Oxy-Fuel Firing in Tilting Rotary Furnaces: What the Evidence Supports | 倾动式回转炉富氧燃烧：证据支持什么

Cold Dross & Residue · 核查修订版 · Sources retrieved 2026-10-30

1. Where the 70% Comes From

The figure is reproducible — which is precisely why it should be treated with care. With methane as fuel, air taken as 21% O₂ / 79% N₂ by volume, and an excess-air factor λ = 1.05, stoichiometric calculation gives:

O2 in oxidiser	Wet flue gas (Nm³ per Nm³ CH₄)	Relative to air	Reduction
21% (air)	10.62	100% — baseline	—
30%	7.77	73.1%	−26.9%
40%	6.10	57.4%	−42.6%
60%	4.43	41.7%	−58.3%
80%	3.60	33.9%	−66.1%
90%	3.32	31.3%	−68.7%
95% (industrial O₂)	3.18	29.9%	−70.1%
100%	3.10	29.2%	−70.8%

From 21% to 100% oxygen is almost exactly a 70% reduction. The coincidence is too clean. This is a stoichiometric ceiling that has been repeated as an operating result.

Three conditions keep a real furnace away from it. Air infiltration — a tilting rotary furnace is not a sealed vessel, and door leakage is routine. Volatile organic compounds from oily and coated scrap burn inside the furnace and add to the gas volume independently of the burner. Excess oxygen control is often deliberately run rich or lean to manage oxidation, moving λ away from 1.05. A field study of Stena Aluminium’s furnace found that unsealed doors let in an unknown quantity of air, placing combustion in a mixed air/oxy-fuel regime and generating appreciable NOx as a result.

Industrial figures, where they can be traced, spread across a wide band:

Source	Furnace	Reported reduction	Nature
Air Products case B	Reverberatory, natural gas	−84% (307,166 → 49,866 SCF/hr)	Vendor model
Linde, 200+ installations	Aluminium melting, various	up to −90% gas and emissions	Vendor disclosure
Linde ALUMELT LTOF	Mainly reverberatory	−80% to −85% per tonne Al	Vendor disclosure
ALTEK, 15 t TRF, USA	TRF, gas/oxygen hybrid	539 m³ gas + 1,078 m³ O₂ per charge (31.25 m³ gas/t)	Case data, not a percentage

A methodological caution: the percentage also depends on the basis. On a dry basis the air-fired figure is about 8.62 Nm³ and the oxy-fuel figure about 1.10 Nm³ — an 87% reduction. Per unit of energy input, oxy-fuel flue gas is about 28% of the air-fired figure, a 72% reduction. Wet or dry, per fuel or per energy: the same furnace can support figures twenty points apart.

2. The 40% Claim, and Why It Fails

No primary source was located for “thermal efficiency increases by 40%”. Two things probably produced it.

First, a conflation of metrics. The closest traceable statement is Linde’s, on low-temperature oxy-fuel: heat transfer to the metal increases by about 40%, and melting rate increases by about 40%. Melting rate is throughput. It is not thermal efficiency. A secondary Chinese-language page carries “melting efficiency up to 50%” and “fuel savings up to 50%” — the same family of numbers with the nouns rearranged, which is how this error propagates.

Second, and more decisively: the absolute values rule it out.

Source	Furnace	Thermal efficiency	Basis
Bartz Maschinenbau	TRF	≈70%	Product literature
ALTEK TTRF	TRF, oxy-fuel or cold-air	75–80%	Product literature
Air Products Table II	Reverberatory, air-fired	42.7% available energy	Model, 2,000 °F flue
Air Products Table II	Reverberatory, oxy-fuel	75.2% available energy	Model, 2,000 °F flue

Aluminium melting furnaces already operate between roughly 43% and 80%. If “40%” were meant as an absolute efficiency, adopting oxy-fuel would represent a decline, not an improvement — which is physically incoherent. As a relative gain, the reverberatory model gives 42.7% → 75.2%, an increase of 32.5 percentage points or 76% relative: not 40%, and from a model, not a rotary furnace.

3. What Actually Produces the Gain

The mechanism is routinely described as “no longer heating nitrogen”. That is only half of it, and the less important half.

