ecologia industrial acero

Upload: ruselkis-josefina-flores-de-gonzalez

Post on 21-Feb-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/24/2019 ecologia industrial acero

    1/66

    THE hTDbSTRIALECOLOGY OF STEEL

    4

    FINAL EPORTTo

    OFFICEOF BIOLooICAL

    AND

    EMENTALRESEARCH

    U.S. DEPA.RT OF ENERGY

    AWARD

    NO.

    DE-FGO2-97ER62496

    BY

    -.

    ,.. . .....

    ~ ~

    ___...

    .._...-,

    ~

    ...., ..._._-

    i .

    - . I. . .

    .

    .

    , .. . . _i_,_

    .

    . . .

    .

    Ahma-i26,2001

    . . ...... . .

    ..

    - ..

    ..

    .

    .-

    .

    -

    . .

    .

  • 7/24/2019 ecologia industrial acero

    2/66

    This report was prepared as an account of work sponsored by an agency of the

    United States Government. Neither the United

    States

    Government nor

    any

    agency

    thereof, nor any of their employees, makes any warranty, cxpnss or implied, or

    assumes

    any legal liability or responsibility for the accuracy, completeness, or use-

    fulness of any inform ation, ap pa rat d, prcduct. or process disclosed, or represents

    that its

    usc

    would not infringe privately owned rights. Reference herein to any

    spe-

    cific commercial product, process, or service by trade name, trademark, manufac-

    turer, or otherwise dots not necessarily constitute or imply its endorsement. recorn-

    mcndktion. or favoring by the United States Government or any agency thereof.

    The

    views

    and opinions

    of authors

    expressed herein do not

    necessarily

    state or

    reflect those of the United Sta tes Government or a n y agency thereof.

  • 7/24/2019 ecologia industrial acero

    3/66

    ._

    . .

    DISCLAIMER

    Por t i ons of this document may be ilkgible

    .

    in electronic image

    produced

    f rom the

    document.

    products. Images

    are

    best available original

    . .

    . .

  • 7/24/2019 ecologia industrial acero

    4/66

    ABSTRACT

    This study

    performs

    an integratedassessment

    of

    new technology adoption in the steel

    industry. New coke,

    iron,

    and steel production technologies are discussed and their economic

    and environmental characteristics are compared. Based upon

    detailed

    plant level data

    on

    ost

    and

    physical hput-output relations by process,

    this

    study develops

    a

    simple mathematical

    optimization model of steelprocess choice. This model is then expanded to

    a

    life cycle

    context,

    accounting for environmental emissions generated during the production andtransportationof

    energy

    and

    material

    inputs into steelmaking. This life-cycle optimization model provides abasis

    for evaluating the environmenM impacts of existing and new ironand steel technologies. Five

    different plant configurations

    are

    examined from conventional integrated steel production

    to

    completely scrapbased operations. Two ost criteria are used o evaluate technoogychoice:

    private and socialcost, with the

    latter

    ncluding the environmen&ldamages associatedwith

    emissions. While scrap

    based

    technologies clearly generate lower em issions in mass

    erm,

    their

    damage cost

    estimates reported in the literature suggests that the social costs

    associated scrap-

    based steel production s slightly higherthan ntegrated steelproduction. Thissuggests adopting

    a W -cycle viewpoint can substantiallyaffect environmental assessmentofnew technologies.

    Finally, this study also examines the impacts

    of

    carbon taxes on steel production

    costs

    and

    technology choice.

    - -

    ._

    _. _ _ .emissions

    of

    sulfur,dioxide-andnitrogen.oxides

    re

    significantly higher.

    Using

    conventional

    . . . . . . . . .

    .

    ..........

    ., .

    _.

    .............. . . . . . . . . . . . . .

    ii

    . . . .

    ....

    -

    . . . . . . .

    . . . . .

    _

    .

    -.

    . . . . . .

    . . . . . . . . . . .

    . .

    . . .

  • 7/24/2019 ecologia industrial acero

    5/66

    Y TABLE OF CONTENTS

    ABSTRACT ................................................................................................................................I

    CHAPTER 1. CNTRODUCTION

    .........................................

    ............................................ ........

    CHAPTER

    II

    .

    NEW

    TECHNOLOGY

    IN

    THE STEEL

    INDUSTRY

    .................................

    3

    COKEhMCING

    .........................................................

    .................................................................... 4

    No~polluting oking

    ofcoal

    .................................................................................................

    6

    Pulverized Coal Injection ...................................................................................................... 7

    DIRECTE D U ~ O N

    ...................................................................................................................

    8

    Gas-Based Direct Reduction..................................................................................................

    8

    Midrex................................................................................................................................ 9

    HYL .....................................................

    :

    ............................................................................

    9

    Atex

    ..........

    :......................................................................................................................

    9

    . . IronCarbide: .........................................................................................................

    10

    Circofer

    &

    Circored

    .........................................................................................................

    10

    Coal-Based Direct Reduction

    ..............................................................................................

    11

    &rex

    ................................................................................................................................ 11

    . . . . .

    HTsmelt

    ............................................................................................................................. 12

    AISI-DOE

    irect Steelmaking........................................................................................ 12.

    DxOSL._tE................................. ....................% ........................................................... 13

    F

    .............................................................................................................................

    13

    Finmet .............................................................................................................................. 13

    OXYGEN TEELMAKtNGAND ........................................................................................ 14

    SCRAP-BASEDETAL RODU~ON........................................................................................ 14

    CASTING ................................................................................................................................... 16

    CHAPTER III

    .

    EMISSIONS FROM IRON AND STEEL INDUSTRY

    ...................

    7

    COKEMAKING............................................................................................................................ 17

    IRONMAKING

    .............................................................................................................................

    18

    STEELMAIUNG

    ..........................................................................................................................

    18

    FACTORS

    FFECTING

    E~SSIONS.............................................................................................

    19

    Raw Material Qualiw.......................................................................................................... 19

    Sulfur

    Content

    ..................................................................................................................

    19

    IronContent ofOre. and Ore Type.................................................................................. 20

    Fuel Choice and Quality .................................................................................................. 20

    .

    Physical

    form

    ......................

    :........................................................................................... 20

    hpmtles

    in

    ~ p u ~

    .........................................................................................................

    21

    Control

    Equipment

    ............................................................................................................... 21

    Choice

    of

    Control Technology......................................................................................... 21

    Efficiency of Control Technology....................................................................................

    22

    Process Characteristics

    ....................................................................................................... 22

    Scrap-Using vs Raw Material Using Technologies......................................................... 23

    Coke

    vs Coal

    Using Technologies

    ...................................................................................

    23

    ..

    . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    . . . . .

    . . . . .

    . . . . .

    . -

    . . . . . . . . . . . . . . . -

    ..-.-.

    -

  • 7/24/2019 ecologia industrial acero

    6/66

    Batch vs Continuous Processes........................................................................................ 23

    Y

    Open

    vs

    Closed Systems.................................................................................................. 24

    Positive

    (High,

    Low)

    Vs

    Negative Pressure Operation

    ...................................................

    24

    Byproduct vs Non-Recovery Coking............................................................................... 24

    Degree of

    Combustion.....................................................................................................

    24

    Duration ofOperations

    ....................................................................................................

    25

    Desired

    Degree ofMetallization...................................................................................... 26

    Product qualiv

    .................................................................................................................

    26

    ReguIat0i-y Standards and Monitoring Stringency .............................................................. 26

    Economic Vmiables

    .............................................................................................................

    28

    High

    temperature

    Vs

    Low

    temperature

    process

    ...............................................................

    25

    CHAPTER

    Y

    .

    THE

    NVIRONMENT & NEW STEEL TECHNOLOGIES

    ...................9

    COKING TEC OL0GI ES ............................................................................................................ 29

    ................................. ......................................................... 32

    ......................................................................................................... 33

    CHAPTER V

    .

    THE

    IFE

    CYCLE

    ECONOMIC MODEL

    .................................................

    5

    THE

    ECONOMICROCESS ODEL............................................................................................

    35

    LIFEYCLE

    MODEL

    ODIFICATIONS

    ....................................................................................... 36

    S~~~ELTING-OLOOIES ........................................................................................................ 29

    . . . .-. -.?

    Process Emission Coeflcients

    .............................................................................................

    36

    .Ups.e.Acti.e. ...=....==..=...=.....-.=.......................................................................... 38

    New

    Technology

    ..................................................................................................................

    39

    Electric Arc

    F

    ....................................................................................................... 39

    Non-Recovery Cokemaking............................................................................................. 41

    DirectReduction.............................................................................................................. 41

    Natural Gas ShadowProcesses.....:................................................................................. 42

    Model

    Calibration

    ...............................................................................................................

    43

    MATHEMATICALT A ~................................................................................................... 45

    CHAPTER VI: MODEL

    SIMULATION

    RESULTS............................................................

    7

    BASELINEECHNOLDGY COMPARISONS................................................................................... 47

    CARBONTAX POL1.............................................................................................................. 49

    Prlvafe Cosf Minlmfiatlon...................... .....................................................................

