【Kiln Furniture Cost Per Ton: Why Pricing Per Piece Misleads Procurement】 窑具采购里最容易被反复推敲的是「一件多少钱」，但把三年的采购额和烧结工序成本摊在同一张表上，往往会发现两件事对不上：单价降了 8%，工序成本反而涨了。问题不在谈判技巧，在计价单位。
In kiln furniture procurement, the most scrutinized number is price per piece. Yet when three years of spending and sintering process cost are laid on one sheet, two figures refuse to reconcile: unit price down 8%, process cost up. The problem is not negotiation skill, but the unit of account.
窑具不是用完就扔的辅材，它的真实成本是被消耗掉的烧结容量。一件匣钵的价值，等于它在寿命周期内承载过的制品总量。
Kiln furniture is not a consumable discarded after one use. Its real cost is the sintering capacity it consumes; the value of a saggar equals the total product mass it carries over its service life.
计价单位：从「元/件」到「元/吨烧结制品」
真正可比的单位应该是元/吨烧结制品：单件价格 ÷（循环次数 × 单钵装载量）。
The comparable unit should be cost per tonne of sintered product: unit price ÷ (cycles × load per saggar).
三种窑具的同口径换算
以年产 5000 吨、单钵装载 5 kg 计，一年要装满 100 万个钵次。三种窑具按同一口径换算：堇青石 280 元/件、60 次循环，折 933 元/吨；莫来石 460 元/件、90 次循环，折 1022 元/吨；碳化硅 1150 元/件、180 次循环，折 1278 元/吨。
At 5,000 t per year and 5 kg per saggar, one million saggar-loads are needed annually. Converted on the same basis: cordierite ¥280/pc over 60 cycles equals ¥933/t; mullite ¥460/pc over 90 cycles equals ¥1,022/t; silicon carbide ¥1,150/pc over 180 cycles equals ¥1,278/t.
碳化硅的循环次数是堇青石的 3 倍，价格是 4.1 倍——寿命没跑赢价格，年采购额反而高 37%。算到这里，多数人会把碳化硅划掉，这正是「元/件」口径最大的问题：它只让你看见这一笔。
Silicon carbide outlasts cordierite threefold but costs 4.1 times as much — service life loses to price, and annual spending runs 37% higher. Most buyers cross SiC off at this point, which is exactly the flaw of per-piece pricing: it shows only one line item.
被漏掉的两笔账：装载率与良率
被漏掉的第一笔是装载率。反应烧结碳化硅抗折强度常见 80–120 MPa，堇青石只有 8–15 MPa，强度余量允许壁厚从 10 mm 减到 6 mm，单窑装载量能提 6–8%，压缩的是窑次与单位能耗。
The first omitted item is loading density. Reaction-bonded SiC typically reaches 80–120 MPa flexural strength against 8–15 MPa for cordierite. That margin allows wall thickness to drop from 10 mm to 6 mm and lifts kiln loading by 6–8%, compressing kiln cycles and energy per tonne.
第二笔是良率。堇青石反复热震后表层剥落，碎屑落进正极材料是整批降级的常见原因，这笔损失记在质量科目里，从不出现在窑具的账上。
The second is yield. Cordierite spalls after repeated thermal shock, and chips falling into cathode material are a common cause of batch downgrade. That loss sits in the quality account and never appears on the kiln furniture ledger.
临界点：良率改善 0.3 个百分点
把多花 172 万与省回的 220 万合起来，净收益约 48 万，临界点落在良率改善 0.3 个百分点——低于这个数，换碳化硅是亏的。
Netting ¥1.72 million extra against ¥2.2 million saved leaves roughly ¥480,000 of gain, with the crossover at 0.3 percentage points of yield improvement. Below that, switching to SiC loses money.
莫来石为什么常被低估
莫来石常常被跳过，是因为它两头不靠：不如堇青石便宜，不如碳化硅耐用。但在碳酸锂体系里它落在一个很舒服的位置，循环次数是堇青石的 1.5 倍、价格只有 1.6 倍，抗热震性优于碳化硅，对窑炉周期短、升降温柔和的产线往往是最稳的选择。
Mullite is often skipped because it sits awkwardly in the middle: not as cheap as cordierite, not as durable as SiC. Yet under carbonate chemistry it lands in a comfortable spot — 1.5 times the cycles at 1.6 times the price, with better thermal-shock resistance than SiC. For lines with short cycles and gentle ramps, it is frequently the most stable choice.
