The pyroprocessing line
Roughly 60% of the plant's capital sits between the top of the preheater and the end of the cooler. It consumes about 90% of the plant's thermal energy and it sets the ceiling on everything else — you cannot sell more cement than this line makes clinker for. It is also where consultants most reliably get it wrong.
On this page · 16 sections
- 01The one idea
- 02The four machines
- 03Where the fuel actually goes
- 04What must be true
- 05What goes wrong here
- 06Scenarios from the field
- 07The numbers that matter
- 08Build the model
- 09Upgrade paths — what you can actually change
- 10The frontier, and who is running it
- 11What's changing now
- 12Check yourself
- 13Mini case
- 14Go deeper
- 15Carry forward
- 16Where this connects
§ 01 The one idea
Everything in the pyro line follows from a single arrangement: material falls down through the system and gets hotter; gas rises up through the system and gets cooler. They meet at every stage and exchange heat.
That is not a detail of the design — it is the design. A cement kiln has to take raw meal at 80 °C and bring it to 1,450 °C, then bring the product back down to 100 °C. If it did that with a burner alone it would cost three times the fuel. Instead, the hot gas leaving the burning zone pre-heats the incoming meal on its way out, and the hot clinker leaving the kiln pre-heats the incoming combustion air on its way out. The flame only has to supply the difference.
Almost every question you will be asked about this line reduces to one sentence: am I getting the heat where I need it, at the right temperature, with enough oxygen, without disturbing the chemistry?
§ 02 The four machines
Preheater — the heat exchanger
Five (sometimes six) cyclones stacked roughly 20 metres apart. Meal enters at the top and falls through each in turn; kiln gas rises through them counter-current. Each cyclone does two jobs at once: it exchanges heat between gas and meal, and it separates the meal back out of the gas stream so it can drop to the next stage. Meal enters at about 80 °C and leaves the bottom at 850–900 °C. By the preheater exit alone, 30–40% of the limestone has already decomposed.
Calciner — where most of the fuel burns
This surprises people: on a modern plant, 55–65% of total fuel is burned in the calciner, not in the kiln. The calciner is a suspension chamber at 850–900 °C, which is exactly the temperature calcium carbonate needs to break down. It is not trying to reach 1,450 °C, so it tolerates lower-grade fuel, alternative fuel and coarser grinding. Meal leaving it is 90–95% calcined and is called hot meal.
This is the reason the modern kiln is short. Once calcination — the largest single heat demand in the process — is moved out of the kiln, the kiln only has to do the sintering. A precalciner kiln makes three to four times the clinker of a long dry kiln of the same length.
Rotary kiln — where the chemistry happens
A steel tube 50–80 m long and 4–6 m across, sloped 3–4% and turning at 3–5 rpm. Material tumbles down it over 25–40 minutes through three zones: calcination completes, then intermediate minerals form at 1,200–1,300 °C, then the burning zone reaches 1,400–1,500 °C where 20–25% of the charge goes liquid and the calcium silicates form. The rotation rolls that partly-molten material into nodules a few millimetres to a couple of centimetres across.
Grate cooler — the second heat exchanger
The cooler is not a cooler. It is a heat exchanger that happens to cool clinker as a side effect. Ambient air is blown up through the falling clinker bed, dropping the clinker from ~1,400 °C to ~100 °C — which is necessary to freeze the mineral phases in place — while heating that air to 1,000–1,200 °C. That hot air then goes back into the process twice over: as secondary air to the main burner, and through a separate duct as tertiary air to the calciner burner.
Tertiary air is what made the precalciner possible. Without a duct carrying hot air directly from the cooler to the calciner, you cannot burn 60% of your fuel there. When a plant's cooler degrades, secondary and tertiary air temperature falls, and the kiln has to make up the difference by burning more fuel to heat cold air. That is why a cooler problem shows up as a fuel problem.
§ 03 Where the fuel actually goes
A modern Indian dry-process line with a 5-stage preheater and precalciner runs at roughly 720 kcal of heat per kg of clinker. The useful way to read that number is in three groups, because each group has a different management implication.
The three groups sum to 795 against an input of ~720. That is not a rounding error and it is worth understanding: the recoverable streams are partly heat that has already been counted doing work upstream and is being re-used inside the plant, not fresh fuel. Read this bar as the shape of the heat flow, and build the real closed balance from plant data in lesson C5. Any published reference balance you are handed — including this one — is illustrative until you close it against a specific line.