    51

    CHAPTERVII.POLICY IMPLICATIONS..

    REFERENCES..,..........

    ............................ ..........W.....................

    ............................................... 53

    APPENDIX

    A:

    ELECTRICITY BALANCE

    .........................................................................

    6

    ...

    .

    APPENDIX B: PROCESSUNITS..........................................................................................

    57

    APPENDIX C: COMMODITIES.......................................................................................... 1.58

    60

    PPENDIX D: EMISSIONS IN THE PROCESS MODEL

    .................................................

    iv

    ....

    . - ..

    1

    . . .

  • 7/24/2019 ecologia industrial acero

    7/66

    LIST OF TABLES

    Table

    1: Resource

    use

    and

    env irom enta l performance ofcok ing technologies

    ........................ 30

    Table 2:

    Resource

    use and

    environmental

    erformance

    of

    iron echnologies

    ............................ 31

    Table

    3:

    Comparison of Alternative DIU Technologies .............................................................

    32

    Table 4: Resource use and environmental performance of steel re-g technologies.............. 3

    Table

    6:

    Comparison of energy and

    material

    use for alternative steel technologies.

    .................

    47

    Table 7: Air and solid

    wastes

    emissions fiom alternative steel technologies intons.................8

    Table

    8:

    Incremental

    costs for

    new steel technologies

    (millions

    of dollars)...............................

    50

    Table

    9:

    Steel technology choice under various

    carbon

    axes

    .................................................... 51

    Table Al:

    Comparison

    OfElectricity Consumption

    in Base

    Case..............................................

    56

    Y

    Table

    5:

    Process

    associations

    between

    LCA and the

    process

    mode1

    ..........................................

    37

    LIST OF FIGURES

    Figure 1

    :

    Future

    steel

    plant

    configurations

    ...................................................................................

    4

    .-. -.Figure2: -Comparison-ofX?~-emissions-acrosslternative echnologies...................................

    19

    F i W 3:

    Coking

    duration

    vs

    wall

    t e r n w e

    ..........................................................................

    25

    Figure 4: Relationshipbetween%green

    coke

    and PAH

    emissions

    ..........................................

    27

    Figure

    5

    .Electricity

    consumption

    per

    percent

    DRI

    e

    ...................................................... 40

    Figure 6

    Processmap ofthe LCA coke and coke oven

    gas

    production.

    ................................... .44

    .

    .......

    _. .. .

    .__-..T~.------- - -_-_I ..... . . . .

    V

    . . . . . . . . . . . . . .

    ..-

    . . .

    -., _ - .........

    .~

    -

    ...........................

    -

    ....... . . . . . . . .

    . . . . . . . .

    ....._lX...

    ...

  • 7/24/2019 ecologia industrial acero

    8/66

    CI&PTER I. INTRODUCTION

    The steel industry provides a classic example of anevolving industrial ecosystem.

    During

    the

    late

    19th and early 20th century, ntegmted steel plants in Pittsburgh and Chicago provided

    neighboring communities with coke oven gas for lighting and district heating. Today, these same

    plants recycle off-gases to generate

    steam

    and electricity

    within

    the plant. By-product coke ovens

    were also a major

    source

    ofpetrochemicals during

    this

    era before modem petroleum refining

    became prevalent. Post consumer recycling is also important with roughly 50 percent

    of

    all steel

    in the

    U.S. oming

    from recycled scrap steel. Like the Kalundborg case discussed by Ehrede ld

    and Gertler (1997), these loop clos ing activities slowly developed over time as firms identified

    and characterized waste sources and

    sinks.

    Steel companies in theU.S. ollowed a similar course

    but were driven more by intense competition from producers both home and abroad.

    Technological innovations have always

    been

    mportant in steelmaking (Barnett and

    - .- ._Cmdall, 1986),,SteeLmills use one of two types of furnaces o make new steel. Both

    f uraaces

    recycle old s teel into new, but each is

    used

    to create different products for varied applications.

    The

    first,

    the basic

    oxygen furnace

    (BOF),

    uses

    about

    28

    percent steel

    scrap

    to

    make new steel.

    The other 72 percent is molten

    iron

    produced h m last furnaces,which require

    iron ore

    fiom

    mines, limestone fiom quarries, and coke from batteries ofovens. The BOF furnace produces

    d o r m and

    high

    quality fiat-rolled steel products used in

    cans,

    appliances, and automobiles.

    The other type of steelmaking f i unace, he electric

    scrap

    to make-new steel;-Steel mim'mills-using-these

    totaiU.S.

    teel production. This steel is used

    such as steel plates, rebars and

    structuralbeams.

    Steel

    than

    ntegrated d s ecause they do not require blast furnacesand coke ovens. Their reliance

    on steel scrap also

    affords

    them anenvironmental advantage in lower energy and virginmaterial

    consumption.

    that can yield relatively highquality sheet steel. Thisadditional competitive forcecomes at a

    time

    when many integrated steel

    firm

    are seriously

    reevaluating

    their plants

    in

    light of the recent

    regulationscontrolling oxic emissions

    iom

    coke

    ovens.

    Most existing methods of producing

    coke generate Iigitive emissions &it

    contain

    potentially carcinogenic substances, such as

    benzene soluble organics (BSOs). A variety of strategies, some entailing additional investment

    andor hi@er operating costs, can educe these emissions. Considine, Davis, and hk akov it s

    (1992) conducted a study estimating the benefits and costs

    of

    coke oven regulations,

    incorporating closure decisions and new technology adoption, and found

    that

    investment in new

    coking technologies

    is

    profitable under the new regulations.

    Inkt nland Steel is currently building a large battery of coke ovens using the

    Thompson non-recovery process, heralded as a possible clean technology breakthrough. This

    design allows the controlled burning of coal that destroys the Benzene Soluble Organics

    (BSOs)

    and other potentially carcinogenic

    compounds

    contained

    in

    the offgases of the coking process.

    There are, however, relatively large amounts

    of

    sulk dioxide emissions h m he waste heat,

    which can be recovered via heat exchangers and used toproduce steam for electricity generation.

    Minimills have entered the last domain of integrated steel, employing thnslab casting

    . .

    .

    _ -

    _ ._..---

    .

  • 7/24/2019 ecologia industrial acero

    9/66

    Iron

    &

    Steel - 2

    0

    There

    are

    several other

    nev

    iron

    and steelmaking technologies that could either

    substantially reduce or eliminate coking coal consumption in the production of steel. Pulverized

    coal injection (PCI), replacing up

    to

    40 percent of the coke needed in ironmaking, is widely used

    inEurope, Asia, and Japan and

    is

    now gaining favor in the United States (McManus

    1992a).

    Natural gas injection isa similar technology being promoted by the

    Gas

    Research Institute

    (Brooks

    1992).

    There are also threenew steelmaking technologies that could totally eliminate the need

    for coke. First, there

    is

    d rect reduction DR), a coal or

    natural

    gas-based

    ironmaking

    process,

    that producqs an iron substitute for scrap in electric arc finances. Another coke eliminating

    option is the Corex process, which does not requi recoke and produces a large volume ofwaste

    heat that

    can

    be used

    to

    cogenerate electricity. Finally, there is direct steelmaking (DSM),

    a

    process that could eliminate the need for coking and ironmaking

    in

    traditional integrated steel

    mills. Unlike

    PCI

    DR, nd CORFX, this echnology is currently not under commercial

    development. -

    these technological options.

    our

    analysis

    is

    based

    upon

    an engineering-txonomic model of steel production with environmental

    coefficients from a life cycie assessment (LCA) of steelproduction tiom primary resource

    extraction

    to

    the plant gate. Our model selects the optimal combination

    of

    activitiestominimize

    cost subject

    to

    a numberof constraints, includingmass and energy balances for

    intermediate

    products. Substitute activities represent new technologies available for possible

    many aggregate process models, cour.mociel is -fora-specific steel plant with cde

    uponactualoperating~ormance.

    The analysispresented below uses thismodel in three

    ways.

    The

    first

    application

    compares the economic and environmental

    performance

    of steel technologies, ensuring

    technical

    feasibility.

    This

    analysis provides insights into the tradeoffs between cost and environmental

    objectives, such as educing greenhouse

    gas

    emissions, toxic discharges, and acidic residuals.r

    The second a p p l i k o n solves the model under

    two

    different definitions of

    cost:

    private and

    social. The latter includes private costs and the environmental damages associated with the LCA

    impacts. This

    allowsus

    o determinewhether steelproduction echnology would

    be different if

    environmenU externalities were internalized through a system

    of taxes or

    permits. Finally,

    we

    examine the impact of carbon

    technological development in the steel industry. A

    discussion

    of fkctok

    affecting emissions appears in Chapter

    III. A

    comparison of the resource

    and environmental characteristics

    of

    alternative steel production paths appears in Chapter IV.

    The development of the economic-engineering process model with Me-cycle impacts is

    ChapterV.

    This

    model is then used in Chapter VI o compare the cost and environmental

    emissions for technically feasible plant configurations using these technologies. The

    final

    chapter

    summafizes our

    main points and discusses the policy implications of

    this

    research.

