Baseline construction: environmental performance
Based on both the specific case studies and on the broader review of the sector provided above, it is
possible to develop a set of recommendations regarding baseline construction for potential steel projects.
Several types of multi-project baselines are developed and discussed here; the analysis also indicates the
kind and extent of the data required.
3.1 Input data
Obtaining energy consumption data useful for standardised baseline construction in iron and steel
production is difficult. Highly aggregated data is available and this can be used to give general indications
about a country’s iron and steel sector, including the extent of the sector’s autonomous energy efficiency
improvements. However, as this data is aggregated for all plant types, it does not indicate whether
improvements are as a result of a change in production routes, product mix or increased efficiency.
Some information is available on observed best practice plants. The data in Table 10 (“best actually
observed” column) represent estimates based on plants that are currently running and of which energy
consumption data are known.


 Data in this table must be treated with caution - plants listed here may be
operating more or less efficiently than indicated in this table. The Corex process is more efficient than the
ISP route. Plants operate typically at 19 GJ/ton, which is very near the design value (CEC, 1999).
Data on global average energy use by production route is scarce. Data might be estimated by extrapolating
from a few important producing countries for which information is available (although, based on existing
literature, even this data is difficult to come across):
1. China, Japan, US and Germany for ISP;
2. China, Japan, US, Mexico and Italy for scrap-based EAF; and
3. Mexico, Venezuela, Iran and Saudi-Arabia for DRI (gas).
Country-specific data on SEC per production route is also patchy. In theory it should be possible, albeit
time-consuming, to gather the data for individual plants. Alternatively, a plant inventory could be carried
out and national energy intensity inferred from weighted average production. For Brazil, only data for the
entire iron and steel sector are available. As indicated before, according to (IISI, 2000) Poland did not
produce any steel via the EAF route in 1998. However, Gilecki (2000) gives energy (electricity)
consumption data for EAF steel production in the same year, thus indicating differences in data that cannot
easily be explained.
The estimated averages used in this study are given in the following table. They should be interpreted with
care and would need international review before being fixed as the values used in establishing baselines. In
the short term, the values as given in the following table can be used as first estimates. 


COM/ENV/EPOC/IEA/SLT(2001)5
28
Table 10: Comparing energy intensities of steel
production in selected countries and processes (GJ/tcs)
Process Best actually observeda Mid-point SEC India d Brazil d Poland
ISP 22 26b 36.4 23 26
Scrap-EAF 7.7 10c 18.8
DRI-EAF (gas) 22 26c 25.3
DRI-EAF (coal) 25 36 35.6
Corex 19
a SEC values represent the lowest SEC observed in a fully operational plant world wide in 1995-1996. Source: (De
Beer et. al., 1999)
b Estimate based on the values of China, US, Japan and Germany.
c Estimate based on the average of the upper and lower values achieved in practice as indicated in (De Beer et.
al., 1999). Because a large share of the production is in developing and OPEC countries, this is expected to be
an underestimate.
d Source: Pyhlipsen (2000).
Emission baselines may also need to take into account improvements in energy efficiency that have
occurred as part of BAU practices in the past. However, such a calculation requires country-specific data
for at least two years and is therefore also difficult to obtain. Some indications for the case study countries
are outlined in Table 11.
As indicated in section 2 and Figure 1, several different final products can be produced from the
intermediate product, crude steel. These final products have different energy intensities (Table 12) and also
different end uses. For example, cold-rolled products are needed for applications such as car bodies or the
outside of refrigerators and their production entails an extra step. To determine total energy consumption in
a cold rolling process, the values for hot and cold rolling should be summed.
Table 11: Indication of past autonomous efficiency improvements
India Brazil Poland
Primary steel production ~0.5-1% per yearc ~0% per yearb -0.6% per yeara
a Source: Worrell et. al. (1997). This figure indicates an increase in Specific Energy Consumption in the period 1980-
1991. Energy efficiency continued to improve between 1991-1995, but at a lower rate (Price et. al. 1999).
b Both Worrell et. al. (1997) and LBNL (1999), indicate only a very small change in efficiency in the Brazilian
iron and steel sector in the period 1980-1995.
c Based on (SAIL, 1996) with indications over the period 1990-1995
COM/ENV/EPOC/IEA/SLT(2001)5
29
Table 12: The differences between hot rolling and cold rolling
with regard to energy intensity on a best practice basis.
Fuel consumption
(GJ/tonne shaped steel)
Electricity consumption
(GJ/tonne shaped steel)
Hot rolling 1.53 0.35
Cold rolling 1.10 0.53
Moreover, the quality of coal and iron ore inputs used can influence the energy consumption of steel
production. Using high sulphur and/or high ash content coal increases the Specific Energy Consumption of
the steel manufacturing process. The quantitative relationship between coal quality and energy
consumption is not exactly known, but for specific cases a rough estimate can be made. In general, a higher
ash content in the coal, or a lower iron content in the ore, will lead to a higher energy consumption
(Worrell, 2000).
Data problems
For all countries studied, acquiring the necessary input data was difficult. In general, national bureaux of
statistics have been the main source of data. The data for India are more complete than those of the other
countries studied (drawing on extensive research as described in Phylipsen, 2000 and fieldwork carried out
in 1998/9 by the main author). Difficulties experienced are production data for electric arc furnaces and
energy consumption data for electric arc furnaces. Data on the product mix from electric arc furnaces are
also difficult to obtain at a national level. On the plant level, these data are available, so this information
could be supplied for a refurbishment CDM project is under development.
For Brazil, data have been gathered within the INEDIS network6
.


 However, the distinction between the
different process routes is relatively difficult to obtain both for production data and energy input data. On a
national level, the product mix and energy inputs are relatively well known for the iron and steel sector as a
whole.
For Poland, the production data from (IISI, 2000) are not in agreement with the data for energy use as
given by the Polish national energy agency (Gilecki, 2000). It is beyond the scope of this case study to
determine the “real” values, so extrapolations have been used.
Data on crude steel production are also not always readily available. However, the values for developing
the baselines could be determined with additional research. Similarly, there are problems with obtaining
the data required for the adjustment of benchmarks for product differences: data are available, but
additional research in the countries is required to assess its accuracy.