In the Air Products comparison, oxy-fuel cut flue gas volume by 84% but flue gas heat loss by only 72% — because CO₂ and H₂O have substantially higher specific heat capacities than N₂, so each cubic metre of oxy-fuel flue gas carries more enthalpy. The larger effect is on the input side: total energy required fell from 26.4 to 17.2 MMBTU/hr, a 35% reduction, while energy delivered to the metal stayed roughly constant at 6.8 MMBTU/hr.

Heat loss through flue gas	Air-fired	Oxy-fuel	Source
Holding/preheat furnace, field	34–62%	≈14%	DiVA thesis, Sapa plant data
Reverberatory, model, 2,000 °F	57.2%	24.4%	Air Products Table II
CH₄, λ = 1, 1,000 °C flue	50% loss (50% efficiency)	24% loss (76% efficiency)	AFRC appraisal

Fuel savings that can be traced: 30% (Linde AIROX staged enrichment), 30–50% (Linde LTOF), up to 50% across 200+ installations, and a maximum of 70% in a single-furnace example in patent US 7,655,067 — where heat input per kg of aluminium fell from about 8,420 kJ to 2,519 kJ. That last figure is a single experimental furnace, spans melting, holding and dross furnaces, and refers to heat input rather than fuel purchased. It should not be quoted as a plant average.

4. What Oxy-Fuel Does Not Do

Recovery figures in this field are unusable without their denominator:

Figure	Denominator	Source
75.7%	Metal directly recovered ÷ dross charged	MDPI Metals 2021, TBRC trial, 35 kg dross + 25 kg salt
98%	Aluminium recovered ÷ aluminium contained in dross	PyroGenesis DROSRITE, vendor claim
98.9%	Aluminium in charge ÷ aluminium in tapped metal	SINTEF, 500 kg pilot, salt-free, non-oxidising
80–90%, or >95%	Not stated	Various vendor pages — unverifiable as published

5. Why It Failed the First Time

Oxy-fuel is not new to secondary aluminium. An industry retrospective records that roughly twenty years ago European secondary aluminium plants widely converted to oxy-gas burners, achieved the expected gains in throughput and consumption — and then reverted to air-fuel burners, because of equipment, refractory and safety problems.

The reason is straightforward. A methane flame in pure oxygen reaches about 2,754 °C against roughly 1,938 °C in air. In a rotary furnace, that produces localised hot spots against the refractory and against the charge, driving oxidation and metal loss. Removing nitrogen is easy; controlling what the hotter flame then does is not.

The patents that followed are therefore not about using oxygen. They are about controlling it:

Patent	Applicant	What it actually claims
US 4,960,460 c. 1990	Alcan International	Plasma torch in a rotary furnace — salt-free, but the heat source is plasma, not oxy-fuel. Cost: electrode maintenance, water cooling.
US 7,655,067 B2 Priority 2004; granted 2010	Air Liquide	Controls combustion by sensing CO and/or H₂ in the furnace or flue; oxidiser O₂ above 10 vol%, preferably above 21%, more preferably above 88–95%; O₂/fuel volume ratio 1–5, preferably 1.5–3. Fuel flow held broadly constant while oxidiser flow is modulated — the point being to suppress aluminium oxidation, not merely to burn hotter.
US 2013/0000450 A1	ALTEK	Door assembly in which a rotating part turns with the vessel and a stationary part does not, sealing the furnace to control atmosphere and suppress thermiting — enabling recovery without salt.
CA 2,930,950 A1	Applicant unverified	Selective oxy-fuel burner and operating method for a rotary furnace, addressing the streaked or uneven heating caused by flame–vessel interaction in a tilting furnace.

Not located, and therefore not asserted here: oxy-fuel patents under Pyrotek, Befesa, Hertwich/SMS, or a patent number for Linde’s ALUMELT/LTOF. Their technologies are documented in vendor literature; the patent numbers were not found.

6. Three Technologies Called “Oxy-Fuel”

Mode	Oxygen level	Consequence
Oxygen enrichment	Below 95%	Flue gas falls along the curve in section 1 — about −27% at 30% O₂, −58% at 60%. Moderate gain, modest modification.
Oxy-fuel	≈95–100%	Flue gas down about 70% theoretically. Full flame-temperature and refractory consequences.
Low-temperature / flameless LTOF	Enriched	Peak flame temperature deliberately held below thermal-NOx formation temperature. The third-generation response to the failure described in section 5.