    - - . - - ~ i

    -.-,

    ia.-*.Z.=-,.

    This study prese

    technology choice in the steel industry.

    The next chapter dis

    . .

    . .

    . .

    .

    .

    .

    .

    . ..

    -.

  • 7/24/2019 ecologia industrial acero

    10/66

    Industrial Ecology

    -

    3

    CHAPTER

    11.

    N~TECHNOLOGYIN

    HE

    STEEL

    INDUSTRY

    Competition and recent environmental regulations are inducing technological innovations

    that

    will

    transformmetal refining and smelting. The steel industry has been evolving fiom

    a

    highly capital intensive, batch processing production technologies to less capital intensive,

    continuous

    processing systems that

    are

    cleaner and

    more

    energy efficient. There

    are

    many

    hdications

    suggesting that this ransformation is accelerating. Moreover, faced with ever more

    stringent environmental controls, many

    firm

    are redesigning their production process

    to

    eliminate pollution or to u t i l i wastes

    as

    resources.

    This strategy is gaining hold in many

    industries and

    is

    likely to become more widespread as firm learn that reducing pollution in some

    cases may lower energy and material costs. Investments in new technology often hold the key to

    these cost savings. Understanding the key characteristicsof these technologies and their

    prospects

    for

    commercial

    development

    is

    the primary objective ofthis chapter.

    _-V~,beginut discussion by

    .providing

    an overview

    of

    the production process

    to

    establish

    a context for

    our

    discussion ofnew ironand steel production technologies. With the recent large

    Capacity

    additions by steel

    minimill

    companies in flat rolled sheet production, there

    is

    increasing

    concern

    about the

    availability ofiron

    bearing raw

    materials. As a

    result, new iro

    technologies, ironwaste recovery systems, and steel scrap purification techni

    developmentthat could potentially offer steelmakersmore l e x i b ~ t yn their raw material

    choices. Further down the production line, the success ofthin

    -

    castingappearstobeusheririg

    in developing &~g~chnnaogies---- --- ----

    - - - =

    --

    The distinction between integrated plants and steel m i n im i l l s is beginning

    to

    blur.

    Most

    steel plants in the United States are either traditional ore based integrated plants producing high

    quality sheet and stripproducts or traditional scrap based electricarc h c e EAF) plants

    producing bars, wire, structural shapes, and other long products. These polar opposites

    appear

    n

    Figure 1.Note the difference in scale and the greater number of steps involves

    with

    integrated

    steel production. Recently, some hybrid plant designs are emerging. Perhaps the best example is

    the Nucor plant in Cradordsv ille,

    Indiana

    that produces sheet products using athin slab caster

    fed

    by

    steel

    produced

    in E A F s using

    a

    mixture

    of

    scrap and

    iron

    carbide, a new ferrous material

    input. Thisplant configuration

    appears

    in the third sectionof Figure 1.

    Another

    variation

    involves integrated firm adopting

    -

    steel technologies, such as direct ironmaking,

    advanced EAFs, and thin slab casting, a configurationknown as mini-integrated (see section 2

    of

    Figure 1). The concern over the availability of

    high

    quality ironmaterial inputs appears to be

    main

    impetus

    or these innovations. Energy costs, capital availability, transportation costs, and

    access toproduct markets are additional considerations. While there may be no unique optimal

    mill configuration, the trend is toward so-called market

    mills

    that supply

    a

    market niche within

    a

    geographical area.

    modernize across the entire production process

    from

    cokemaking

    to

    rolling and finishing

    operations. For nstance, Wand Steel constructednon-recovery

    coke

    ovens and most ntegrated

    firms

    in the US are substantially raising their rates

    of

    pulverized coal injection into blast

    furnaces. Minimills continue to advance electric arc

    furnace

    echnology. There is

    a

    g r o h g ist

    New te c h d o g y adoption continues across all plant types. Integratedmills continue

    to

  • 7/24/2019 ecologia industrial acero

    11/66

    Iron

    & Steel

    -

    4

    of new ironmaking technologies designed to provide a substitute for

    high

    quality scrap, which is

    in greaterdemand since

    d s

    re employing new casting technologies to produce high

    quality products for appliance and automotive markets. Innovations in electricarc furnaces, such

    as oxygen

    injection,

    arematching thisnew flexibility in raw material supply.

    are

    discussed in the

    following sections.

    Figure

    1:

    Future

    steel plant configurations

    ConventionaI Integrated Procaw:

    4

    mtll ion

    tons

    peryear

    Sw

    -

    Mini

    Integreted P&tion.-

    1-2

    miIIion tonsperyear

    These innovations

    +

    Continuous

    m

    Thinslab

    AdvancedScrqp Based Production.-1-2

    mill ion

    tonsperyear

    -

    Obsolete Mclter

    Proccssin

    castin

    TmditionuISeng,Based Producton.-0.5-1

    mill ion

    tonsperyear

    PCI

    - pulverizedeoal njection

    EAOF -HeCtric a r ~xygen

    f i vnact

    New

    OSM--cOnvtntional

    orcontinuous

    refining with

    scrap

    preheating

    AdvancedMeltg - Fossil fuel

    or hybrid

    melterwith scrap preheating

    SOW: -(FnUhan.RJ.,d I.,-1995) .

    .

    COKEMAKING

    Coke is

    abasic

    material used o manufacture iron n the s teel industry.

    Iron

    naturally

    occurs

    n oxide

    ores,

    and

    a

    chemical reaction

    called

    reduction

    is

    necessary

    to

    remove

    the

    oxygen,

    leaving the ironinmetallic form. Carbon bonds strongly withoxygen, and coke, the residual

    char

    after heat forces

    he organic volatile matter fiom

    coal, is

    composed primarily of f i xed

    carbon.

    In

    the

    blast f urnace,

    he carbon rom coke reducesironores into molten ironor pig iron.

    Subsequently, controlled oxidation forces

    added

    oxygen

    to

    react with large

    k o u u t s

    of xcess

    carbon eaving anironproduct with about 1% carbon-steel.

    .,

    ,.

    .....

    ...

    _,.

    -

    ..

    .

    ..

    , .

    .

    ..

    . . .

    . .

    .

    ... .. ....... - .-

  • 7/24/2019 ecologia industrial acero

    12/66

    I

    Industrial Ecology

    -

    5

    Since the nineteenth century, the steel industry

    has

    been manufacturingcoke

    in

    by-

    product recovery ovens, designed to capture the volatile matter driven from coal during coking.

    When cooled, the by-product matter condenses into coal tar,a source of many chemical products

    such as synthetic tars, plastics, and crude light oil.

    Operators

    of by-product recovery coke oven batteries are contendingwith several market

    forces and regulatory issues. One-fifth of

    the

    coke oven batteries in the United States

    are

    operating well beyond their expected productive life of

    thi rt y

    to thirty-five years (Peters 1992, .

    18). Emissions

    from

    by-product recovery coke ovens contain everal carcinogenic compounds

    such as benzene soluble organics @SO), which affect public health. The Clean Air Act

    Amendments

    CAAA)

    f 1990 pecify minimumemission limits, Maximum Available Control

    Technology (MAC) and the more stringent Lowest Achievable Emission Rate (LAER),

    intended to reduce air

    toxic

    emissions by 90% by 2003. Ifresidual emissions do not provide

    an

    ample

    margin

    of safety to protect the most exposed individual, EPAwillpromulgate even

    more stringent

    conmlson

    coke oven emissions.

    .. ..

    - ~ -r 1-T .-.--

    I n a C iceoven emission controls

    on

    he united

    States steel i n d w , Considine,

    Davis,

    nd Marakovits

    (1993a)

    use

    an

    engineering-economic

    approachto

    measure the economic consequences of regulation

    and

    its effect

    on

    adoption in the

    steel

    industry.

    In

    a

    later,

    p r e w tudy

    estimating

    he

    costs

    coke oven emission controls, hey incorporate untxx&ty surzounding key parameters, such as

    EPAs controversialestimateofunit

    risk,

    the p

    one UIiit of coke oven emissions (Considhe

    distributions for eight parameters and generate one hundred

    stochastic analysis. As

    in

    the earlier study,

    emissionlimits

    on

    he steel industry in 1995and 1998, espectively, and compare the results with

    the base year solutions to estimate the industry costs due to regulation.

    In the earlier d y s i s , Considine et

    al.

    conclude that coke oven

    emission

    controls would

    accelerate the current trend towards electric 8ty: steelmaking and nonpolluting coking

    technologies (Considine et al. 1993% pp. 452-3). However, lengthy plant construction lead times

    would createcoke

    shortages,

    causing

    imports

    o rise. The older coke batteries shut downunder

    MACT, but newer,

    low cost

    batteries shut down in 1998 or exceeding LAER missionlimits.

    Steel producers have continually adopted innovative steelmaking technologies. Aging

    coke oven batteries and declining cokingcapacities over time have been forcing integrated

    producers to examine alternatives to traditional coking and ironmaking practices (Davis and

    Considine 1992, . 57). In addition, environmental regulations on coke oven emissions may

    stimulate technological innovation though regulators intend the MACT standard to encourage-

    increased use

    of

    proven technologies

    Graham

    and Holtgrave

    1990,

    p.

