Skip to main content

Energy mass balanceCostingEquipment sizingE-fuels

Fischer-Tropsch syncrude separation and fractionation

LTFT syncrude split into tail gas, e-naphtha, e-kerosene, e-diesel and wax for the hydrocracker by Anderson-Schulz-Flory cuts, with reaction water, reboiler duty from the DOE refinery bandwidth, Kreutz 2008 cost law and separation cost per t.

  • CurrencyEUR 2023
  • Battery limitfully financed
  • 62 documented assumptions
  • Updated2 Nov 2026

Mass and energy balance

What the asset draws and delivers for the case and load you choose.

Reference case

Share of the year at full feed. 90 % for a synfuel plant (Kreutz et al. 2008); lower if the FT loop follows variable power.

Flows

  • In: Reboiler heat (steam)—
  • In: Electricity for pumps and air coolers—

Asset

Fischer-Tropsch syncrude separation and fractionation

  • Out: E-kerosene cut C9-C16 (straight-run, before hydrotreating)—
  • Out: E-naphtha cut C5-C8—
  • Out: E-diesel cut C17-C20 (before hydrotreating)—
  • Out: FT wax C21+ to the hydrocracker—
  • Out: Tail gas C1-C4 to the recycle or fuel loop—
  • Out: FT reaction water (oxygenates in solution) to treatment—

Annual quantities at the current usage.

All hydrocarbons formed in the FT reactor (C1+), reaction water excluded. 20 t/h is roughly a 1,000 MW-electrolyser e-fuel plant. The cost law is sized on the FT flow to HC recovery (syncrude x the FT flow ratio): at the default ratio 3.15 a second train starts above about 61 t/h of syncrude, and below about 2.0 t/h of syncrude (6.35 t/h of FT flow) the cost law is extrapolated.

Kerosene cut
— t/h
Wax to hydrocracker
— t/h
Reboiler duty
— MW
Trains in parallel
— -

CAPEX

Specific cost at this size, what backs it, and how it behaves with scale.

Specific CAPEX at this size

—

Reference size 2.02 t/h

Kreutz 2008 Table 24

Model value: Kreutz 2008 HC recovery plant cost law with BOP and indirect costs, escalated to 2023 EUR. Customized: enter an all-in quote per t/h of syncrude.

Specific CAPEX by syncrude feed (hydrocarbons, water excluded)

Log in to view this sensitivity chart.

  1. Ex-factory
  2. Installed
  3. Fully financed· this figure
CAPEX at the battery limit—

Included

  • Equipment and installation
  • Indirect and owner's costs
  • Contingency

Not included

  • Nothing

Added at project level

  • Nothing: project mode uses the figure as is

Project figures

Values passed to project mode when this asset is added to a project.

Concept sketch, not to scale · drag to rotate

Reboiler heat (steam) in
—
Electricity for pumps and air coolers in
—
E-kerosene cut C9-C16 (straight-run, before hydrotreating) outMain product
—
E-naphtha cut C5-C8 out
—
E-diesel cut C17-C20 (before hydrotreating) out
—
FT wax C21+ to the hydrocracker out
—
Tail gas C1-C4 to the recycle or fuel loop out
—
FT reaction water (oxygenates in solution) to treatment out
—
CAPEX to the battery limit
—
Fixed O&M
—
Variable O&M, excluding feedstocks
0 EUR/year
Asset lifetime
20 years

Log in to add this asset to a project.

Separation cost

A standalone indicator for a quick read. Project mode recomputes it for the whole system.

—

  • Separation cost: capital—
  • Separation cost: O&M—
  • Separation cost: steam—
  • Separation cost: electricity—

Linked to the balance above

Share of the year at full feed. 90 % for a synfuel plant (Kreutz et al. 2008); lower if the FT loop follows variable power.

Anderson-Schulz-Flory growth probability of the FT catalyst. 0.85 gives a kerosene-rich, low-wax syncrude; 0.95 sends over 70 wt% to wax. Collis et al. 2022 use 0.934.

Kreutz et al. 2008 size the HC recovery plant on the whole raw FT reactor product (unconverted syngas, inerts, water and hydrocarbons), not on the syncrude alone. Default 3.15 from their BTL-RC-V stream table (raw synthesis product 32.74 kg/s per 10.4 kg/s of C1+ hydrocarbons), the design closest in scale; their other designs give up to 5.04.

Utilities, offsites and buildings as a share of the process plant. Default 23.2 % from Kreutz et al. 2008 Eqn 4 at their 601 MW biomass scale.

Price of the reboiler heat. Default: 41-bar steam, 17.7 EUR/GJ (Collis et al. 2022, Table 5). Lower if the column reboils on FT reaction heat.