7. Salt-Flux Practice: The Parameters Oxy-Fuel Sits On

Oxy-fuel changes how heat arrives. It does not change what the salt has to do: cover the metal against oxidation, lower interfacial tension so droplets coalesce, and dissolve Al₂O₃.

Parameter	Value	Source
Salt composition	70 wt% NaCl + 30 wt% KCl, with CaF₂; eutectic 667 °C	German standard practice, TMS 2002
Alternative	Equimolar NaCl–KCl, eutectic 667 °C	KTH thesis
Operating temperature	≈750 °C setpoint; up to 800 °C in trials	MDPI Metals 2021; KTH
Salt factor — rotary drum	0.9–1.8	TMS 2002
Salt factor — tilting rotary	0.4–0.8	TMS 2002
Consequence	Over 1 tonne of salt cake per tonne of dross treated; landfill prohibited in the EU	Industry review

The salt factor distinction matters more than it appears: the same phrase “rotary furnace” covers two machines with salt consumption differing by a factor of two to four. Any comparison of routes must state which.

8. Salt-Free Routes: Oxy-Fuel Has One Confirmed Instance

Process	Heat source	Recovery	Note
ALUREC	Oxy-fuel burner	Not disclosed as a rate in the source consulted	Confirmed. Rotary tilting converter with oxy-fuel burner and exhaust port on the same side. Nitrogen-free, small-volume flue gas at elevated temperature; burner brings the refractory wall to about 1,000 °C in a short time. Lower flue volume and higher flame temperature together make the process more energy-efficient.
DROSRITE	Plasma / electric arc	Up to 98% of contained aluminium	Vendor claim. Seven systems in Saudi Arabia, up to 45,000 t/year of dross.
SINTEF pilot	Graphite plasma torch	98.9% of charged aluminium	500 kg scale, salt-free, non-oxidising; no refractory wear reported.
ALTEK TTRF	Oxy-fuel or cold air	Not disclosed	Salted or dry-flux operation; thermal efficiency 75–80%; cycle up to 3 h.
Hertwich URTF	Oxy-fuel	Not disclosed	3–20 m³ (6–40 t), 1.5–8 t/h, with oxygen lance and online CO measurement; configured for low-salt, salt-free ALUREC, or calcium aluminate.

A correction to a claim sometimes repeated elsewhere. The assertion that a salt-free rotary furnace at Voerde in Germany used oxy-fuel burners rather than plasma torches could not be confirmed. What is documented: Corus Aluminium Voerde was one of three parties that jointly developed the URTF; the URTF uses about one third of the salt charge of a conventional rotary salt furnace — that is low salt, not salt-free; and the two URTFs at Stena run AGA oxygen/propane burners. No primary source was found connecting Voerde specifically with a salt-free furnace, or comparing burner technologies at that site. ALUREC is the instance that can be cited.

9. Cycle Time and Energy: Two Cases, and Why They Cannot Be Compared

Case	ALTEK, Middle East (2010)	ALTEK, Kentucky USA (2013)	SA-Foundry catalogue
Configuration	25 t, 40 m³, 5.4 MW	15 t	Tilting rotary, 3 / 5 t Al
Firing	Single burner, air / natural gas	Natural gas / oxygen hybrid	Gas or liquid fuel
Pure melt time	3.5 h	—	≈2 h
Cycle, tap to tap	4.7 h	3.0 h	≈3 h
Throughput	7.3 t/h (>30,000 t dross/yr)	>100 t per 24 h (≈4.2 t/h)	—
Energy	75 kWh/t	31.25 m³ gas + 62.5 m³ O₂ per tonne	400–550 kWh/t

The 3.0-hour oxy-fuel hybrid cycle looks faster than the 4.7-hour air-fired cycle until throughput is checked: the air-fired 25-tonne furnace produced more metal per hour. Cycle time is set by capacity, charge composition and work organisation. Firing mode is one variable among several, and these two cases cannot isolate it.

The energy figures diverge for a related reason. The 75 kWh/t case processes primary dross, high in contained metal. The 400–550 kWh/t catalogue figure covers mixed charge bearing oxides and oil. Neither is a general TRF constant. The 15-tonne case, at roughly 312 kWh per tonne from gas, sits inside the catalogue band and is the more transferable of the two.

10. NOx: Theory Says Down, Practice Is Conditional

Removing nitrogen from the oxidiser removes the feedstock for thermal NOx — in principle. In practice the outcome depends on two things that vary furnace by furnace.