    243-4).

    Howevb,

    environmental regulations raise the costs of coke production. For example, only twelve percent

    of

    the batteries existing in 1991 can economically achieve the less stringent MACT emission

    l imts,

    and 83% of those batteries requi rerebuilding (Peters 1992,p. 15). As a result,

    environmentalcontrolscould accelerate the steel industrys aatural rate of technology adoption

    towards processes that reduce or even eliminate coke requirements.

    . . . . .

    ~. _ . _

    . . .

    . .

  • 7/24/2019 ecologia industrial acero

    13/66

    Iron& Steel

    -

    6

    The steel industry has se vpil technological choices to reduce coke oven emissions. The

    viability of each option varies by coke plant based on he age and conditionof the batteries,

    existing coke oven emission levels, availability of capital, plant location, and relative material

    and energy prices, among others. In addition to importing foreign coke, steel producers

    can

    retrofit or rebuild existing coke ovens,

    as

    well as

    install

    new w e ta k in g batteries (Considine et

    al. 1993%p. 445). Retrofitting involves replacing the coke ovens doors and their jambs. New

    jambs may necessitate replacement of the refractory bricks, depending

    on

    he condition

    of

    the

    oven walls (Struthers Corporation 1991, p. 14). A pad-up rebuild involves demolishing the

    existing

    battery to

    he

    concrete

    "pad" or foundation and building a new battery of the same

    dimensions in its place (Struthers Corporation 1991, p. 15).

    Besides additional investment in the wet-coking technology, steel producers have several

    new processes

    to

    consider, which we describe below. Coke producers

    can

    nvest innonpolluting

    coking echnology calledJewell. Alternatively, the industry

    can

    adopt ironand steelmaking

    processes, such

    as

    pulverized coal injection, d rect reduction, and scrapbased steelmaking,

    whichsave-oreliminate,wke@to

    coke oven emissions

    to

    re@

    Nonpolluting Coking of Coal

    nonpolluting coke oven presently being

    used

    in

    Vansant, Virginia

    by Jewell Coal and Coke

    Company, owned by Sun Coal Company. Ironically, the Jewell-Thompson

    to theObSXete-ve oven ~ ~ ~ ~ g ~ ~ e e ~ ~ ~ e ~

    oven, the Jewell oven operates onnegative pressure, which

    minimizes

    coal

    tars

    and

    gas

    combust nside the oven and flues (Knoezner et

    al.

    1

    process recovers

    no

    by-products, eliminating the need for by-product recovery facilities

    and

    he

    disposal

    of hazardous wastes. Provisions for the recovery of excess heat will permit

    cogeneration

    of

    electricity (Knoezner

    et al.

    1992,

    p.

    50).

    ;nallv, fcoke producers cannot economicallycontrol

    can

    cease operations.

    CAAA requixe the sdministrator

    of

    EPA to evaluate the Jewell-Thompson oven, a

    A Jewell oven is about twelve feet wide and forty-five feet long,

    accommodating

    a coal

    charge of twenty-five to B y tons (Knoezner et al. 1992, p. 50). Oven charging is by

    a

    cOnveyOr

    machine

    through he pusher

    side door

    and not from above (Knoemer

    et

    al. 1992, p.

    52).

    The

    operatoradmits

    a l i mtedamount of air

    into

    the oven to combust some

    of

    the volatile matter

    being driven fiom the coal, which generates heat required for coking. The partially combusted

    gas

    escapes

    through

    ole flues

    located

    below the oven

    floor,

    combusting further within the flues

    and creating additional heat underneath the coal bed (Knoener et al. 1992, p. 5 1). Coking time

    can

    range between wenty-four to forty-eight hours (Knoezner et

    al.

    1992, p.

    50). Pushing

    and

    quenching

    operationsare

    similar to the wet-coking process (Knoezner et

    al. 1992, p. 52).

    Other

    no-recovery coke even designs include the European

    Jumbo

    Coking Reactor and

    the American Calderon coking process cunently under development may completely eliminate

    all cokesv en emissions

    (Wrona, 1997,

    p.

    60). Dry

    quenching

    is

    another technique to

    reduce

    emissions, recovering heat fiom the hot coke during quenching to generate steam. Thisprocess,

    however, uses more electricity thanother coking processes and requires greater inputs of coking

    coal. Another process is form coking that produces briquettes by drying and partially oxidizing

    ._%_.

    -.

    . .

    . . . . . . _

  • 7/24/2019 ecologia industrial acero

    14/66

    Industrial

    E C O ~ O ~ Y

    7

    lower

    quality

    noncoking p u l v e q d coal with

    steam

    in a fluidized-bed reactor, followed by

    carbonization at higher temperatures (Lankford, et, al, 1985, p. 142). Both of these processes

    offer a relatively environmentally acceptable means

    of

    producing coke and are extensivelyused

    in Japan and Russia but have yet to find extensive use in

    North

    America.

    Aging cokemaking facilities, stringent emission controls

    on

    coke oven emissions, and

    perhaps increased scarcity

    of

    metallurgical coal

    in

    some

    areas

    have forced plant

    operators

    o

    examine alternatives to traditional ironand steelmaking (Davis and Considine 1992, p.

    57).

    Below we discuss several new technologies that lower coke requirements or eliminate the need

    for coke.

    Pulverized

    Cod

    Injection

    coke. To

    reduce

    ironore to

    its

    metallic

    form

    suitable for production, ironmakers can inject

    pulverized coal

    intoa

    blast fUmace proper to lower coke requirements (Davis and Considine

    1992,p.

    3-7).

    One pound ofpu ive rhd coal can

    replace one

    pound of coke

    as

    a fuel and

    as

    a

    reducing agent (Unsworth et

    al.

    1989,

    p.

    1-20).

    Because pulverized

    coal

    cannot

    substi=

    for the

    permeability of porouscoke or support the blast f i vnaceburden, it cannot completely

    replace

    coke.

    Pulverized coal injection

    (PCI)

    nd continuous casting reduce the need for metallurgical

    -

    The U.S.Department of Energys

    project atBethlehem Steel Corporations Burns

    techn d o . - . -

    used in many plants today. In the process, both

    blast furnace in place of naturalgas (or o

    generated by the blast furnace itselfremain

    exiting the blast firmace is clean, containing

    no

    measurable SO2 or NO,. Sulfur

    removed by the limestone flux and bound up in the

    slag,

    which

    is

    a marketable

    addition to the net

    emission

    eduction

    realized

    by coke displacement,

    high

    blast furnace

    production.

    Coal injection

    also

    allows the use of a wide range of relatively

    inexpensive

    coals,

    in contrast

    to

    coke,which

    can

    only

    be

    made fiom certainhigh

    quality

    coals.

    Two

    high-capacity

    blast furnacesat the Burns Harborplant, eachwith aproduction

    capacity of 7,000 net

    tons per

    day of hot metal, have been etrofitted with the coal injection

    technology.

    The

    two units

    will use about

    2,800

    tons/day of

    coal

    during

    full operation,

    eplacing

    about 40% of the coke needed in the furnaces. Bituminous coals with sulfur content ranging fivm

    0.8% to 2.8% fiom West

    Virginia,

    Pennsylvania, Illinois, and Kentucky are to be used. A western

    sub-bituminous coal having 0.4-0.9%

    sulfur

    might be ested also.

    Construction was completed

    in

    February 1995. Bethlehem Steel submitted

    a

    public

    design report in March 1995. Start-up testing

    has

    been completed, and the plant

    is

    complete.

    Operational testing began n November 1995. Furnace C hasbeen operated with an average coal

    injection rate

    of 275

    lbdnet

    ton

    ofhot metal, using low-volatile bituminous coals. Bethlehem

    Steel has determined

    that

    thisinjection ate will be the new operating baseline for Furnace

    C

    for

    all future test coal comparisons. Furnace C also

    has

    been operated with a coke rate of

    approximately 650 lbdnet ton

    of

    hot metal without coal injections, down from

    770

    lbdnet

    ton.

    ...

    , . . .- .

    ...

    .

    ,

    . .

    . . ..

    . I .~

  • 7/24/2019 ecologia industrial acero

    15/66

    Iron & Steel - 8

    Furnace D

    as

    been operated

    with

    a coal injection rate of approximately 190 lbdnet

    ton

    of hot

    metal, which is above its design N i t of 180 lbdnet ton. Bethlehem Steel has completed

    repairs

    to a coalpreparation plant necessitated by tramp organics in recent coal supplies. Bethlehm Steel

    plans to increase substantially the coal feed rate through all 52 tuyeres for comparison

    with

    he

    baseline standard of 275 lbdnet tonof hot metal

    on

    Furnace C.

    The

    granular

    coal injection project at Bethlehem Steels

    Bums

    Harbor plant

    in

    Northern

    Indiana

    offers he U.S. steel ind- a way to become more co mpti tive while impmving its

    environmental performance. The echnology can be applied to essentially allU.S. blast furnaces.