Price of the pump and air-cooler electricity. Default 52.4 EUR/MWh (Collis et al. 2022, Table 5).

Separation cost by capacity factor

Log in to view this sensitivity chart.

Benchmarks

This model's figures against published references and reference cases.

ASF C22+ at alpha 0.934, %

52 Collis 2022 / de Klerk 2011

Reference cases at this size and usage

Log in to compare the reference cases.

All results

Every model output at the current settings, grouped, with the sensitivity charts.

62 outputs · 7 groups · 20 also shown above, dimmed

Separation cost

7
Separation cost
ComputingEUR/t
Separation cost: capital
ComputingEUR/t
Separation cost: O&M
ComputingEUR/t
Separation cost: steam
ComputingEUR/t
Separation cost: electricity
ComputingEUR/t
Separation cost per t of liquid products
ComputingEUR/t
Yearly separation cost
ComputingEUR/year

Product cuts

16
Kerosene cut (C9-C16)
Computing%
Kerosene cut
Computingt/h
Naphtha cut (C5-C8)
Computing%
Naphtha cut
Computingt/h
Diesel cut (C17-C20)
Computing%
Diesel cut
Computingt/h
Wax to hydrocracker (C21+)
Computing%
Wax to hydrocracker
Computingt/h
Tail gas (C1-C4)
Computing%
Tail gas
Computingt/h
Stabilised C5+ liquid
Computing%
Distilled liquid
Computingt/h
Kerosene share of the liquid
Computing%
Mean carbon number
Computing-
ASF C22+ at alpha 0.934
Computing%
Gap to the published C22+ share
Computing%

Reaction water

4
Reaction water per t of hydrocarbons
Computingt/t
Reaction water
Computingt/h
Dissolved oxygenates, upper end
Computingt/h
Paraffin end-group factor
Computing-

Heat and electricity

7
Reboiler duty
ComputingkWh/t
Reboiler duty
ComputingMW
Distillation heat per t of liquid
ComputingkWh/t
Heat above the practical minimum
Computing%
Electricity per t of syncrude
ComputingkWh/t
Electricity demand
ComputingMW
Low-grade heat rejected
ComputingMWh/year

Yearly flows

10
Load hours
Computingh/year
Syncrude treated per year
Computingt/year
Kerosene cut per year
Computingt/year
Naphtha cut per year
Computingt/year
Diesel cut per year
Computingt/year
Wax to hydrocracker per year
Computingt/year
Tail gas per year
Computingt/year
Reaction water per year
Computingt/year
Reboiler heat per year
ComputingMWh/year
Electricity per year
ComputingMWh/year

Size and CAPEX

13
Total plant cost (all-in)
ComputingEUR
Specific CAPEX
ComputingEUR/(t/h)
Process plant
ComputingEUR
Balance of plant
ComputingEUR
Indirect costs
ComputingEUR
FT flow to HC recovery
Computingt/h
Trains in parallel
Computing-
HC recovery flow per train
Computingt/h
Cost reference size in syncrude
Computingt/h
Train size over cost reference
Computing-
Cost law extrapolated
Computingflag
Escalation USD2007 to EUR2023
Computing-
BOP and indirect multiplier
Computing-

Finance

5
Capital charge rate
Computing1/year
Annual capital charge
ComputingEUR/year
O&M per year
ComputingEUR/year
Steam cost per year
ComputingEUR/year
Electricity cost per year
ComputingEUR/year

Sensitivities

5 charts
Separation cost by component vs capacity factor

Log in to view this sensitivity chart.

Separation cost and specific CAPEX vs syncrude feed

Log in to view this sensitivity chart.

Product cuts vs chain growth probability

Log in to view this sensitivity chart.

Reboiler duty and reaction water vs chain growth probability

Log in to view this sensitivity chart.

Separation cost vs steam price

Log in to view this sensitivity chart.

About

Calculator context

Separation section downstream of a low-temperature Fischer-Tropsch (LTFT) reactor: hot and cold separators remove the wax, the condensate and the reaction water, a stabiliser sends the C1-C4 to the tail-gas loop, and an atmospheric column splits the stabilised liquid into e-naphtha (C5-C8), e-kerosene (C9-C16), e-diesel (C17-C20) and a C21+ wax cut for the hydrocracker.

Product cuts. The syncrude follows the Anderson-Schulz-Flory distribution on a weight basis, so one input, the chain growth probability alpha (0.85 to 0.95), sets every cut through the closed-form cumulative fraction. The five cuts close the hydrocarbon balance exactly. Counter-test: at the alpha of 0.934 used by Collis et al. 2022, pure ASF gives 56.9 wt% C22+ against the 52 wt% of straight-chain C22+ alkanes they quote from de Klerk; the gap is expected because ASF counts every C22+ species.