Three things follow, and they are the whole reason a consultant needs this table:
- About 73% of the heat cannot be reduced without changing the product. Calcination alone is 410 kcal/kg and is set by how much limestone the clinker recipe requires. This is why "get to 680 kcal/kg" is a credible target and "get to 500" is not — the floor is physics, somewhere around 550–600 for this configuration.
- The recoverable 200 kcal/kg is not lost. Preheater exhaust dries the raw meal first; whatever survives that feeds waste heat recovery. Cooler exhaust does the same. A consultant who calls the 125 kcal/kg preheater exit "the biggest loss" will be corrected immediately.
- The improvement levers are almost all in the small red block. False air, cooler tuning, refractory condition, hot clinker — unglamorous, cheap, and where the money actually is. Intuition points at the burning zone because it is the hottest and most visible part. Intuition is wrong.
§ 04 What must be true
Everything before this section explains how a pyroprocessing line works. This one says what has to be true for it to work well — the conditions physics imposes regardless of who built the kiln, in what decade, or in which country. A line meeting all eight runs near its practical floor. A line failing one of them cannot be fixed by anything downstream of the failure.
Use it as a checklist. Each row is a thing you can verify on site in a day, and each has a published target.
| What must be true | Why | Target | How you verify it |
|---|---|---|---|
| 1. Every stage exchanges heat counter-currently | A cyclone stage only transfers heat where hot gas and cold meal are in intimate contact moving in opposite directions. Gas that bypasses meal does no work and still costs fan power. | 40–60 °C hot-end approach |
Stage-by-stage gas and meal temperatures. A stage where the gas-to-meal approach has widened is a stage that has stopped working — look for a collapsed splash plate, a worn dip tube or a blocked meal pipe before you look anywhere else. |
| 2. The top cyclone separates | Meal that escapes the top stage leaves with the gas, gets collected downstream and is fed back. That material is heated twice and the circuit pays twice. | >93–95% top-stage efficiency |
Dust loading at preheater exit against kiln feed rate. A rising exit dust loading at constant feed is the tell, and it usually precedes a visible problem by months. |
| 3. Calcination is nearly finished before the kiln inlet | The calciner exists so that decarbonation — the single largest heat demand in the plant — happens in a cheap vessel at 850–900 °C rather than in the kiln's hottest, most refractory-limited zone. | 90–95% degree of calcination |
Loss on ignition of hot meal sampled at the kiln inlet. Below 90% the kiln is doing the calciner's job, and it is doing it in the worst possible place. |
| 4. There is liquid phase in the burning zone | Alite does not form by solid-state reaction at any practical rate. It forms through a melt, and the melt quantity is fixed by the silica and alumina moduli set back in A3. | 20–28% melt at 1,450 °C |
Litre weight (1,250–1,400 g/l), free lime and coating behaviour together. Thin coating with rising free lime means too little melt; ring formation means too much, or the wrong viscosity. |
| 5. Heat in the clinker comes back | Clinker leaves the kiln at ~1,350 °C carrying a large fraction of the heat you paid for. Secondary and tertiary air are the only route back into the process. Heat that leaves in the cooler vent or in hot clinker is simply gone. | 950–1,100 °C secondary air |
Secondary air temperature, cooler specific air consumption and clinker discharge temperature. Cooler grate loading below the CII design figure of 45 TPD/m² — benchmarked Indian plants run 41–47, overloaded ones 50–70. |
| 6. What enters leaves | Alkalis, sulphur and chlorine volatilise in the burning zone and recondense upstream. If the input exceeds what clinker and dust can carry out, the difference accumulates in a loop until something blocks. | Cl <0.015% in raw meal S/A molar 0.8–1.2 |
Hot-meal chloride and the sulphate-to-alkali balance. These are the numbers that decide the ceiling on alternative fuel (A6 §06), and they are measured at the kiln inlet, not in the clinker. |
| 7. Air enters only where you invited it | False air is ambient air heated to process temperature for no return. It dilutes the gas, cools the exchange, and loads every fan downstream of the leak. | <5% per section |
Oxygen rise measured across each section — kiln inlet to preheater exit, mill inlet to mill outlet. The leak is almost always at an inspection door, an expansion joint or a worn kiln seal. |
| 8. The line is stable before it is optimised | A kiln's fuel rate is set by its worst hour, not its average one, because free lime must stay under limit on the bad hours too. Everybody pays for that defence every hour. | LSF SD <1 at kiln feed |
Standard deviations, not means — free lime, kiln feed LSF, kiln inlet O₂, fuel rate. Halving a deviation buys heat rate with no equipment change at all (F6 §03), and it is the cheapest heat rate in the plant. |
Run the eight rows in order and score each one verified good, verified bad, or not measured. The third category is usually the largest, and it is the finding that matters most — a plant that cannot tell you its degree of calcination at the kiln inlet is a plant that has never located its own losses.