    It should be applicable to any rank coal commercially available in the United States

    that

    has

    a

    moisture contentno higher than 12%. The environmental impacts

    of

    commercial application

    come .primarily

    from

    a reduction in the need for coke, the production of which can elease

    emissions

    of

    sulfurdioxide and

    air

    toxics.

    Also,

    because 8 wide range of relatively low-cost

    coals can

    be

    used

    to

    replace processed coke, the ironmaking process is less expensive. The cost

    of more expensive &els such as

    MW

    as and oil

    can

    be avoided.

    - -

    yx-- - Iu.x- .-7 1-

    -- --I-

    Industrial Ecology - 19

    produce zinc-rich dust that also indudes lead and cadmium. During steelmaking various kinds of

    scrap and oily mill scale are generated. Home scrap, generated during various finishing and

    shaping operations, is recycled back to the EAFBOF. The pickling process during steelmaking,

    in which the contaminated coating on steel is removed with HCl, generates water emissions with

    suspended and dissolved solids.

    Figure

    2:

    Comparison of C02 missions across alternative technologies

    Kilogramsof carbon dioxide per ton hot rolled

    strip product

    2500

    ntegrated Steel DRI+EAF based EAF based

    I

    Mill Mill Minimill

    Source: VAI (Sapphire: Advanced Solutions for

    Waste

    Free Iron

    &

    Steel Plants)

    FACTORS

    3 W X G

    EMISSIONS

    The conversion of iron ore to steel is a complex process involving various processes, raw

    materials, equipment, and physicdchemical reactions. Emissions during these processes are a

    function of several factors, the more important of which are discussed briefly.

    RawMaterial Quality

    coal, iron content of ore,BTU (or fixed carbon) content of he l, and impurities in raw materials.

    These are discussed below:

    Su&r Content

    The sulfurcontent of coal

    has

    a direct bearing on emissions of

    SO,

    and H2S during the

    coking process. Russell and Vaughan (1

    976)

    report that

    as

    an approximation, the percent

    sulfur

    in coke oven

    gas

    is 1.7 times the percent sulfur in coal input to coking units.

    Sulfur

    n coalalso

    affects the sulfurcontent in pig iron, which necessitates a hot metal desulfurizing step during

    steelmaking.

    Quality parameters that influence environmental releases include the sulfbrcontent

    of

  • 7/24/2019 ecologia industrial acero

    27/66

    Iron Steel - 20

    Iron Content

    of Ore

    and Ore

    T p e

    The iron content

    of

    ore determines the bulk that needs to

    be

    crushed, transported and

    handled. These operations

    are

    associated with energy use as well

    as

    C02 and PM emissions.

    Hence processes that use less ore,

    ceterispuribus, are

    associated with less emissions and

    higher

    energy efficiency.

    FuelChoiceand Qual

    The quality of industrial fuels

    -

    coal, fuel oil, esidual oil, and natural

    gas

    - determine the

    amounts of criteriapollutants, and metallic hazardousair pollutant emissions from industrial

    combustion sources within an ironand steelmaking plant. For example, the BTU content of a

    coal determines the

    total

    mass of fie1thatmust be transported, crushed and conveyed.Hence,

    low

    BTU

    coal, even ifpriced

    at

    a discount in the market, exacts its full cost, albeit

    on

    he

    environment.

    YI..--T.-_

  • 7/24/2019 ecologia industrial acero

    28/66

    Industrial Ecology

    -

    21

    Coal too

    is

    available

    in

    many s%s ranging fiom -6mm to -32 mm. Transport and handling

    of

    iron

    ore fines and pulverized mal releases some fine ironore and coal particles in to the

    immediate atmosphere. Processes that use lumpy ores and coarser coals, or processes that bypass

    the use of raw

    materials

    will naturally be associated with less particulate emissions.

    hl J W ' & S ill hp

    Chattexjee (1993) reports that the amount of slag generated during

    ironmaking

    is directly

    related to the amount of gangue present in the iron feed in the heat, The ash content of

    coalis

    a

    fm o r in the generation of slag in blast furnace operations. Since the removal of

    ash

    silica)

    generates COZ during

    slag

    fo m ti on , higher silica content

    in

    coal and ironore generates

    i n m e n t a l C& emissions. Chatterjee (1993) M e r sserts that good quality

    coals

    with low

    levels

    of

    ash,

    s u b

    nd volatile matter are likely to yield higher fuel efficiency, and require less

    additives

    (fluxes

    and detd fbrm m

    ).

    Similarly, EPA (1999)

    reports

    ead emissions

    in

    particulates

    from EAF operationsare a h c t i o n of the lead in the

    scrap

    charged to the process.

    determine

    emissions. A

    case in point

    is

    the emissions

    of

    polyaromatic

    hydrocarbons

    PAH)

    emissions during the quenching operation in cokemaking. Quenching

    pedorined

    with

    con amhated waters releases order of magnitude higher amounts of PAH thanquenching with

    clean water.

    -

    ---Inthwcases, the strength and quality of reagents

    as

    well

    as

    operating practices may

    -----.---x-.- -.--_--- - - ----__-

    ~ -

    ,._--. --

    ontrol Equipment

    Both the choice and efficiency'ofpollution control equipment has an important impact

    on

    emissions n ironand steel production. There is substantial diversity in the typeand efficiency of

    pollution

    control equipment in the industry that complicates estimation of emissions at the

    aggregate level.

    Choke

    of Control

    Technology

    Numerous pollution abatement technologies are available. These technologies are

    typically

    specific

    o one medium - air, water, or solid. Within each medium however, several

    variants exist thatcontrol one or multiple pollutantstodifferent degrees. Particulate abatement

    equipment include simple cyclones, high efficiency cyclones, electro-static precipitators (ESP)

    with a continuugi. of efficiencies,scrubbers, wet ESPs, fabric filters,and ceramic

    filters.

    Cyclones achieve

    about

    a 90%cbntrol

    of

    particulates by

    mass.

    However, their

    capacity to

    control

    finer particles

    (lidWaste Emissions

    Ted

    Blast Furnace

    1600

    up

    to 33% with PC

    0.3 -0.45

    22-34

    0.3

    07 I

    __ i b )

    760 - 1410

    12.25 - 100.7

    SOO(c)

    6.2097

    180(c)

    3.8(d)

    0.25

    17.5 - 25

    1487 (1344

    -

    160s

    1175 (1050

    -

    134:

    0

    60

    656.8 (525

    -

    788)

    0.6

    0.4 - 0.6(a)

    130

    53

    114

    16.5 - 17.28

    60

    0.01

    0.04

    0.42

    50

    145

    95:

    60E

    0.6

    estimatebasedon conversion

    of

    volume share CO2 to mass using exhaust flow of 1650-1728m3/thrn

    FromTable2 t is evident that COREX has several desirable features. First, its iron

    ore

    nput

    coefficient

    (1344

    -

    1609Kgs per

    ton) is modestly

    lower

    than he 1600Kgs of ironore

    required

    for blast furnace reduction. It requires modestly higher electricity and more oxygen thandoes the

    blast h c e eduction process. Second, the process does not require coke, and hence eliminates,

    .

    _.

    ..

    .

    .

    .

    .

    .

    . .

    .

    ..

    .

    .

    .

    .

    .

  • 7/24/2019 ecologia industrial acero

    39/66

    Iron& Steel -

    32

    inl

    1

    .4

    0

    at leastpartidry, the emissions wsociated with cokemaking.

    It

    is also environmentally superior to

    blast f urnacetechnology.

    This

    advantage derives fiom the partial use of hydrogen as a reductant

    rather thancomplete reliance on

    carbon.

    PM and

    SO,

    missions are significantly lower, though

    NOx emissionsmay be higher than hat fiom the blast fimace. In addition, he

    use

    of

    coal release

    some methane during the reduction. C O W lso

    boasts

    of significantreductions

    in

    ammonia,

    phenol and sulfide discharges

    to

    water effluents

    during

    the

    reduction

    process.

    The

    echnology,

    however, produces higher amounts of

    slag,

    possibly due

    to

    the direct inputs of impurity-

    containing coal in to smelting reduction fiunace.

    leso urce Requirements

    ____-.-.._._-I

    Iron

    ore

    Coal

    Coke

    .

    Electricity

    o w e n

    Natural

    Gas

    Water

    Labor

    ir Emissions

    co2

    PM/Dust

    SOX

    NO,

    Methane

    raterEmissions

    Ammonia

    BOD

    Sulfides

    Phenol

    >lid

    Waste Emissions

    Slag

    Dust and

    Sludge

    DRI

    T E C O L O G ~ S

    There

    are

    at least half a dozen, and probably a dozen competing DRI technologies. The

    evaluation here is limited

    to

    those that have

    are

    commercially established.

    All

    DRI rocesses in

    Table

    3

    consume approximately the same amount of ironore. FASTMET and Circofer use coal

    8s

    the reductant,

    wlkreas

    other technolo

    Table

    3:

    Comparison of Alternative

    DRT

    Teclqologies

    use

    natural

    gas for

    providing

    thermal

    energy

    and

    -1 -r

    - - ~. -....