Reaction water. One mole of water per mole of CO converted (ENI EP 2 403 807), corrected for the paraffin end groups at the mean carbon number: about 1.26 to 1.27 t per t of hydrocarbons for a cobalt catalyst. Iron catalysts make less through water-gas shift.

Heat and electricity. Transposed from atmospheric crude distillation in the DOE 2006 refining energy bandwidth: current average energy (case 1) or practical minimum (case 2), applied to the C5+ liquid only. This is an envelope, not a column design; no printed source gives the condenser duty, so the model only reports the heat input as low-grade heat rejected.

Cost. Total plant cost from the Kreutz et al. 2008 'HC recovery plant' row (0.724 MUSD2007 at 14 thousand lb/h, exponent 0.70, trains of up to 426 thousand lb/h, the largest Kreutz cost in their Table 26 although Table 24 prints 200, 0.9 for multiple trains), sized like Kreutz on the whole raw FT product sent to HC recovery: 3.15 t per t of syncrude by default, derived from their BTL-RC-V stream table (their other designs give up to 5.04), with balance of plant and 32 % indirect costs, escalated by CEPCI to 2023 and converted at the ECB rate. The separation cost per t of syncrude adds the Kreutz capital charge rate (15.41 %/year) and O&M (4 %/year) to the steam and electricity. With the defaults (20 t/h, alpha 0.90, 90 % capacity factor, current average energy) the model gives about 23.2 EUR/t, a little over half of it steam, and a total plant cost of about 8.2 MEUR. Below 6.35 t/h of FT flow per train, about 2.0 t/h of syncrude at the default ratio, the law is extrapolated downwards and flagged.

Model

132 variables — inputs, calculations and outputs, with their dependencies.

132 variables shown of 132
VariableValueUnitDepends on
20t/h—
90%—
0.9-—
1——
0——
412000EUR/(t/h)—
3.15t/t—
23.2%—
17.7EUR/GJ—
52.4EUR/MWh—
VariableFormulaUnitDepends on
if((<)+(>)+(<)+(>)+(<=)+(>)+(<)+(>)+(<)+(>)+(<=)+(<)+(<)+(>)+(<)+(<)>,,)bool
^(-)*(*(-)+)-
^(-)*(*(-)+)-
^*((+)*(-)+)-
^*((+)*(-)+)-
*/kWh/t
*klb/h
***/EUR/(t/h)
clamp(*(/)^(-)*,,)*^(-)EUR/(t/h)
VariableFormulaUnitDepends on
/(-)-
(-)*%
(-)*%
(-)*%
(-)*%
*%
*%
/max(,)*%
**t/h
**t/h
**t/h
**t/h
**t/h
*t/h
^(-)*(*(-)+)*%
-%
*/(*+)-
*t/t
*t/h
*(+)*t/h
if(==,,*/)kWh/t
(/max(,)-)*%
*kWh/t
*/*kWh/t
**MW
**MW
**h/year
*t/year
*t/year
*t/year
*t/year
*t/year
*t/year
*t/year
*MWh/year
*MWh/year
+MWh/year
*t/h
/max(,)t/h
max(,ceil(/))-
/t/h
/(*)-
if(<,,)flag
//-
(+*)*(+*)-
if(>=,,)EUR/(t/h)
*EUR
/EUR
**EUR
(+)**EUR
*1/year
*EUR/year
**EUR/year
***EUR/year
*EUR/year
/max(,)EUR/t
/max(,)EUR/t
**EUR/t
**EUR/t
+++EUR/t
/max(,)EUR/t
+++EUR/year

Assumptions

62 assumptions used in the calculations

  • Lower limit in the input guards and the flags.

    0
    Numerical constant
  • Unit multiplier in the ASF cumulative fractions, the guards, the train count and the Customized switch.

    1
    Numerical constant
  • Converts a fraction to a percentage and caps percentage inputs in the guard.

    100%
    Numerical constant
  • Converts a percentage (0-100) to a fraction.

    0.01
    Unit conversion
  • Protects the per-tonne divisions against a zero denominator.

    0
    Numerical constant
  • Hours in a non-leap year, the basis of the load hours.

    8760h/year
    Unit conversion
  • Converts kWh/t x t/h (kW) to MW.

    0.001MW/kW
    Unit conversion
  • Converts the electricity price from EUR/MWh to EUR/kWh.

    0.001MWh/kWh
    Unit conversion
  • Converts the yearly heat from MWh to kWh before pricing it per GJ.