• Flame temperature. Methane in pure oxygen reaches roughly 2,754 °C against 1,938 °C in air. Local hot spots can cross the threshold at which thermal NOx forms rapidly, above about 1,300 °C.
• Sealing. Where doors leak, nitrogen re-enters and the theoretical advantage disappears. This is what the Stena field study observed.

Linde’s LTOF addresses this directly by holding peak flame temperature below the formation threshold. No absolute NOx concentration in mg/Nm³ measured on an aluminium rotary furnace was located. That figure is not published here.

11. What This Page States and Does Not

Claim	Status	Basis
Flue gas −70%	Theoretical	Reproducible at 70.8% for 21% → 100% O₂ at λ = 1.05. Industrial range 50–90%. Not an operating figure without stated conditions.
Thermal efficiency +40%	Withdrawn	No primary source. As an absolute value, contradicted by measured 70–80%.
Fuel saving 30–50%	Traced	Linde vendor disclosure; 70% is a single-furnace patent example, not an average.
Melting rate +40%	Traced	Linde LTOF. Throughput, not efficiency.
Recovery improved by oxy-fuel	Not supported	Oxidation loss 0.5% at both 50% and 90% O₂; no difference.
Cycle time shortened by oxy-fuel	Not isolated	The air-fired 25 t furnace had higher hourly throughput than the 15 t hybrid.
ALUREC is oxy-fuel salt-free	Confirmed	Primary literature; Hertwich URTF configured for salt-free ALUREC.
Voerde salt-free oxy-fuel	Unverified	Not asserted. URTF at Voerde documented as one-third salt, i.e. low salt.
NOx concentration, mg/Nm³	Unverified	Mechanistic analysis only; no measured value for an aluminium rotary furnace.
Investment payback, break-even oxygen price	Unverified	Depends on gas price, oxygen price and existing heat recovery. No primary data located.

Primary Sources

• Air Liquide, US 7,655,067 B2 — Method for processing aluminium in a rotary or a reverberating furnace (priority 2004-02-25; granted 2010-02-02)
• Alcan International, US 4,960,460 — Plasma arc dross treatment (c. 1990)
• ALTEK, US 2013/0000450 A1 — Tilting rotary furnace system and methods of aluminium recovery
• CA 2,930,950 A1 — Selective oxy-fuel burner and method for a rotary furnace
• Air Products, Oxy-Fuel Technologies and Strategies for Secondary Aluminum Melting Operations — Table II, four-case comparison
• Linde, ALUMELT low-temperature oxy-fuel (LTOF) technical literature; fueling-growth aluminium brochure citing 200+ installations
• TMS Light Metals 2017 — aluminium oxidation loss measured at 50% and 90% O₂
• TMS 2002 — salt factor by furnace type; German salt practice
• MDPI Metals 2021 — TBRC trial, 35 kg dross + 25 kg salt, recovery by denominator
• Chalmers University thesis — Stena Aluminium URTF, AGA oxy/propane burner, air infiltration and NOx
• SINTEF — 500 kg salt-free plasma pilot, 98.9% recovery
• AFRC, Oxygen Technologies Appraisal for Copper Scrap Smelters — stoichiometric flue gas basis and efficiency comparison
• ALTEK case studies — 25 t Middle East (2010) and 15 t Kentucky (2013) installations
• Hertwich / SMS group — URTF product documentation
• SA-Foundry — tilting rotary furnace catalogue, 400–550 kWh/t
• Bartz Maschinenbau — TRF product literature, thermal efficiency ≈70%
• Nokta Metal — industry retrospective on the reversion from oxy-gas to air-fuel burners

Calculation basis for section 1: CH₄ + 2O₂ + 7.52N₂ → CO₂ + 2H₂O + 7.52N₂ (air) and CH₄ + 2O₂ → CO₂ + 2H₂O (oxygen); air at 21% O₂ / 79% N₂ by volume, argon excluded; λ = 1.05; complete combustion; wet basis at 0 °C and 1 atm. Real natural gas containing ethane and higher hydrocarbons gives slightly higher flue volumes.