The order is not arbitrary. Rows 1 to 3 govern whether the heat exchanger works at all, and nothing downstream compensates for a failure there. Rows 4 to 6 govern whether the chemistry and the recirculating loops are inside their limits. Rows 7 and 8 are the two that cost the most and get measured the least.
These eight have not changed since the first suspension preheater at Beckum in 1953, and they will not change. Every technology in §10 is an attempt to do one of them better — none of them removes a row.
§ 05 What goes wrong here
The pyro line has perhaps a dozen recurring failure modes and four of them account for most of the money. What makes them hard is not that they are subtle — it is that each one presents as a fuel problem, and the fuel gauge is the same gauge for all of them. Diagnosis is a matter of reading several instruments as a set.
| What goes wrong | The tell | What it's worth | Commonly misdiagnosed as |
|---|---|---|---|
| False air ingress | O₂ rising at kiln inlet and preheater exit, no fuel change; ID fan power up; preheater exit temperature up | CII: +0.5 pp cyclone-exit O₂ = 12 kcal/kg at the top stage, up to 60 kcal/kg at stage five | Burner problem, or "the kiln is just old" |
| Cooler deterioration — grate plate wear, poor air distribution, snowman, red river | Secondary and tertiary air temperature falling; clinker exit temperature rising | India actual: a cooler upgrade recovered 19–31 kcal/kg and 1.75 kWh/t | Fuel quality; burner setting |
| Volatile build-up — sulphur, chlorine, alkali cycles | Rising preheater pressure drop; recurring blockages; coating in the riser; rings in the kiln | Blockage stops. Sulphur-to-alkali ratio outside 0.9–1.1 is the leading indicator | Mechanical problem; "bad batch of fuel" |
| Refractory and coating loss | Localised shell hot spot 30–60 °C above normal; kiln drive amperage instability | A 5-day refractory-driven stop at a Gujarat plant cost ₹4.5 crore and 10,000 t | Chemistry problem |
| Cyclone pressure-drop creep — dip tube burn-out, partial blockage | ΔP rising toward the top of the 600–800 mmWG band; fan power up with no output gain | EPA: 0.5–0.6 kWh/t per 50 mmWG; FLSmidth top-stage retrofit 8–10 kcal/kg | Fan problem |
| Chemistry instability from upstream | LSF standard deviation and free lime both moving; R90 drifting | Absorbed as fuel and as refractory life, invisibly | A kiln problem. It is a silo or quarry problem |
False air is the one to lead with, and the reason is the CII figure in that table. A half-point of extra oxygen at the top of the tower costs 12 kcal/kg, but the same half-point at stage five costs up to 60 — because air leaking in low down has to be dragged up through the entire tower and heated at every stage. Where the leak is matters five times more than how big it is. One documented Indian audit replaced a single kiln inlet seal and cut false air from 24,377 kg/h to 6,076 — worth 5.5 kcal/kg and 0.24 kWh/t from one component.
The second thing worth internalising is that refractory life is now a fuel-choice consequence, not a maintenance constant. Burning-zone lining runs about 12 months on coal, drops to 8–10 months on RDF or petcoke, and recovers to 15–24 months with alkali-resistant or MgO-spinel brick. Indian plants spend ₹20–40 per tonne of clinker on refractory, so a 5,000 tpd kiln spends ₹6–7 crore a year — a number that sits in the maintenance budget and is almost never set against the fuel saving that caused it to rise.
§ 06 Scenarios from the field
Everything in the last section sounds like it should produce a big number. It doesn't, and the sector's own data says so.