    1000kWh/MWh
    Unit conversion
  • 1 kWh = 3.6 MJ = 0.0036 GJ; prices the reboiler heat at the steam price per GJ.

    0.0036GJ/kWh
    Unit conversion
  • Converts the DOE 2006 distillation energy from thousand Btu per barrel to kWh per barrel.

    0.293071kWh/kBtu
    Unit conversion
  • Converts the syncrude feed in t/h to the thousand-lb/h basis of the Kreutz cost row ("Mlb/hr" with M = thousand, as in "MM SCF/hr" for million).

    2.2046klb/t
    Unit conversion
  • Converts the Kreutz base cost from million USD to USD.

    1000000USD/MUSD
    Unit conversion
  • Lower limit of the syncrude feed; a smaller unit is a pilot, not a costed plant.

    0.2t/h
    Model domain
  • Upper limit of the syncrude feed: 3 to 4 HC recovery trains, within the 9 trains Kreutz cost for CTL-RC-V.

    150t/h
    Model domain
  • Lower limit of the ASF growth probability; below it the product is mostly gas and naphtha, not a distillate-and-wax syncrude.

    0.85-
    Model domain
  • Upper limit of the ASF growth probability.

    0.95-
    Model domain
  • Reference case 1: atmospheric distillation at the US refinery current average energy (DOE 2006, CAE).

    1
    Model convention
  • Reference case 2: atmospheric distillation at the DOE 2006 practical minimum energy (PME), a best-practice heat-integrated column.

    2
    Model convention
  • First carbon number of the stabilised liquid and of the naphtha cut.

    5carbon number
    DOE 2006 / method
  • First carbon number of the kerosene cut (naphtha is C5-C8).

    9carbon number
    Collis 2022
  • Last carbon number of the kerosene cut; diesel starts at C17.

    16carbon number
    Collis 2022
  • Last carbon number of the atmospheric diesel cut; C21+ is the wax bottoms sent to the hydrocracker.

    20carbon number
    DOE 2006 Table A1
  • Carbon number of the de Klerk C22+ counter-test.

    22carbon number
    Collis 2022 / de Klerk 2011
  • Alpha of the published counter-test: the model is evaluated at this alpha and compared with the published C22+ share.

    0.934-
    Collis 2022
  • Published C22+ share used as an independent counter-test of the ASF closed form.

    52%
    Collis 2022 / de Klerk 2011
  • Water made per t of CH2 units formed (nCO + 2nH2 -> (CH2)n + nH2O); the model corrects it for the paraffin end groups.

    1.284t/t
    EP2403807 / stoichiometry
  • Molar mass of a chain CH2 unit.

    14.027g/mol
    IUPAC
  • Mass of the two end hydrogens of an n-paraffin CnH2n+2.

    2.016g/mol
    IUPAC
  • Upper end of the dissolved alcohols in the reaction water (informative output for the water treatment).

    5%
    EP2403807
  • Upper end of the dissolved acids in the reaction water.

    1.5%
    EP2403807
  • Fired heat per barrel of feed of an average US atmospheric crude column; read as the reboiler duty of the syncrude column in case 1.

    106.8kBtu/bbl
    DOE 2006 Table A2
  • Practical minimum energy per barrel, used as the reboiler heat of case 2 (best practice, heat-integrated column).

    50.4kBtu/bbl
    DOE 2006
  • Pumps and air coolers of the column, per barrel of feed.

    2.3kBtu/bbl
    DOE 2006 Table A2
  • Converts the per-barrel energy of DOE 2006 to a per-tonne basis.

    0.143t/bbl
    DOE 2006 Table A1 (derived)
  • Base cost of the hydrocarbon recovery plant (separators, stabiliser, fractionator), before BOP and indirect costs.

    0.724MUSD2007
    Kreutz 2008 Table 24
  • Reference train size of the cost row, in thousand lb/h.

    14klb/h
    Kreutz 2008 Table 24
  • Reference train size in t/h of FT flow to HC recovery, the anchor of the 0.70 scaling law; below it the law is a downward extrapolation (flagged).

    6.35t/h
    Kreutz 2008 Table 24
  • Reference size of the 0.70 scaling law expressed in syncrude, the unit of the size slider, at the default ratio. It moves with the ratio slider (6.35 / ratio, 1.26 t/h at 5.05); the output capex_ref_syncrude_t_per_h gives the live value.

    2.02t/h
    Kreutz 2008 Tables 12 and 24 (derived)
  • Scale exponent of one HC recovery train: the specific CAPEX falls as size^-0.30.