---

中文摘要 · 倾动式回转炉富氧燃烧

富氧燃烧常被引述两个数字：烟气量减少 70%、热效率提升 40%。前者作为理论极限成立但作为工业数值不成立；后者无法追溯到任何一手来源，且作为绝对值与实测相矛盾。

1. “70%”从何而来。以 CH₄ 为燃料、空气按 21% O₂/79% N₂、过量系数 λ=1.05 的化学计量计算：21% O₂ 时单位燃料湿烟气量 10.62 Nm³，30% 为 7.77（−26.9%）、40% 为 6.10（−42.6%）、60% 为 4.43（−58.3%）、80% 为 3.60（−66.1%）、95% 为 3.18（−70.1%）、100% 为 3.10（−70.8%）。从 21% 到 100% 恰好约 70%，这个吻合”过于完美”——它是化学计量上限，被当作运行结果反复引用。真实炉况有三个偏离源：① 空气泄漏——倾动回转炉非密闭容器，炉门漏气是常态；② 含油涂料料产生的 VOC 在炉内二次燃烧，与烧嘴无关地增加烟气量；③ 过量氧控制常刻意偏离 1.05。Stena Aluminium 的现场研究即发现炉门不密封导致未知量空气进入，燃烧实际处于空气/富氧混合状态并因此生成可观 NOx。可核实的工业数值分布很宽：Air Products 反射炉案例 B 为 −84%（307,166 → 49,866 SCF/hr，厂商模型）；Linde 200+ 装置汇总为烟气与排放最高 −90%；Linde LTOF 为每吨铝 −80%~−85%；ALTEK 15 吨案例给的是 539 m³ 天然气 + 1,078 m³ 氧气/炉（31.25 m³ 天然气/吨），而非百分比。基准差异亦须注意：按干基计算空气助燃约 8.62 Nm³、纯氧约 1.10 Nm³，降幅 87%；按单位输入热量计算纯氧烟气约为空气的 28%（降 72%）。湿基/干基、每燃料/每能量——同一座炉可以支撑相差二十个百分点的数字。

2. “40%”为何不成立。未找到任何一手来源。两个可能的成因：① 指标混淆——最接近的可追溯表述是 Linde 低焰富氧”传向金属的热量增加约 40%、熔炼速率提高约 40%”，熔炼速率是产能而非热效率；某中文行业页面写作”熔化效率最高 50%、节省燃料最高 50%”，是同一族数字换了名词，这正是误传的扩散方式。② 绝对值直接证伪——Bartz 载 TRF 热效率约 70%，ALTEK TTRF 载 75%–80%，Air Products 反射炉模型空气助燃 42.7%、纯氧助燃 75.2%（可用能量占比，烟气 2,000°F）。铝熔炼炉本就运行在约 43%–80%，若”40%”指绝对效率，则采用富氧意味着效率下降而非提升，物理上说不通；若指相对提升，反射炉模型给出绝对提高 32.5 个百分点、相对 76%，既非 40% 也非回转炉实测。该论断予以撤回。可陈述的是：TRF 热效率据厂商载约 70%（Bartz）至 75%–80%（ALTEK）；富氧的作用应用能量平衡表述，而非单一百分比。

3. 增益究竟来自哪里。通行说法是”不再加热氮气”，但这只是次要的一半。Air Products 对比中，纯氧助燃烟气量降 84%，而烟气热损失仅降 72%——因为 CO₂ 与 H₂O 比热容显著高于 N₂，单位体积携热更多。更主要的作用在输入端：总需能量从 26.4 降至 17.2 MMBTU/hr，降 35%，而传给金属的能量基本不变（6.8 MMBTU/hr）。烟气热损失对比：DiVA 论文 Sapa 现场数据空气助燃 34%–62%、氧燃约 14%；Air Products 反射炉模型 57.2% vs 24.4%；AFRC 在 CH₄、λ=1、烟气 1,000°C 下给 50% vs 24%（即效率 50% vs 76%）。可追溯的燃料节省：30%（Linde AIROX 分段富氧）、30%–50%（LTOF）、200+ 装置汇总最高 50%，以及专利 US 7,655,067 单炉实施例最高 70%（每千克铝热输入从约 8,420 kJ 降至 2,519 kJ）——后者是单炉实验值、混合了熔炼炉/保温炉/铝渣炉三种炉型、且指输入热能而非采购燃料，不可作为工厂平均值引用。

