INTERSTAGE

This option involves installing converter between the penultimate and final stages of the syngas compressor (Fig. 8). It allows a rate increase if the machine has a power limitation.

Flow through the final stage is reduced because of the removal of the methanol and water formed. This allows more syngas to be supplied to the suction of the machine, to bring it back to maximum power.

It is important to minimize pressure drop in the converter and other equipment to realize the full potential of this retrofit, hence a radial-flow converter design is suitable.

PURGE GAS

In general, valuable unreacted carbon oxides are lost, along with inerts, in the loop purge stream. A tube-cooled converter can be installed on this stream to make additional methanol (Fig. 9).

A good equilibrium yield of methanol can be achieved because purge gas initially has a low methanol concentration and the reactor can be operated at loop pressure and with a low exit temperature. And because pressure drop is not usually critical in the purge stream, the converter diameter and catalyst volume are quite small.

For instance, the tube-cooled converter for the purge of a 2,000 mtd plant operating at 100 bara requires only 3-4% of the catalyst volume in the main converter. The loop can continue operating at 100 bara, in which case output is increased by 40 mtd.

Alternatively, the loop pressure can be reduced, allowing an increased syngas flow to be delivered from the compressor, and increasing output by as much as 130 mtd.

This retrofit may be especially beneficial for loops that have low carbon efficiency. It is simple and can be installed with minimum disruption to plant operation.

PARALLEL LOOP

To achieve increases of 20% or more, extra converter capacity, additional heat-exchanger surface area, and increased circulation capacity are almost always necessary. Substantial extra syngas feed to the loop will therefore be available.

This retrofit can be used for any of the four converter types. Using a steam-raising converter can be convenient because the reaction heat can be recovered as steam, to be exported to the site steam system. This scheme avoids having to integrate the new loop with existing saturator water or boiler feedwater systems.

Operators with two parallel loops have an extra degree of flexibility. The proportion of syngas delivered to each loop, depending on catalyst performance, can be varied to optimize production.

When choosing between converter types or retrofit schemes, operators must consider the consequences of all options on other loop equipment, and their effects on the overall flowplan.

Copyright 1993 Oil & Gas Journal. All Rights Reserved.