Apr-2006

Refinery CO2 challenges: part II

CO2 emissions and hydrogen production costs in hydrogen plants. Understanding how the impact of unit configuration, operating conditions, feed type, fuel costs and other cost factors set the cost of hydrogen production

Joris Mertens, KBC Process Technology

Article Summary

Everyone in the refining business is aware that hydrogen demand is on the rise and will continue to go up in the foreseeable future. Consequently, an increasing amount of the hydrogen consumed at refineries originates from dedicated hydrogen production facilities (internal or over-the-fence) rather than from naphtha reforming or as a petrochemical by-product. Steam reforming generates the vast majority of this hydrogen.

But what sets the cost of hydrogen production through steam reforming? And do we understand the impact of unit configuration, operating conditions, feed type, fuel costs and other cost factors? In addition, a new factor has entered the cost matrix: CO2 emissions. How large are these and how will CO2 emissions from hydrogen production affect the cost of H2 production?

Efficiency
An important parameter in the production cost and CO2 emissions of refining units is (energy) efficiency. For steam reformers, energy efficiency can be defined as follows:
- Energy content of H2 produced
- Net energy required to produce H2

The lower heating value of pure hydrogen is 2.58GCal/kNm3 (3MWh/ kNm3), or 28.66GCal/t (33.3MWh/t). The net energy required for hydrogen production consists of the following:
- Calorific value of the feed minus the calorific value of the non-hydrogen molecules in the hydrogen product stream
- Total fuel fired to the furnaces minus the calorific value of the fired PSA purge (if relevant)
- Calorific value of steam used in the CO2 absorber regenerator (if relevant)
- Calorific value of steam export (to be subtracted)
- Other energy items: compression duty for gas feeds, pretreating costs of naphtha feed, air blower duty, power required to recycle H2 product, and energy requirements of PSA operation. These cost items tend to be small and can normally be ignored.

A unit efficiency over 80% is high, while efficiencies below 70% should be considered poor. Table 1 lists and quantifies the main parameters that will affect steam-reforming unit efficiency.

Hydrogen plants are designed with two different purification systems: pressure swing adsorption (PSA) and CO2 absorption (Benfield). Figures 1 and 2 show the simplified flow diagrams for typical H2 plant configurations with CO2 absorption and PSA respectively.

The energy consumption of PSA systems is negligible. Hence, the regeneration of the absorbent makes the units with CO2 absorption intrinsically less efficient than hydrogen plants equipped with PSA purification. The 8–10% efficiency loss resulting from the absorbent regeneration shown in Table 1 corresponds with a relatively high regenerator reboiling duty of 0.85GCal per tonne of CO2 desorbed.
However, absorbent regeneration is done through reboiling at low temperatures, requiring relatively low-level heat (below 250ºC). Consequently, if there is excess low-level heat in the refinery (for example, low-pressure steam), the lower efficiency of units with purification through CO2 absorption will not lead to an economic penalty.

PSA-designed units are generally more efficient than units with CO2 absorption. This is not only due to the fact that PSA does not require heat input for absorbent regeneration, but also because the units generally are more recent (and therefore more energy-efficient) designs. PSA-designed units tend to have larger, more expensive furnaces than units with CO2 absorption. This cost is normally offset by other cost items required in the units with CO2 absorption, such as the absorber system and low-temperature shift reactor. Flue gas temperature reduction considerably increases unit efficiency. This can be achieved by installing air pre-heat or economisers.

The shift reactor effluent contains large quantities of uncondensed injection steam. The inlet temperature of the air or water coolers after the last shift effluent reactor is a good indicator of how much of the injection steam heat is being recovered. KBC has seen temperatures varying from 110–230ºC, but typically around 150–160ºC. The efficiency improvements in the recovery of injection steam heat listed in Table 1 correspond with a reduction in the air/water cooler inlet temperature of 100ºC (for example, from 210–110ºC) and depends on the steam-to-carbon ratio and feed type.

In units with CO2 absorbers, the shift effluent is an ideal stream to provide heat to the absorber regenerator. Units with PSA purification do not have this obvious low-level heatsink. De-aeration of the boiler feed water is an alternative way to recover low-level heat.

The feed type also has an impact on unit efficiency. This is largely due to the fact that chemical steam consumption is higher for heavier feeds. The heat required for naphtha feed vapourisation or natural gas feed compression is small. Compression of natural gas feed from 5–30 barg consumes around 0.023GCal per kNm3 of hydrogen produced, which corresponds to 0.5% efficiency.

The direct impact on unit efficiency of other operating parameters such as steam-to-carbon ratio, reformer furnace inlet temperature and PSA hydrogen recovery is even less significant. In principle, more injection steam heat will be recovered at higher steam injections, which will compensate for the additional energy required for the extra injection steam. However, it should be noted that the recovered heat is at a lower level than the heat required to produce the injection steam.

Steam exports
As higher unit efficiency results from better injection steam or furnace flue gas heat recovery, steam export is correlated to the overall unit efficiency. Several other factors that only have a small impact on unit efficiency do have a major impact on the steam output of the unit: