The Compagnia Hidrogeno de Paraguana (CHP) plant -- PDVSA Paraguana Refining Center in Venezuela

New Hydrogen Plant Design Achieves Low Cost and High Efficiency

by James D. Fleshman, Lawrence J. McEvoy, Thrun Vakil and Vinay Khurana

July 1999    Cryogas International      (Subscribe)

In September 1997, a new grassroots hydrogen plant was commissioned by a joint venture between BOC and Foster Wheeler. The Compagnia Hidrogeno de Paraguana (CHP) plant was completed two months ahead of schedule and under budget. It is currently delivering up to 50 mmscfd of high purity hydrogen to PDVSA's Paraguana Refining Center in Venezuela.

PDVSA operates two of the world's largest refineries at Amuay and Cardon on the Paraguana Peninsula. Both of them are complex refineries that make a wide range of products for Venezuelan domestic use and for export to the U.S. market. In 1994, the Amuay refinery (Lagoven at the time) already had over 100 mmscfd of H2 capacity, and more was needed to meet U.S. gasoline specs.

Choice of Build, Own and Operate (BOO)

The user generally has three options for meeting the growing needs for hydrogen:

• Expand existing hydrogen production or recovery
• Install and operate a new plant for incremental hydrogen needs, or
• Purchase hydrogen from an industrial gas supplier (BOO).

Until recently refineries normally have selected revamping or building their own new plant, particularly where a reliable hydrogen supply is critical to plant operations and profitability. In current times however, with depressed crude prices, tougher environmental regulations and pressure on margins, many users are looking to reduce capital expenditure by purchasing hydrogen "over the fence" from industrial gas companies. This is particularly true where hydrogen can be secured from a company with a proven track record for reliable supply.

The last option is based on hydrogen being supplied by an industrial gas company that would build a plant near or on the user's premises. The gas company would provide the capital and manage the design, engineering and installation of the plant. The plant normally is owned and operated by the gas company in return for a hydrogen supply contract. This is the option (BOO) chosen by PDVSA for the hydrogen supply at CHP.

Benefits to the user in choosing BOO H2 Supply schemes include:

• No upfront capital requirements for the user; available funds can be used elsewhere for other critical needs;
• Lower cost due to use of gas industry rather than refinery standards;
• Increased reliability due to continuous improvement in plant design resulting from incorporating "Best Operating Practices;"
• Lower unit H2 costs due to steam and utility integration, high reliability and preventive maintenance; and
• The possibility of economies of scale by supplying multiple customers from a single plant.

Plant Design Strategy

Hydrogen production is dictated by the following factors:

• Capital Costs: This is typically determined by how well the process is optimized to fit the customer needs and includes engineering, equipment and construction costs.
• Utility Costs: These include fuel feedstock, electric power, cooling water, purge nitrogen, and export steam. They must be integrated into the process and matched to the client's needs and availability.
• O&M Costs: This depends on the number of operators required; maintenance staff is dedicated to that plant or shared with common pool, sparing philosophy, etc.
• Regulatory and area development costs, which include disposal of spent catalyst environmental emissions and noise, flare requirements, permitting, soil and site preparation, must also be considered.

Based on typical U.S. conditions and natural gas feedstock, capital cost is the controlling factor for smaller plants, whereas for large plants, feed and fuel costs govern.

Given the size of the CHP plant, one would expect feed and fuel to be the primary factor in product cost. However, low cost natural gas at this site reduced the contribution of utility cost. This provided an opportunity to make a large change in hydrogen cost by reducing the capital cost of the plant.

Engineering Approach

The market for build-own-operate hydrogen plants requires that in order to minimize capital cost the plant be optimized for the particular site, but at the same time engineering costs be minimized. The solution was to develop a design that could be modified readily to suit other sites. At the time of the award, Foster Wheeler was already engaged in a value engineering program for hydrogen plants, and many of the results were incorporated into the CHP facility.

In order to avoid recycling of engineering and ensure that the latest technology was used, the design team met with equipment vendors early in the process design phase. This also allowed the different standards of BOC, Foster Wheeler, and the vendors to be merged to ensure maximum reliability at the lowest cost.

Having vendor information available during the process design also allowed changes in process conditions to optimize equipment cost. For example, it was possible to make a large reduction in the size of the raw gas air cooler by increasing the allowable pressure drop.

Process Description

The plant uses steam reforming with purification by a PSA unit to produce 50 mmscfd of hydrogen at 99.5 percent purity and 400 psig. The product is then compressed to 970 psig to supply PDVSA. Figure I below shows a diagram of the unit.