Read the first tile carefully, because the easy misreading is the flattering one. 707 and 694 are not sector averages — they are the best-achieved 5-stage figures in CII's benchmark. So that 1.8% is not laggards failing to catch up. It is the frontier itself barely moving: India's best 5-stage kiln got 13 kcal/kg better in eleven years, with capital available, under a regulatory efficiency scheme, and with every incentive to try. Step changes of 10–15% are not what happens. What happens is a register of dozens of small items, each worth a fraction of a percent, applied every year without stopping.
Two separate gaps follow, and conflating them is a common error. The frontier moves slowly — 1.8% in eleven years. The distance from average to frontier is much larger: 740 against 694 on thermal, about 46 kcal/kg, and 73 against 56.1 kWh/t on electrical. On a 6,000 tpd line those are roughly ₹17 crore and ₹8 crore a year. Neither is a project. Both are a register — and a consultant who arrives looking for one big lever will find nothing and conclude the plant is already efficient.
CII labels row 14 "Overall SEC Average" for the 88 → 73 kWh/t series. The thermal rows carry no such label — and they cannot be averages, because the same document puts the sector average at 740 while the 5-stage row reads 694. A sector average cannot sit above both of its cohorts. The first draft of this lesson quoted 707 → 694 as a sector average. It is a best-achieved value, and the error survived until a reader checked it. Whenever you quote CII, establish whether the row is an average or a benchmark before you use it.
What a register actually looks like
Indian plants publish these. Every CII energy-award submission is a plant listing every project it did, with investment and saving, project by project. Eight of them:
| Plant | Period | Projects | Investment | Saving | Result |
|---|---|---|---|---|---|
| JK Cement, Muddapur | FY21–23 | 149 | ₹2.21 cr | — | Plant power 65.57 → 63.61 kWh/t cement; 5.03 kWh/t of SEC impact |
| Chettinad Cement, Kallur | FY21–23 | 73 | ₹0.88 cr | ₹11.1 cr/yr | 6.41 kWh/t of SEC impact; average project ₹1.2 lakh |
| Orient Cement, Devapur | FY20–23 | 46 | ₹0.60 cr | ₹3.6 cr/yr | 56.4 → 51.2 kWh/t cement |
| UltraTech, Reddipalayam | FY14–17 | 35 | — | ₹1.95 cr/yr | 746 → 689 kcal/kg and 72 → 64 kWh/t |
| Ramco Cements, Jayanthipuram | FY21–23 | 32 | — | ₹105.5 cr/yr | 744 → 733 kcal/kg; 83.27 → 79.35 kWh/t. Seven of the 32 cost nothing |
| JK Cement, Nimbahera | FY20–23 | 32 | ₹123.8 cr | ₹55.5 cr/yr | 773 → 756 kcal/kg; 93.29 → 70.3 kWh/t |
| Ambuja/Adani, Maratha | FY22–24 | 32 + 32 | ₹180 cr | ₹60 + ₹6 cr/yr | 32 capex projects and 32 zero-investment proposals worth ₹6 crore a year |
| JK Lakshmi, Durg | 16 months | ~12 | $0.42 m | $0.40 m/yr | 78.56 → 72.9 kWh/t; 715 → 704 kcal/kg. Payback 12.67 months |
Read the first row again. JK Cement Muddapur ran 149 separate projects over three years, spent ₹2.21 crore in total, and moved specific energy consumption by 5.03 kWh per tonne. That is an average of 0.034 kWh/t and about ₹1.5 lakh per project. The single largest item in the entire three-year register — a nozzle-ring modification on cement mill 3 — accounts for 10.6% of the total. The other 89% is 148 items nobody would put in a board pack.
Then read the Maratha row, which cuts the other way and should not be hidden. ₹180 crore of capex returning ₹60 crore a year is a three-year payback — comfortably outside the 1.2-year median threshold the literature says firms apply. Large Indian producers evidently do fund longer-payback capital when it is strategic. The hurdle-rate finding later in this section describes a tendency, not a law, and a consultant who treats it as a law will under-recommend.
These registers come from CII National Award for Excellence in Energy Management plant submissions, which are the single best public evidence of how cement plants actually improve. As of August 2026 the CII award archive no longer resolves publicly — the deep links are still indexed but return the CII-GBC homepage. So cite these as "CII National Award submission, [plant], [year]" and keep your own copy; do not send a client a link that will 404.