这套模型的三个盲区
这套算法也有管不到的地方。截面超过 1800 mm 的悬臂梁、辊棒类构件，实际失效往往不是磨损而是高温断裂，一次断裂的停窑损失足以抹平三年成本差；定制窑具交期长于标准品，安全库存占用的资金成本也不会出现在任何比价表里。所以模型算出的临界点，应当被当作「值得做一次对比试验」的信号，而不是可以直接签字的采购结论。
This model has blind spots too. For cantilever beams and rollers above 1,800 mm, the actual failure mode is high-temperature fracture rather than wear, and one fracture can wipe out three years of cost advantage. Custom furniture also carries longer lead times, and the capital tied up in safety stock never appears on a comparison sheet. The crossover the model produces should therefore be read as a signal to run a comparative trial, not as a procurement conclusion ready for signature.
樊亚东行业观察
樊亚东行业观察：碳化硅窑具这几年被过度神化，不少选型方案直接跳过成本测算。真实拐点不在窑具本身，而在物料腐蚀体系——氢氧化锂与高镍路线下，碱蒸气会与堇青石中的硅铝组分反应生成低熔点玻璃相，匣钵循环次数从 60 次掉到 15–25 次。同一套公式重算，堇青石折 2800 元/吨，碳化硅仍在 1500 元/吨附近，价差方向直接反转，而且不需要任何良率假设来支撑。所以「碳化硅划不划算」没有统一答案，决定它的是物料腐蚀性、单钵装载量和对制品洁净度的敏感程度；用碳酸锂体系积累的旧台账去做新产线决策，是最常见的隐性亏损来源。
Industry observation by Fan Yadong: Silicon carbide kiln furniture has been over-mythologized in recent years, and many selection plans skip the cost model entirely. The real inflection point is not the furniture itself but the corrosiveness of the material being fired. Under lithium hydroxide and high-nickel routes, alkaline vapor reacts with the silica-alumina phases in cordierite to form low-melting glass, cutting saggar cycles from 60 to 15–25. Rerun the same formula: cordierite rises to ¥2,800/t while SiC stays near ¥1,500/t — the gap flips outright, with no yield assumption required. So whether SiC pays back has no universal answer; what decides it is corrosiveness, load per saggar and sensitivity to product cleanliness. Using a ledger built on carbonate chemistry to decide a new line is the most common source of hidden loss.
窑具在烧结成本里通常只占 5%–8%，但它撬动的是窑炉周转、能源消耗与制品良率三块大头。
Kiln furniture usually accounts for only 5–8% of sintering cost, yet it levers three far larger items: kiln turnover, energy consumption and product yield.
实操落地建议
- 询价表增加「元/吨烧结制品」一列，并强制供应商填报循环次数的测试条件，没有测试条件的循环次数等于没有数字；
- 碳酸锂、氢氧化锂、高镍三条线分开建寿命台账，不要合成平均数，合并统计会把最贵那条线的信息抹掉；
- 把制品良率写进窑具供应商考核，窑具的价值不是「用了多少炉」，而是「少报废了多少吨」。
Practical recommendations
- Add a cost-per-tonne column to the RFQ sheet and require suppliers to state the test conditions behind cycle counts — a cycle number without test conditions is no number at all;
- Keep separate service-life ledgers for carbonate, hydroxide and high-nickel lines instead of averaging them; merged statistics erase the information from the most expensive line;
- Put product yield into supplier assessment. The value of kiln furniture is not how many cycles it survived, but how many tonnes of scrap it prevented.
总结
总结，窑具是能改变工序经济性的变量，不是一件只影响辅材单价的商品，计价单位换掉的那一天，很多「更便宜」的选择会自动消失。
In summary, kiln furniture is a variable that can change process economics, not a commodity affecting only a consumable line item. The day the unit of account changes, many cheaper options disappear on their own.