    -- ---1-

    I - . --_

    Units

    . -

    K g a m

    Tondton

    KwH

    per

    ton

    (N)m3lthm

    GcaVthm

    m3Iton

    Manhrs per thrr

    TonsiVlm

    glthm

    m3tthm'

    g/thm

    mgMl

    g m m

    Ton/thm

    Kglton

    DIRECTLY

    REDUCED IRON TECHNOLOGIES

    ClRCOFEl

    1442

    380

    0

    90

    190

    -

    195

    0

    2

    0.315

    - . -_-_-

    50mgMm3

    10- 100

    1.1mgnitre

    CIRCORE

    1483

    0

    0

    100

    0

    2.75

    1.5

    0.31

    5

    -_..- ---?

    50mg/Nm?

    10

    86ngIJ

    FASTMEl

    1380- 143(

    270 - 380

    0

    60 - 90

    0

    0.6

    -

    0.65

    1-1.5

    0.2

    -

    0.3

    1.05-

    1.18

    H n

    11

    1450

    0

    0

    90

    0

    2.55

    -

    2.8&

    1.6

    - 1.8

    0.34

    3.039kgftor

    0.2

    0.05kgEton

    26.7

    -

    m4

    0

    0

    1.97

    i

    -

    ..

    .

  • 7/24/2019 ecologia industrial acero

    40/66

    Industrial E co lo ~ y 33

    U n i l

    carbon

    & hydrogen for reductio% The

    DFU

    processes consume different combinations

    of

    reductants, oxygen and energy and other inputs. Midrex DRI

    is

    relatively electricity intensive

    thanother

    DRI

    technologies. However, consistentwith higher labor c o s t s in the US, abor input

    at 0.16 manhoursper ton ron is the least among he set

    of

    DRI echnologies. Unfortunately,

    emissions

    coefficients are missing for many echologies.

    This

    is acritical

    idonnationgap

    that

    must be filled to

    facilitate

    comparison

    of

    nvironmental performance.

    Technology

    BOF QBOP

    EAF

    1EEEFI G ECHNOLOGIES

    ._ -_--

    -~ ~ m

    m3lton

    KwH per

    ton

    Nm3/thm

    Manhrs per thm

    Manhrs per thm

    Basic Oxysen Furnace @OF) technology once

    dominated

    the steel industry in theUS.

    The EAF ased minimills have, however, gained a substantial shareof he steel refiningmarket.

    A few

    steel mills employ the Q-BOP rocess.Table

    4

    presents he resource use and emissions

    coefficients or BOF, QBOP, nd EAF technologies.

    Table 4: Resource use and environmental perormance

    of

    teel refining technologies

    _----

    0.932- -

    101-201

    0.097

    0.32

    I

    STEELMAKING TECHNOLOGIES

    0.58 -

    38

    0.066 -

    1.w

    0.07

    0.066

    -0.54

    esource Requirements

    IrOdSeiiiii

    Nitrogen

    Electricity

    Oxygen

    Labor

    Maintenance

    ir Emissions

    c 0 2

    - . . ..

    I

    M/ PMlO

    Tondthm

    Lb per thm

    Lb per thrn

    Lb per

    thm

    g/thm

    Ton/thm

    Kgbn

    olid Waste Emisslons

    0.0068

    - 28.5 0.056(a

    0.1 17 - 0.220

    0.0038

    0.02-0.1

    100-440

    8-62

    Ranges

    often

    indicate differences in controlequipment

    0.8

    1.13

    30

    0.048

    0.967

    430 -600

    350

    110-420

    20 -40

  • 7/24/2019 ecologia industrial acero

    41/66

    Iron

    & Steel

    -

    34

    The Q-BOP pparently iqa little more efficient at converting iron to steel. It is

    also

    very

    energy

    efficient,consuming only 30 KwH* per tonof steel, which is at least

    an

    order of

    magnitLlce lower

    thanBOF

    nd

    EAF

    processes.

    EAF

    generate approximately the same amount of slag, dust and sludge. Emission coefficient

    ranges

    overlap for these competing technologies, negating any basis for

    ranking

    them.

    coking,

    ironmaking and

    steelmaking

    echnologies revealsmany datagaps, especially pertaining

    to emissions coefficients. The only notable observation is the a p k n t superiorityof COREX

    over the traditional blast furnace reductionof iron

    both

    in

    term

    of resource use and

    environmentalpedormance.

    Data for comparison of environmental p e r f o m c e ishard

    to

    come by. The BOF and the

    The above examhation of the resource use and environmental performance of alternative

    . . . . . . . . . . . . . . -.

    .......

    i ..-.. . . . . . . . . . . . . .

    . . . . . . . . . . .

    . . . . . . . . . . . . . . . .

    8 according to

    the

    EPRI

  • 7/24/2019 ecologia industrial acero

    42/66

    Industrial Ecology - 35

    CHAPTERV.

    k~

    IFECYCLECONOMIC MODEL

    The sequential

    natureof

    production w ith multiple joint products

    has

    motivated m y

    researchers ta develop linear programming or process models of the steel production process.

    Early studies include

    Tsao

    and Day (1970) and Russell and Vaughan (1976) while more recent

    efforts include those by Sparrow,

    et.

    al. (1984) and ,Considhe, et.al. (1992). These models are at

    the industry level with aggregate inputsutput relations determined either fiom representative

    process

    uni ts

    or

    h m

    ggregate industry data.

    In

    contrast,

    this

    study develops a detailed process model for

    a

    specific steel plant, the

    Mon Valley Steel Works,owned and operated by U.S.Steel Corporation

    (USX).

    The Mon

    Valley Works ncludes the Clairton Coke Works, he largest coke oven batteries in the U.S., he

    historic Edgar Thompson blast fixrnace and BOF mill developed by

    Andrew

    Carnegie, and the

    Irvin finishing

    mill.-

    Thiscomplex

    is

    somewhat unique because it generates excess coke, which

    is

    transferred

    to otherUSXmills and sold to other steel producers. Nevertheless, thisplant is

    representative

    of

    many integrated

    steel

    mi l l s

    in the

    U.S.

    because technical efficiencies

    of

    coke

    ovens, blast furnaces,and other equipment do not vary substantially between firms.

    The model development occurred in two phases. The first step entails the development of

    8 simple

    hear

    programming model based upon estimates

    of

    the costs by process based

    upon data

    provided by Don

    Barnet$

    (1997).

    This

    dorm ation, however, does not provide a

    very

    detailed

    picture-of

    ~ - ~ p r o d ~ & : i a r i dl~cric-@~-bdanr@Sietweenhe

    various process units

    in the plant. The

    ife

    cycle inventory

    data

    collected by Rhodes et. al.

    (2000)

    provides information

    to develop these balances and

    to

    estimate and the environmental emissions by process. The ife

    cycle assessment(LCA)ncludes

    raw

    material extraction through steel product manufacturig.

    Both sources

    of

    infoxmation provide a fairly detailed view of he steel production process,

    including

    iron

    and

    coal

    mining, coking, blast fiunace production, steel production,

    and

    rolling

    and finishing operations.

    naEECONOMIC

    ROCESSMODEL

    world.

    His

    data includes input-output relations, factor prices, and capacities by process.

    Companies operating hese plants voluntarily subscribe

    to

    his service under he condition that

    their detailed process cost

    data

    remainsconfidential. Based upon this idormation, Barnett

    provides analysis that allows companies to detemine where their plant costs

    rank

    n relation to

    others in he industry.

    Barnett calculates the average cost of production by process by

    summing

    the product of

    the inputsutpu t coefficient and the respective input prices, which provides

    an

    average cost per

    unit

    of

    output. The

    per

    unit

    costs

    of intermediate inputs, such as coke and pig iron, hen

    essentially become input prices for

    downstream

    processes, such

    as

    steelmaking.

    'Barnett (1997) collects detailed process

    data

    for many integrated steel plants around the

    Barnett provided uswith his data for the

    Mon

    Valley Steelworks. This allowed the

    specification

    of a

    simple linear programming model with 16 process activities. The

    first

    two

  • 7/24/2019 ecologia industrial acero

    43/66

    Iron& Steel - 36

    include screening of metallurgicql coal and

    the

    production of coke fiom conventional by-product

    slot coke ovens. The coefficients

    of

    these activities include coal, labor, energy, and maintenance

    requirements per unit of coke produced

    as

    well as an estimate of the amount of coke oven off-gas

    generated. The next group of activities involve

    iron

    production, including sintering, pulverized

    coal injection, and blast

    f urnace

    operation. The third group of activities entails steelmaking;

    including basic oxygen f i unaceoperation,

    vacuum

    de-gasification, and continuous casting. The

    final

    group of activities includes rolling and f i s h i n g operations hat produce hot-rolled band,

    cold rolled steel sheet, and galvanized steel products. Our base model solution provides an

    estimate of total variable production costs very close to the costs estimated by

    Bamett

    (1997).

    LIEE

    CYCLE

    MODELMODIFICATIONS

    The life cycle inventory

    data

    provides considerable detail that

    permits

    significant

    enhancement of the process model initially developed using the idormation rovided by Bamett.