Compression

The pressure of the natural gas feedstock could be as low as 375 psig, while H₂ was to be delivered to Lagoven at 970 psig. Allowing for pressure drop through the hydrogen plant, two compression stages were necessary. In order to minimize power requirements only a single stage of product compression was used. This required a relatively high operating pressure for the hydrogen plant, with a suction pressure for the hydrogen compressor of 400 psig. A separate stage of feed compression was also used.

The two services were combined in a single multi-service machine, with one natural gas cylinder and three hydrogen cylinders in parallel. In order to ensure uninterrupted service, a spare 100 percent compressor is provided.

Feedstock Desulfurization

After the feed is compressed, zinc oxide is used to remove traces of sulfur from the gas. After review of Lagoven's experience with feed purification, it was concluded that the natural gas feedstock contained only light sulfur compounds that could be removed by zinc oxide, without recycled hydrogen or a hydrogenation reactor. To provide additional protection, the temperature of the desulfurization reactor was increased to 750^oF rather than the typical 700^oF in order to assure complete removal of the sulfur compounds. Because of the low levels of sulfur involved, only a single zinc oxide reactor was necessary.

Reforming

The design took advantage of improvements in reforming technology to assure that hydrogen could be reliably produced at the high operating pressure. This included the use of Foster Wheeler's Terrace Wall^TM reforming furnace with HP-modified catalyst tubes. Burners in these furnaces fire upward along the sloped walls above each terrace: the uniform heat flux results in significantly longer catalyst life. In addition, since two firing levels are used, heat input can be adjusted as catalyst activity changes.

The reformer furnace has two radiant cells. The catalyst-filled tubes are placed in a single row in the center of each cell; process flow is downward through the tubes where the reforming reaction takes place. The flue gas heat leaving the radiant cells is recovered in a common convection section, which is mounted above the radiant cells. Heat recovery coils include feed gas preheat, mixed feed (gas + steam) preheat, steam generation and steam superheat.

The reformer furnace process outlet is cooled in the process gas waste heat boiler (WHB) by exchange with circulating feedwater to produce steam. By locating the WHB between the two cells, it can be "close-coupled" to the outlet manifolds, avoiding the need for a long transfer line.

A single steam drum is mounted on the furnace that serves the WHB and the reformer furnace convection steam coils.

Natural Draft Operation

Plant economics dictated that the reformer furnace would not require a combustion air preheat system. The Terrace Wall configuration, with upward firing burners and upward flue gas flow, made it possible to build the furnace as a natural draft unit. There are no fans, either forced draft, or induced draft. Absence of such rotating equipment improves the plant reliability.

Burners

In the Terrace Wall configuration, the burners fire along the refractory-lined walls, essentially parallel to the catalyst tubes. This keeps the flame fully "controlled" against the hot walls, stabilizing the flame and avoiding flame impingement on the tubes. PSA tail gas provides the bulk of the heat input to the furnace, with natural gas for startup and control.

The firing arrangement provides very stable firing of the low pressure, low BTU tail gas fuel from the PSA unit. Once the reformer is heated, the unit can operate with 100 percent PSA fuel, if needed, even with the low NO_x burners required for CHP. It is always necessary to be able to regulate total firing in order to control the furnace.

However, the ability to handle 100 percent tail gas firing saves fuel and provides a useful operating margin. As operating conditions are changed, the composition of the tail gas varies, often "backing out" natural gas fuel. The Terrace Wall unit can maintain a stable flame while firing a high percentage of tail gas fuel, reducing the need to vent tail gas in order to maintain control of the furnace.

Plot Space

The Terrace Wall reformer forms a very compact unit, without compromising accessibility or operability. The basic layout is determined by the two radiant cells: the convection section and the stack mount on top, with the process gas waste heat boiler located underneath. The steam drum is mounted on top of the reformer.

Terminal points for all piping connections to the rest of the unit were made at a single point: this allowed engineering for the reformer and the rest of the plant to proceed independently.

Construction Features Modularization

The arrangement of the radiant cells permits maximum shop assembly and modularization, limited only by shipping and delivery clearances. For this project, each of the two radiant cells was shipped in five modules. Each module was completely shop-assembled with steel casing and structure, refractory lining, burner tiles and the complete radiant coil assembly (radiant inlet header through radiant outlet header) in place. The 10 modules were erected in less than a week. Field welds were limited to eight butt welds for the inlet headers and eight for the outlet headers. The convection section was also modularized.

Fuel Efficiency

As in most designs, the design of the reformer involved considerations of capital cost and fuel efficiency. In this case, the relatively low fuel costs could not justify the expenditure necessary for combustion air preheat or pre-reforming. Low steam values also meant that there was less incentive to generate steam for export: energy savings would have to come from high reformer efficiency.