4. 富氧不做什么。它不提高金属回收率。TMS《Light Metals》2017 对三次回转炉熔炼实测（空气 / 50% O₂ / 90% O₂），铝氧化总损失在两种氧气浓度下均为 0.5%（以出铝量为分母），二者无差异。盐浴回转炉的回收率由盐、温度与界面张力决定，与氧浓度无关；该研究还发现瞬时氧化速率在炉次末期温度峰值时最高，指向温度控制而非燃烧气氛才是杠杆。回收率数字无分母则不可使用：MDPI《Metals》2021 TBRC 试验 75.7%（直接回收金属÷投入渣量，35 kg 渣 + 25 kg 盐）；PyroGenesis DROSRITE 98%（回收铝÷渣中含铝）；SINTEF 500 kg 中试 98.9%（投入料中铝÷出铝，无盐非氧化气氛）；各类厂商页面的 80%–90% 或 >95% 均未说明分母，无法采用。

5. 为什么第一次失败了。富氧对二次铝并非新事物。行业回顾记载：约二十年前欧洲二次铝厂普遍改用氧-气烧嘴，取得预期的产能与消耗改善，随后因”设备、耐火材料和安全性问题”全部回退到空气-燃料烧嘴。原因很直接——CH₄ 在纯氧中火焰温度约 2,754°C，空气中约 1,938°C；回转炉内由此产生局部高温热点，作用于耐火材料与炉料，加剧氧化与金属损失。去掉氮气容易，控制变热的火焰做什么不容易。后续专利因此不是关于”使用”氧，而是关于”控制”氧：Alcan US 4,960,460（约 1990）用等离子炬在回转炉中无盐处理，热源是等离子而非富氧，代价是电极维护与水冷；Air Liquide US 7,655,067 B2（优先权 2004，授权 2010）通过检测炉内或烟气中 CO 和/或 H₂ 控制燃烧，氧化剂 O₂ >10 vol%（优选 >21%，更优选 >88%–95%），O₂/燃料体积比 1–5（优选 1.5–3），燃料流量基本恒定而氧化剂流量受调——其目的在于抑制铝氧化而非单纯烧得更热；ALTEK US 2013/0000450 A1 以炉门组件（旋转部分随炉体转、静止部分固定）密封炉内气氛以抑制铝热反应，从而无需加盐回收铝；CA 2,930,950 A1 针对回转炉的选择性氧燃烧嘴，解决倾斜炉火焰-炉体交互导致的条纹状不均匀加热。未查到故不作断言：Pyrotek、Befesa、Hertwich/SMS 名下的氧燃专利，以及 Linde ALUMELT/LTOF 的专利号——其技术载于厂商文献，但专利号本次未定位到。

6. 被统称为”富氧”的三种技术。富氧（O₂ <95%）：烟气量沿第 1 节曲线下降，30% O₂ 时约 −27%、60% 时约 −58%，改造幅度小；纯氧（O₂ ≈95%–100%）：理论降约 70%，火焰温度与耐材后果全部显现；低焰/无焰（LTOF）：刻意将火焰峰值温度控制在热力型 NOx 形成温度以下，是对第 5 节所述失败的第三代回应。

7. 富氧所依附的盐浴工艺参数。富氧改变热的送达方式，不改变盐必须做的事——覆盖金属抑制氧化、降低界面张力使金属滴聚并、溶解 Al₂O₃。盐配比：70 wt% NaCl + 30 wt% KCl 加 CaF₂（德国标准实践，TMS 2002），共晶 667°C；替代方案为等摩尔 NaCl-KCl（共晶 667°C，KTH）。操作温度：设定点约 750°C，试验最高约 800°C。盐/料比必须区分炉型：回转鼓式炉 0.9–1.8，倾动式回转炉 0.4–0.8（TMS 2002）。后果：每处理 1 吨渣产生逾 1 吨盐饼，欧盟已禁填埋。同一句”回转炉”覆盖了盐耗相差二至四倍的两种机器，比较工艺路线必须说明是哪一种。