    The first improvem&t involves the inclusion of emission coefficients by pn>cess. In general, the

    K d % y B

    &&

    those defined by Rhodes et.al. The required

    aggregations

    re discussed in the next section. Unlike

    Barnett

    (1993, the

    life

    cycle inventory

    data

    ncludes emissions generated in upstream activities, such as coal and

    ironore

    mining, which

    are

    discussed in the second sub-section below. As the previous discussion indicates, our

    surveys

    of the industry did not yield a complete emissions profile ofnew technologies, which necessitates

    several

    assumptionsthat are discussedbelow.

    Finally,

    a carefbl

    comparison of

    emissionsreported

    by Rhodes et. al. with those fiom

    our

    model provides

    a

    good check for 8ccuracy. The complicating factor in

    this

    comparison, however,

    arises fiom the hypothetical nature of the plant output in their study. As mentioned above, the

    Mon

    Valley steelworks is imbalanced, producing substantiallymore coke thanrequired,

    As

    a

    result, he

    Rhodes

    tudy only included those emissions associated with coke requirements at the

    plant.

    Other

    plants consum ing the excess coke would include, at least in theory, the associated

    emissions. From a plant optimization viewpoint, not accounting for these emissions and in

    particular

    the off-gases

    that

    coke ovens generate would be a

    serious

    distortion. Thismodel

    calibration

    issue

    and related

    matters

    are

    discussed in

    more detail below.

    Process Emission Coefficients

    In calculating the emission coefficients, the first step was to match processes

    in

    the life

    cycle inventory data by Rhodes to those

    in

    the base process model

    so

    that emission coefficients

    from the

    Rhodes

    study

    could be properly assigned to processes within the process model.

    Comparison of the two models yielded the following associations and subsequent assignments

    shown

    in Table 1.

    Where single LCA processes

    are

    associated with multiple processes, emission

    coefficients have only been attached to the irst of the associated processes listed inTable 1.The

    units

    of emissions were given

    in

    a variety

    ofunits,

    and allwere converted

    to

    a

    pound

    ~ e rhort

    ton basis tomatch process levels

    as

    defined in the existing process model. All emissionswere

    then added to the existing set of commodities.

    ,

    ---...

    -----------

    -___.?-I.iT--I__- --

    -

    -

  • 7/24/2019 ecologia industrial acero

    44/66

    Industrial Ecology

    -

    37

    COKE-SLOT

    SCREEN

    SINTER

    ~ ~~

    Table 5: Process,associations between LCA nd the process model

    Making

    Coke

    Coke

    Sintering

    GAL-HD

    8305

    HOt-DipGalvanizing '

    @&)

    Galv Hot dipped 8324

    Coke Oven Products

    (mg/MJ)

    No

    changes were made

    to

    the input profiles

    materials,

    abor, operating costs), except

    in a

    very few specific cases noted below. The Barnett data did not include the

    three

    power plants at

    the

    Mon

    Valley works. Their steam output significantly

    alters

    the energy flows

    within

    the plant.

    The ife cycle datacollected by Rhodes, however, provide a more accurate representation of the

    plant's energy flows

    The first step in ncorporating the power plants was to add them

    as

    new processes, and to

    complete'their input

    /

    output profile. The power plants

    are

    at the coke ovens, the steel works,

    and the

    rollingmill.

    Second, the emission coefficients were calculated. This

    was

    straighgorward,

    and coefficients were taken fkom the LCA Model and converted to pounds per MJ of steam

    d. Steam is tracked in UT hroughout the model.

    Third, power plant fuel inp& and by-product electricity generatio

    added to the

    input/output profiles

    in

    the proper

    unts.

    The steam output

    fiom

    the three power plants

    is

    clearly

    assigned to consumption points within the plant

    and,

    therefore, each power plant's steam

    output

    is

    designated accordingly. These three items were added to the set

    of

    ommodities and the set of

    intermediate products.

    Since steam was not yet entered as an input into any process, this was the next and final

    step to complete for the power plants. There are only four processes that use steam

    as

    an nput,

    and these processes were amended to include a steam input in the appropriate

    units

    of

    MJ

    team

  • 7/24/2019 ecologia industrial acero

    45/66

    Iron

    Steel - 38

    per shorttonof process level. Adetailed electricity balance, which appears in Appendix A,

    indicates that kilowatts of consumption and production

    are

    vlrithin5 'percent of each other, which

    may reflect transmission losses and slight discrepancies between the LCA data and Bamett's

    technical coefficients.

    The

    main

    solid waste fiom steelmaking is the slag h m he blast iiunaceand fiom the

    steel furnace. Waste slag

    has

    no

    e-use within the plant, and all of the slag exits the steel

    mill.

    Waste slag is commonly sold for use in road building and cement production. In the LCA model,

    waste slag totals790,000

    MT

    per year. Most of the slag is sold for productive use, and

    approximately 5% is landfilled as solid

    waste.

    In the LCA report,Table 3-4, page 30,the slag

    figure

    reported only includes the landfilled quantity, not the slag sold for productive use.

    The LCA slag output coefficients were apparently adjusted so that only the landfilled

    portion

    is represented. In the present model, al l slag coefficients were readjusted upwardusing

    the implied

    waste

    ratio. Slag exiting the plant

    as

    waste is

    accounted

    for as a solid waste and

    landfill

    costs accrue

    to

    the total cost ( t h e

    are no

    other costs associated with landfilled slag),

    while slag sold earns

    a

    credit that accrues

    to

    the total cost.

    Upstream

    Activities

    mining, iron pelletproduction, fluxmining, zinc mining, and electricity.generation. The addition

    outphfthese five processeso

    ~ e . . ~ ~ u - ~ ~ u ~ ~ ~ ~ p ~ e ~ ~ f o - r - ~ h - p , r - ~ s s ~

    as

    straightforward.

    Emissions were converted-

    to pounds

    pei

    short

    ton

    emissions were added to he

    commodities and emissions sets as required. Labor, energy, and material inputs were not

    included because he

    process

    model optimizes plant activities not the entire vertically integrated

    supply chain.

    The challenge

    is

    to

    decide where to draw the line

    as

    we move upstream and m e r

    The next modification includes the addition of

    six

    upstream processes:

    coal

    mining, iron

    outside of the steel mill. We have bcluded ironand coal mining operations, pellet production,

    flux mining, and

    zinc mining

    as hese are key inputs. But should we include the inputs that are

    required

    to

    execute the pellet production process? Fuelsare consumed, but they are

    not part of

    the plant's optimization

    hction,

    i.e. the

    cost

    of

    upstream

    inputs is

    accounted

    for in the price of

    the product purchased. In essence, what processes do

    we

    exclude? The answer to this question

    can be based onavariety of arguments, but sophisticatedarguments were not necessary

    to

    reach

    a

    conclusion. This isdue to the fact that the LCA documentation does not permit further

    upstream analysis or inclusion. Although items such as "2237-Fuel

    Oil

    Produce/Deliver" is

    itemized as an input to several processes, there is no corresponding inputloutpdemission profile

    for

    this

    process, hence we cannot include it directly.

    This

    orces a different approach.

    When we examined our model's performance in replicating LCA totalemission figuresby

    source, we found our model

    was

    quite accurate in all respects, but slightly underestimated

    . emissions for these five upstream processes which rely on nputs like "2237" above, where

    no

    inputloutputlemission profiles are available. Therefore, based on inal delivered quantities in the

    hypothetical case, we adjusted he emission coefficients on these five processes slightly upward

    in order to calibrate the model based on he

    total

    emissions reported in SCSI (1998, p. 29-30).

  • 7/24/2019 ecologia industrial acero

    46/66

    Industrial Ecology

    - 39

    The external electricity gqneration emission profile for the Middle Atlantic region is

    calculated based

    on

    he LCA report. The Rhodes report provided a basis for estimating emissions

    per

    ki lowatt of purchased electricity

    New Technology

    New technology

    is

    represented in the model

    as

    process options. For example, the model

    allows two different types of coke production: conventional by-product coking and non-recovery

    cokemaking.

    Several W er en t kinds of iron echnologies

    are

    available: Corex and

    two

    coal and

    three gas-based

    DRI

    echnologies. To absorb any possible DRI production, the model allows

    switchingfrom basic oxygen

    steel

    k c e s o five different EAF fimaces,

    each

    with different

    DRI-scrap blends. Similarly, the electricity co-generating plants have the option of usingby-

    product gases or purchased

    natural

    gas.

    Some emissions coefficients for the new technologies were estimated due to incomplete

    survey results. Newtechnologies

    are

    relatively early

    in

    their development, and process

    coefficients varied

    among

    datasources and specific and proposed installations. We have

    attempted

    to

    reconcile these variations and the coefficients sometimes represent averages,

    estimates, and interpolations. We believe that we have achieved an accurate accountingof the

    input,

    output,

    and emission profiles of the new technologies. The new technologies include

    electric arc furnaces,non-recovery cokemaking, and direct iron reduction.