The solution was to utilize a modified Terrace Wall arrangement that was developed for use in very high fuel cost areas to improve efficiency, originally with combustion air preheat. Fuel savings will depend on the particular design: for the CHP plant the reduction in total fuel firing is estimated to be four percent. The design does not affect reliability, since the maximum tube metal temperature remains unchanged.

Shift Conversion

Carbon monoxide is reacted with steam to form additional hydrogen in the high-temperature shift converter.

A single-stage reactor using iron/chrome catalyst with a copper promoter was selected for shift conversion. The iron-based catalyst provided ease of operation, reliability and the low cost of a single-stage shift system.

A second stage of shift conversion was considered, but would have added considerable complexity to the plant, since the copper-based catalyst requires a separate startup circuit. Figure 4 shows the additional equipment needed to support a low temperature shift converter. In addition, the copper-based catalyst is more sensitive to poisoning or upsets than the hightemperature shift catalyst.

Since residual CO leaving the shift converter is recovered in the PSA unit as reformer fuel, there is only a small gain in plant efficiency if a second stage of shift is added. This small improvement in utilities was not found to be enough to justify the increase in capital cost.

Heat Recovery

Downstream of the shift converter, the gas is cooled against boiler feedwater, then against air and cooling water.

By working with heat exchanger vendors during the process design, and by increasing the allowable pressure drop, it was found that the size of the air coolers could be cut from four bays to two bays.

While the heat recovery system at this site is relatively simple, it was recognized that the design needed to be re-usable for areas with higher energy cost. Therefore, the plot plan allows for addition of a more complex heat recovery system with minimum impact on the other areas of the plant.

After cooling of the raw gas and separation of process condensate, a ten-bed PSA unit is used to purify the raw H₂. Tail gas from the PSA unit is used as reformer fuel. The project team and UOP (the PSA vendor) worked together to optimize the design of the PSA unit, matching the hydrogen recovery to plant requirements in order to get the most cost-effective design.

Schedule

The original schedule for the plant was 24 months from award to completion. The plant was complete two months ahead of schedule, and was completed under the construction budget. A number of factors were responsible for achieving the short schedule and the savings in capital cost.

Project management concentrated on reducing engineering time. The combination of BOC's experience in operation and Foster Wheeler's in engineering of synthesis gas plants allowed the design to be fixed at an early stage.

As was noted earlier, the project team met with vendors early in the design stage to agree on process and equipment requirements, and on design standards. This minimized changes later in the project.

In the field, modularization was used in constructing the reformer and PSA unit to reduce field labor.

Standardization

It was recognized at the start of the project that the competitive market for build-own-operate plants would require minimizing engineering for each project, even though there is considerable variation in site conditions between plants.

The plant was therefore divided into two sections. The first was a group of high-cost equipment items that need to be designed for each project. The second was the central core area that requires a large amount of engineering although the installed cost was relatively low. The layout allows this central area to be re-used with minimal changes from project to project, while the high-cost equipment packages each attach to the central core at standard interfaces.

Once this was done, the plant was laid out in "engineering modules" so that equipment could be added to a particular area without redesign of other modules.

The use of advanced engineering tools also helped to shorten the schedule and minimize engineering costs. 3-D CAD minimized piping effors and will aid in reusing the design in the future. Smart P&ID's link engineering and instrumentation to help in procurement and tracking of bulk materials and instrumentation.

Plot Plan

The Compact layout reduces site requirements as well as piping and structural steel. High-cost piping was identified early in the design, and piping layout could then focus on reducing the length of high-temperature or alloy lines.

The combination of value engineering, more compact plot plan, short construction schedule and modular reformer construction resulted in a significant reduction in plant cost. We estimate that the installed cost of the CHP plant was 23 percent below that of previous plants of this size.

Operations

The plant has shown an average service factor of 99 percent since the initial shakedown period, and has met guarantees for both capacity and efficiency.

BOC operates a number of other H₂ plants that have shown better than 99 percent reliability. Considering the operating experience, the design, provided spares and instrumentation, this plant is expected to operate with better than 99 percent reliability from its second year of operation onward.

About one week of scheduled down time per year is anticipated for preventive maintenance and equipment turnaround for this plant (This was set to suit PDVSA's hydrogen network: only minimal modifications would be required to match a run length of up to three years.)

Editor's note: A version of this paper was presented earlier this year at the National Petrochemical & Refiners Association Annual Meeting in San Antonio, Texas.

James D. Fleshman is a chiefprocess engineer and Lawrence J. McEvoy is a marketing manager with the Fired Heat Division of Foster Wheeler USA, Clinton, N.J.

Tarun Vakil, engineering technology manager and Vinay Khurana, Hyco product manager, are with BOC Gases of Murray Hill, N.J.