(331b) Development of PSA System for the Recovery of Carbon Dioxide and Carbon Monoxide From Blast Furnace Gas in Steel Works

Development of PSA
System for the Recovery of Carbon Dioxide and Carbon

Monoxide from Blast
Furnace Gas in Steel Works

H.Saima, Y. Mogi, T. Haraoka,

Environmental Process Research
Department, Steel Research
Laboratory, JFE Steel Corp,

Climate exchange by carbon
dioxide became the serious and global problem.  In Japan, steel works discharge carbon dioxide about 15% of
total discharging amount.  From
this point of view, The Japan Iron and Steel Federation (JISF) starts a new
project, called  "COURSE 50" (CO₂
Ultimate Reduction in Steelmaking process by innovative technology for cool
Earth 50).  The authors are trying
to develop separation and recovery technology of carbon dioxide and carbon
monoxide from Blast furnace gas by utilizing PSA technology.

This trial seemed to be easy
by binding PSA for carbon dioxide and one for carbon monoxide.  However, the amount of Blast furnace
gas is enormous, such as more than several million m³/hr and the
concentrations of both carbon dioxide and monoxide are only around 20%.  Furthermore, it is necessary to built
up low cost process for the commercialization.

At first the authors compared
13 kinds of adsorbent from the market by isothermal curve and choose 2 kinds of
adsorbent.  These adsorbents are
tested by PSA testing apparatus in the laboratory.  The inner diameter of adsorption tower in testing apparatus
is 43mm and the height of it is 500mm long.  We found Zeolum F-9 was suitable for carbon dioxide
separation because of high adsorption capacity and adsorption selectivity.

From the importance of
recovery cost for the commercialization, the authors estimated the recovery it
at laboratory PSA apparatus, which was shown in photo 1, in the various
separation conditions.  Higher
adsorption pressure leads higher recovery ratio although it also leads higher
power consumption at blower.  Simultaneously,
lower desorption pressure leads higher recovery ratio and higher power
consumption at vacuum pump. As the results of experiment and power estimation
at blower and vacuum pump, it is suitable that adsorption pressure is around
150-200kPa (absolute pressure) and desorption pressure is around 10kPa
(absolute pressure).  Furthermore,
cycle time is key factor of cost for construction.  The authors thought the standard cycle time was 630
seconds.  As the results of
laboratory PSA apparatus, which was shown in Figure 1, any change in separation
ability was recognized among the results of cycle time at 450, 630 and 720
seconds.  And slight change was
observed in the results of cycle time at 300 seconds.  The effects of CO2 concentration in raw gas were also
tested.  Higher CO2 concentration
leads higher CO2 recovery ratio. As shown in Figure 2, when CO2 concentration
is 32% in raw gas, CO2 recovery ratio rises up to 1.6 times higher than one in
original CO2 concentration even the ratio in CO2 concentration in raw is only
around 1.4.  This change is
important because CO2 concentration in blast furnace gas is thought to be
increased in the future development. 
Injection of hydrogen to blast furnace is studied in the same
project.  If hydrogen is injected,
CO2 in the blast furnace gas will be increased because of operation conditions
of blast furnace.  By these
results, recovery cost estimated to be 63% of original cost suggested.

Base on these results, bench
scale PSA plant called "ASCOA-3" (Advanced Separation system by Carbon Oxides
Adsorption) was constructed as shown in Photo 2.  The planned capacity of ASCOA-3 is 3 tons of CO2 per
day.  The inner diameter and height
of adsorption tower are 0.6m and 1.2m, respectively.  Figure 3 shows the flow diagram of ASCOA-3.  The actual blast furnace gas is
compressed up to 300kPa, cools to 10 degree centigrade and removes water and
sulfur compounds by adsorption with silica-alumina gel and activated carbon
before feeding to PSA unit.  At the
first operation called RUN100, ASCOA-3 shows its enough ability.  CO2 recovery rises up to 4.2tons per
day and CO2 recovery ratio and purity exceeds the target.  At the second and third operation
called RUN 200 and 300, various operation conditions were tested and actual
power consumption was measured in these operations.  Especially in RUN 300, CO2 recovery rises up to 6.2 tons per
day with shorter cycle time (225 seconds) and high CO2 concentration in blest
furnace gas (34%) even at low adsorption pressure (150kPa).  From these results, recovery cost (63%
of original one), which is suggested in laboratory PSA apparatus, is confirmed.  And further operation called RUN 400 is
now started to reduce cost to the target (50%).

Furthermore, 9 thermocouples
were installed in a adsorption tower. 
They can measure the temperature of center, half radius and wall side of
3 height.  At adsorption step and
rinsing step, rise of temperature which is caused by adsorption of carbon
dioxide is observed from bottom to top. 
On the other hands, temperature goes down simultaneously at desorption
step.  Adequate adsorption, rinsing
and desorption time is easily known by this method.

Those results ware obtained at
small PSA testing apparatus compared with commercial sized plant. The diameter
and length of adsorption tower in commercial plant is presumed to be 6.5m and
20m, respectively.  The volume of
adsorption tower of commercial plant is near 1 thousand times larger than that
of ASCOA-3.  Especially, there are
possibilities that the gas pressure and the velocity in the adsorption tower
will not be uniform because the adsorption tower is enormously large as
mentioned.  The gas pressure
distribution of lower face of adsorbent is calculated with "Fluent".
Unexpectedly the results of calculation clearly show that pressure distribution
is homogeneous and the maximum difference is only 0.03kPa as shown in Figure
4.  This means there needs no
special design to the adsorption tower from the viewpoint of pressure
distribution.

  COURSE
50 project including this study is sponsored by NEDO (New Energy and Industrial
Technology Development Organization). 

Photo 1 Laboratory PSA Apparatus

Figure 1 
Influence of Cycle Time

Figure 2 Influence of CO2 Concentration in Raw Gas

 Photo
2  Bird View of ASCOA-3

 Figure
3  Flow Diagram of ASCOA-3

Figure 4 
Calculation Results of Pressure Difference in Commercial Adsorption
Tower