    0.7-
    Kreutz 2008 Table 24
  • Converts the syncrude feed to the size basis of the Kreutz HC recovery cost row, which is the total FT reactor effluent sent to HC recovery. Without it the cost law is evaluated on a flow 3 to 5 times too small and understates CAPEX.

    3.15t/t
    Kreutz 2008 Table 12 (derived)
  • Lower guard of the FT flow to HC recovery per t of syncrude.

    1t/t
    Model domain
  • Above 426.2 thousand lb/h (193.3 t/h) of FT flow to HC recovery the plant is split into parallel trains: about 61 t/h of syncrude at the default ratio 3.15.

    426.2klb/h
    Kreutz 2008 Tables 11 and 26 (derived)
  • Cost of n identical trains = cost of one x n^0.9.

    0.9-
    Kreutz 2008 Eqn 3
  • Engineering, contingency and other indirects on top of the installed process plant plus BOP.

    32%
    Kreutz 2008 Table 24
  • Cost index of the Kreutz base year.

    525.4-
    CEPCI
  • Cost index of the 2023 currency year of the calculator.

    797.9-
    CEPCI
  • Converts 2023 USD to 2023 EUR.

    1.0813USD/EUR
    ECB
  • Lower guard of the specific CAPEX, only against out-of-domain values.

    10000EUR/(t/h)
    Model guard
  • Upper guard of the specific CAPEX, only against out-of-domain values.

    5000000EUR/(t/h)
    Model guard
  • Yearly capital charge per EUR of total plant cost, used as the annualisation factor of the standalone separation cost.

    15.41%
    Kreutz 2008 Section 5.1
  • Yearly operation and maintenance, labour included, as a share of TPC.

    4%
    Kreutz 2008 Section 5.3
  • Cut fractions come from the closed-form cumulative W(>=N) = alpha^(N-1) [N(1-alpha) + alpha]. The five cuts telescope and sum to 1, so the hydrocarbon balance closes exactly. Real LTFT syncrude deviates from ASF (methane excess, C2 deficit, olefin re-insertion); the cuts are a screening estimate.

    —
    Flory 1936 / Collis 2022
  • Water per t of hydrocarbons = 1.284 x (14.027 n / (14.027 n + 2.016)), with n = 1/(1-alpha) the number-average carbon number of an n-paraffin product. Exact for a paraffin-only ASF product on a mole basis; an olefin-rich product tends to 1.284. Cobalt (non-shifting) catalyst assumed.

    —
    Stoichiometry
  • Only the stabilised C5+ liquid is distilled, so the reboiler duty per t of syncrude is the per-t-of-liquid figure times the C5+ fraction. The DOE figure is fired fuel at the crude heater; booking it as reboiler heat at the steam price is conservative by the furnace efficiency. No printed source gives the condenser duty: the model only states that the heat input leaves as low-grade heat (energy balance).

    —
    DOE 2006
  • TPC = 0.724 MUSD2007 x (FT flow to HC recovery per train / 6.35 t/h)^0.70 x n^0.9 x CEPCI 797.9/525.4 / 1.0813 USD/EUR x (1 + BOP) x (1 + 32 %), with the FT flow = syncrude x hc_recovery_flow_to_syncrude_ratio (Kreutz size the plant on the raw FT product, not the syncrude). TPC is all-in (installed, BOP, engineering and contingency), so the meta declares battery_limit fully_financed. Owner's costs and interest during construction are carried by the capital charge rate, not by TPC.

    —
    Kreutz 2008
  • Default of the balance-of-plant input.

    23.2%
    Kreutz 2008 Eqn 4 (derived)
  • Default of the capacity factor input; an e-fuel plant fed by variable power with limited H2 storage can run lower.

    90%
    Kreutz 2008 Section 5.2
  • Defaults of the steam and electricity price inputs. Cooling water (0.35 EUR/GJ in the same table) is not costed because no source gives the condenser duty; at the reboiler duty it would add well under 1 EUR/t.

    —
    Collis 2022 Table 5
  • Lifetime exported to project mode (meta.asset.lifetime_years).

    20years
    Kreutz 2008 Table 30
  • The syncrude feed is not declared as an inlet (no chainable commodity); the wax cut is exported as heavy_ends for the hydrocracker, the C1-C4 as offgas, the kerosene cut as jet_fuel before hydrotreating. Steam heat and electricity are inlets priced by the project, so the meta books no variable opex.

    —
    Model convention
  • Below 6.35 t/h of FT flow per train, about 2.0 t/h of syncrude at the default ratio 3.15, the 0.70 law is extrapolated downwards; for skid-mounted modular units it likely understates the cost. The flag below_reference_scale_flag says when.

    —
    Kreutz 2008