8. 无盐路线：富氧有一个已证实实例。ALUREC（已确认）——回转倾动式转炉，配氧燃烧嘴且排烟口同侧；氧燃烟气不含氮、体积很小，减少的烟气量与升高的火焰温度共同使工艺更节能，烧嘴可在短时间内将转炉耐火壁加热至约 1,000°C。DROSRITE（等离子/电弧，厂商称回收渣中含铝最高 98%，沙特已部署 7 套、年处理可达 4.5 万吨渣）；SINTEF 500 kg 中试（石墨等离子炬，无盐非氧化，投入料中铝回收 98.9%，未报告耐材磨损）；ALTEK TTRF（氧燃或冷风，可加盐或干熔剂，热效率 75%–80%，周期 ≤3 h）；Hertwich URTF（氧燃，3–20 m³ 即 6–40 吨，1.5–8 t/h，配 O₂ 枪与在线 CO 测量，可选低盐/无盐 ALUREC/铝酸钙三工艺）。一处需纠正的流传说法：“德国 Voerde 的无盐回转炉用富氧烧嘴而非等离子炬”未能核实。可查证的事实是：Corus Aluminium Voerde 是 URTF 的三方联合开发方之一；URTF 仅使用传统回转盐炉三分之一的盐量，即低盐而非无盐；Stena 的两台 URTF 使用 AGA 氧/丙烷烧嘴。未找到把 Voerde 与”无盐炉”直接对应的一手文献，也未找到该站点烧嘴技术路线比较的资料。应引用的是 ALUREC。

9. 周期与能耗：两个案例为何不可比。ALTEK 中东 2010（25 吨、40 m³、5.4 MW、单烧嘴空气/天然气）：纯熔炼 3.5 h、周期 4.7 h、7.3 t/h、年处理逾 3 万吨渣、75 kWh/t。ALTEK 美国肯塔基 2013（15 吨、天然气/氧-气混合）：周期 3.0 h、31.25 m³ 天然气 + 62.5 m³ 氧气/吨、24 小时逾 100 吨（约 4.2 t/h）。SA-Foundry 目录（倾动回转炉 3/5 吨铝）：纯熔炼约 2 h、周期约 3 h、400–550 kWh/t。3.0 小时的富氧混合周期看似快于 4.7 小时的空气周期，但一算产能就反过来了——25 吨空气炉每小时产金属更多。周期由容量、料型与作业组织共同决定，燃烧方式只是变量之一，这两个案例无法把它单独分离出来。能耗数字分歧同理：75 kWh/t 那例处理的是一次渣（含金属量高），400–550 kWh/t 覆盖含氧化物与油污的杂料，两者都不是 TRF 的通用常数；15 吨案例按天然气折算约 312 kWh/t，落在目录区间内，是两者中更具可迁移性的一个。

10. NOx：理论下降，实践视条件而定。去掉助燃剂中的氮，原则上消除了热力型 NOx 的氮来源。但实际结果取决于两项因炉而异的因素：① 火焰温度——CH₄ 在纯氧中约 2,754°C、空气中约 1,938°C，局部高温热点可越过热力型 NOx 急剧生成的门槛（约 1,300°C 以上）；② 密封性——炉门漏气则氮重新进入，理论优势消失，这正是 Stena 现场研究观察到的。Linde 的 LTOF 直接针对此点，将火焰峰值温度控制在形成温度以下。未找到铝回转炉氧燃 NOx 的绝对排放浓度（mg/Nm³）实测数据，本页不予给出。

11. 本页陈述与不陈述。烟气 −70%：理论值（21%→100% O₂、λ=1.05 可复算 70.8%），工业区间 50%–90%，未给适用条件不作运行数值。热效率 +40%：撤回（无一手来源，且作为绝对值与实测 70%–80% 矛盾）。燃料节省 30%–50%：可追溯（Linde 厂商披露；70% 为单炉专利值非平均值）。熔炼速率 +40%：可追溯（Linde LTOF，属产能非效率）。富氧提升回收率：不成立（50% 与 90% O₂ 氧化损失均为 0.5%）。富氧缩短周期：未能分离（25 吨空气炉小时产能反而更高）。ALUREC 为氧燃无盐：已确认。Voerde 无盐氧燃：未证实，不作断言。NOx 浓度 mg/Nm³：未证实（仅机理分析）。投资回收期与盈亏平衡氧价：未证实（取决于天然气单价、氧气单价与既有热回收水平，未找到一手数据）。

计算基准：CH₄ + 2O₂ + 7.52N₂ → CO₂ + 2H₂O + 7.52N₂（空气）与 CH₄ + 2O₂ → CO₂ + 2H₂O（纯氧）；空气按 21% O₂ / 79% N₂ 体积比，不计氩；λ=1.05；完全燃烧；湿基，0°C、1 atm。含乙烷及更重烃的实际天然气烟气量略高。