    In order to constrainthe addition of new

    .-_

    echnologies

    ___ _j

    to - r j

    easonable levels,

    a

    set of capital

    comtr&&we;e ~~-~h~w~echnologys assigned a lump sum capital cost in aprice

    table. These capital costs

    are

    aken h m Considine et

    ut.

    (1992). The plant

    can

    be constrained

    to

    a

    fixed amount of total capital

    outlay

    (say

    $500

    million over the planning horizon . But these

    costs

    are not incurred as lump sums, instead we assume they are financed by some combinafion

    of debt and equity, and we amortizepayments o these sources based on

    a

    twenty year lifeand a

    average

    cost

    of capital of ten percent in the base case. Costs must be

    amortized

    and included

    in

    this

    manner o support the structure of the optimization

    of

    annual operating

    costs.

    Electric

    Arc Furnaces

    Obviously, electricarc furnaces0re not new but they do representan alternative

    not

    employed in the

    existing

    plant. Moreover, EAFs

    can

    be fed different

    ratios

    of scrapand

    directly reduced irondepending upon the relative

    cost

    these twomaterials. Data by Barnett for

    the Cravdordsville plant operated by Nucor, Inc. provide the labor, energy, and materials

    requirements for the E M options. Emissions are based upon hose presented above.

    Direct

    CO

    and C02 missions are assumed to be negligible. The range of the PMlO coefficient was very

    large, and using an average has significant impacts in a social planning

    context.

    A figureof 0.05

    was used in consideration of BF and BOF coefficients.

    To

    permit evaluation of different input ratios of scrap and DFU, ultiple E M rocesses

    are

    included in the model to represent the different fixed input ratios.

    An

    analysis was made of

    the effect of the input ratio of these two items has onother factors, mainly electricity

    consumption. There

    are

    many important factors that affect the

    amount

    of energy consumption in

    the

    EAF

    as a function

    of

    percent DRI charged. Some of these include:

  • 7/24/2019 ecologia industrial acero

    47/66

    Iron

    &

    Steel

    -

    40

    Physical form of the

    DIU

    (particle size)

    Friability (leads to excess fines and yield decline)

    Temperature of the DRI harge

    Batch versus Continuous Charging

    Composition (unreduced iron oxides, FeO)

    Gaugue

    (mainly silica and

    alumina)

    We assume continuous charging at low-medium charge rates

    of 1040%.

    Batch charging

    generally leadsto lower productivity, while continuous charging generally leads to improved

    productivity (Stephenson,

    1980).

    The inclusion of

    DRI

    n the

    EAF

    charge may affect electricity

    consumption positively or negatively based on he relative influence of the tors itemized

    above,

    it

    is the net effect that is important for this study.

    As

    described in Stephenson

    (1980,

    p. 1 o), "continuous

    ofDRIresults

    n

    several improvements

    in

    the melting processwhich tend to

    e t he detrimentaleffects of

    DRI

    composition. The improvement inpower consumption

    occurs at relatively small additions, in the range of 10 to

    30

    percent

    of

    the charge, where savings

    onand help

    Figure 5. Electricity consumption per percent

    DRI

    harged

  • 7/24/2019 ecologia industrial acero

    48/66

    Industrial

    E C O ~ O ~ Y41

    inpower consumption of

    3

    to log ercent have been reported when the optimum quantity of DRI

    is used. When more

    than

    30percent DRI is continuously fed, the overall effectis toward

    increased power consumption relative to an a l l scrap charge. With DRI constituting

    60-75

    percent of the charge, the electric power consumption experienced by most commercial DRI

    users is

    in

    the range of

    550-650

    kwh

    ton

    apped.

    An

    approximation of this relationship is

    depicted in Figure

    1.

    The electricity consumption coefficients applied in the process model

    are

    based onFigure

    1. It is

    important

    to note

    that in addition to the factors mentioned above regarding sensitivity of

    energy requirements, other items such as electrode, oxygen, and lime consumption may go up or

    down. Also,yields may vary

    as

    well.

    In

    order to keep the model concise, and for lack of

    scientific description of the complex interrelationships of these factors, none

    of

    these other

    fkt ors have beenaccounted for in the

    EAF

    profiles.

    Non-Recovery

    Cbkemaking

    power generation. The input

    data

    are

    updated h m he study by Considine et

    al.

    (1

    992),

    collected h m lant operators. Steam output canbe used downstream in the BF, since he Jewell

    process does not require steam

    as

    an input and we typically consider conventional by-product and

    Jewell

    to

    be mutually exclusive. Emissionsare based

    upon

    Table 1above by adding emissions

    fiom

    charging, coking, pushing, and quenching to calculate a totalemission (averaging where

    this

    emission coefficient. Power generation emissions

    are

    e based upon the previously

    discussed Unit emissions

    for

    the existing power plants. PMlO emissions are estimated fiom the

    by-product cokingemissions coefficients. Finally, our coefficients assume

    that

    a l lCO and C&

    emissions h m ewell

    are in

    the offgas stream, which

    are

    emitted during the power generation

    process.

    -

    ThenonLiveryCokenAcingproces-involves

    coking, flue-gas desulfurization, and

    needed). - p o l Y ~ ~ ~ c ~ Y ~ ~ ~ ~ - ~ - ~ ~ ~ e ~ oue.to.high uncertainty of

    Direct Reduction

    The first

    direct

    reduction process includes COREX. The resource requirements

    are

    h m

    Considine

    et

    aZ.

    (1992)

    Unlikeconventional blast furnaces, COREX, requires

    ironore not

    pellets

    and c811accept

    a

    wide range of coal types, including steam coal. Emissions based

    upon

    Table

    2 above except that CO emissions

    are

    estimatedbased upon the blast furnace coefficient and coal

    consumption.

    The Unit

    resource

    equirements fo r the remaining direct reduction technologies are based

    upon Table 3above. For the Circofer process, CO, NO,, and PMlO emissions

    are

    based on

    Corex coefficients, using coal

    as

    a reference. The

    CG,

    lag, Phenol, C1, and

    3

    e m i s s i o 9 - p

    fiom the manufkhuer. Unit methane emissions are an average of the other DRI technologies.

    The C@ emissions coefficient for the Circored process is from estimates by the technology

    vendor. NOx and PMlO coefficients

    are

    based upon two other natural gas based processes,

    Mdrex and HYLIIIprocesses for which data are available. Emissions of COY O,, NO,, PMlO,

    and CH4 by the coal-based Fastmet process are estimated based upon Circofers unit

    coal

    consumption. C Q missions fiom the HYLIII and M drex process are estimated based upon the

    unit gas consumption of the Circored process. The process emissions coefficients for PM10, SOX,

  • 7/24/2019 ecologia industrial acero

    49/66

    Iron& Steel - 42

    NO,, and

    CO

    are very close to thsose from a DRI plant permit application in Convent Louisiana in

    1997. VOC and dust figures were taken fiom the permit and added to the emission profile for a l l

    five DRI processes.

    Reduction gases are produced on site in a gas reformer. Reducing gases are generated

    from a mixture of

    naturiil

    gas

    and recycled top gas fiom the reduction

    furnace.

    The t h d

    efficiency of the reformer is enhanced with a heat recovery system. Sensible heat is recovered to

    preheat the combustion air used in the reformer burners and the mixture of top gas and naturiil

    gas fed to the refomer.

    The majority of the top gas fiom the reduction furnace is recycled either fo r reinjection to

    the f urnace after cleaning) or for heat recovery. In addition, top-gases are consumed within the

    plant

    in

    avariety of other utilities services such

    as

    steam generation, electricity generation, and

    otherpower and preheating activities. Inthis analysis, we ssuIl[ie that a l l such utilities

    are

    required components of the DRI module

    as

    a whole. That is, these off-gases cannot

    be

    exported

    to

    other

    uses

    within an integrated production process.

    What we are

    concerned with

    are

    those

    excess gases, however

    small,

    that

    are

    available for

    export

    and their respective heat values.

    But

    due to the variety of arrangements of reformer,

    furnace,

    and

    utilities, one

    cannot h o w he

    amount of off-gas available for export without a detailed schematic

    of

    a plant.

    Estimates

    of

    export gas heat d u e s are equally elusive due to changes in heat value, pressure, and temperature

    as op-gases

    pass

    hrough

    a

    maze of reuses.

    Multiple technical s o ~ $ W e r e - ~ ~ l ~ ~ ~ - y i e l ~ g ~ - f e w ,ide ranging estimates for

    export gas volume and heat values. For gas-based processes, estimates ranged fiom 1.188

    mmBTUIton metal to 5.840

    mmBTU/ton

    metal, depending on he particular plant schematic. For

    gas-based

    DRI processes,

    this

    study

    assumes

    2.5 &TU

    off-gas ton

    Fe

    nd 0.5

    mmBTU /ton Fe

    or

    coal-BasedDRI

    processes.

    The gas-based figure compares with a figure of 4mmBTU/tonmetal

    exported fiom the Blast Furnace.

    In

    the present model, the Blast Furnace sim ilarly recycles

    much of its top

    gas,

    and

    is

    actually

    a

    net electricity produce