Two specific cautions in the register below. The Ramco cooler-silencer figure of 2.6 kWh/t clinker comes from the National Productivity Council's 2017 Good Practices Manual, not from CII, and it is large relative to a total cooler-area consumption of roughly 5 kWh/t — verify it against the original before you build a case on it. And JK Cement Mangrol's 37.3 → 27.9 kWh/t came with throughput going 100 → 227 tph: that is a capacity project that also improved specific power, not an efficiency optimisation, and presenting it as the latter misrepresents it.
The register — documented items, named plants
These are not generic opportunities. Every line has a plant behind it or sits in CII's published coefficient table for Indian conditions. This is the menu the model in §08 is built from.
| Thermal item | kcal/kg | Capex | Where it's documented |
|---|---|---|---|
| Replace burnt or missing dip tubes, lowest cyclones | 14.3 | ₹60 L | CII coefficient |
| Cut excess primary air by 10 percentage points | 14.3 | nil | CII coefficient |
| Substitute LD slag for limestone in the raw mix | 10.0 | nil | Dalmia Bharat, ROI 4.67 months of kiln running |
| Correct the preheater ΔP profile across all five stages | 8.0 | ₹375 L | UltraTech Reddipalayam, CFD study |
| Cut excess air by 10% | 7.2 | nil | CII coefficient |
| Restore a missing meal flap | 4.8 | ₹5 L | CII coefficient |
| Heat-resistant paint, preheater cyclones and kiln shell | 4.0 | ₹97 L | JK Nimbahera (1,720 t fuel/yr); Orient Chittapur (2.03 mo) |
| Cross-belt analyser on the raw mix | 4.0 | ₹300 L | CII coefficient |
| Graphite kiln-inlet seal replacing the pneumatic seal | 3.5 | ₹25 L | CII coefficient; 4–5 year life |
| Lower the dispersion box height in the riser duct | 3.5 | ₹15 L | CII coefficient |
| Cut dust loss from the top-stage cyclone | 3.5 | ₹20 L | CII coefficient |
| Low-conductivity kiln-inlet refractory, 1.4 → 0.6 W/m·K | 2.5 | ₹60 L | CII coefficient; shell below 250 °C |
| Cooler air-distribution pattern, 4th chamber | 2.0 | nil | Chettinad Kallur; clinker 190 → 180 °C |
| Rotor weigh feeder for coal firing, ±10% → ±1% | 2.0 | ₹300 L | CII coefficient |
| Coal mill table speed reduced, less water spray | 1.5 | nil | Ramco Jayanthipuram |
| Electrical item | kWh/t cli | Capex | Where it's documented |
|---|---|---|---|
| Remove cooler-fan silencers and seal cooler leaks | 2.60 | nil | Ramco Alathiyur; JK Nimbahera got ₹33.6 L/yr for ₹0 |
| Bell-mouth inlets on the cooler fans | 1.49 | ₹4 L | Orient Devapur (published as 1 kWh/t cement) |
| GRR slip control → VFD on the preheater ID fan | 0.96 | ₹800 L | Ramco Alathiyur, −200 kW |
| Fan inlet box geometry enlarged | 0.50 | ₹9 L | Maratha CW, 3-month payback; UltraTech Reddipalayam by CFD |
| Idle running eliminated — bag houses, transport, coal circuit | 0.30 | nil | Orient Chittapur; ACC Sindri |
| Standby auto-changeover page in the CCR | 0.20 | nil | ACC Sindri, ₹36 L/yr for ₹0 |
| Compressed air: 6.55 → 5.8 bar, seven leaks arrested | 0.15 | nil | CII compressed-air audits, ₹45 L/yr for ₹0 |
| Redundant dampers and orifice plates removed | 0.12 | nil | Dalmia Dalmiapuram |
| Cooler ESP field-current optimisation | 0.10 | nil | JK Nimbahera, ₹6.35 L/yr for ₹0 |
What capable plants got wrong
The second half of the picture, and the more useful half for a consultant. These are not badly run plants — several are award winners, and the errors below were found by the audits that won the awards. Competence does not prevent any of them.
| What was wrong | What it was worth | Why it survived |
|---|---|---|
| The cooler was never re-rated after the kiln was uprated. CII's design guidance is grate loading below 45 TPD/m²; it records overloaded plants running at 50–70 | Up to 55% above the design limit, with increased kiln heat loss attributed directly to it. The counterweight matters: CII's own benchmarked plants run 41–47, so this is a laggard problem, not a sector-wide one | The uprate project ended at the kiln. Nobody owned the sentence "and therefore the cooler is now undersized" |
| Silencers on the cooler fans. Fitted for noise, pure parasitic pressure drop, never questioned | 2.6 kWh/t clinker at Ramco Alathiyur — roughly 3–4% of a modern plant's entire electrical consumption. JK Nimbahera: ₹33.6 lakh a year, recovered for zero | It was there when everyone currently employed arrived. Nothing on any screen shows what it costs |
| A working cement mill, making its number. Holtec found ball charge at 80% of design, too many 100 mm balls, five broken liner plates, separator vent air at 4% of total flow — and the mill in manual | 42.73 → 33.60 kWh/t and 135 → 160 tph — 21% of the power and 19% of the output, and nothing was bought: no roller press, no new mill. Two caveats: the study dates from 1998, and both states are at ~2,700 cm²/g Blaine, so use the 21% relative gain rather than the absolute kWh/t | The mill hit its tonnage. Not one of the five was a design flaw — every one was a maintenance or selection decision that accumulated |
| A 1,500 m³/h compressor serving a 65 m³/h duty at UltraTech RDCW, losing 3.3 bar of a 6 bar setpoint in 222 m of undersized pipe, run 24 hours for an 8–10 hour requirement | −145 kW, eight-month payback on a dedicated small compressor | Four failure modes stacked in one asset: mis-selection, a distribution system eating half the pressure, a setpoint raised to compensate, and idle running. Each hid the next |
| A correctly-sized fan in the wrong duct. Yurt Çimento's ID fan lost over 15% in pressure to inlet geometry — vortex formation and flow asymmetry, not the impeller | CFD redesign: −26.2% shaft power, +19.6 points of efficiency, 2,527 MWh a year | From the control room, a correctly-sized fan in a wrong duct is indistinguishable from a correctly-sized fan |
| Measuring the fuel to ±10%. CII: conventional coal weigh feeders are accurate to ±10%; rotor feeders to ±1% | Not directly costed — but a ±10% error on fuel makes every heat balance, every kcal/kg number and every "the kiln is fine" conclusion unfalsifiable | The feeder works. It reports a number. Nobody asks what its error band is |
IFC states it plainly: "Optimum design margins for capacity and heat should not be greater than 10 percent, resulting in improved power consumption of up to 25 percent." Check the scope before quoting it — that sentence sits in IFC's section on high-efficiency fans and variable-speed drives for mill vents, so the 25% is fan power, not plant power, and the guidance is about sizing a mill-vent fan. It remains the sharpest published statement of the problem; it is not whole-plant design advice, and a plant engineer will know the difference. The mechanism is named by the fan standards body itself — DOE and AMCA describe designers who "compensate for uncertainties in the design process by adding capacity," stacked on top of a fouling allowance and an expansion allowance. Three margins, each individually defensible, multiplied together.
And the price of carrying them is not obvious: a fan running at 75% of its flow costs about 80% of rated power on a damper and about 42% on a variable-speed drive. The waste peaks precisely where a modestly oversized fan spends its whole life. Nothing on the control room screen says so.
Why it persists at competent firms
This is not an engineering question and you should not treat it as one. The evidence is unusually good.
- Roughly half of everything a competent engineer identifies is never done — but the exact figure turns on a definition, so state which one you mean. The US Industrial Assessment Center database holds 23,120 assessments and 169,952 recommendations as of August 2026. Counting only what was implemented gives 47–50%; a DOE-commissioned evaluation counting "implemented or with a concrete plan within one year" gives about 60%, and 45% for energy-management recommendations specifically. Whichever you use, the state-level spread on identical methodology runs from 38% to 65% — organisational, not technical.
- Firms apply a 1.2-year median payback threshold to energy projects — an implied hurdle rate of 65–80% — while accepting far lower returns on capacity projects. Anderson and Newell, studying 38,920 projects across 9,034 plants, also found firms weight a rupee of capex 40% more heavily than a rupee of saved energy. That asymmetry is the quantitative signature of capital and operating budgets being separated.
- The stated hurdles are discounting and people, not knowledge. A 2026 study of 106 on-site assessments across 74 firms found identified potential of 17% of consumption, at a median cost of saved energy well below what those firms pay for energy — and concluded that "high discounting and a shortage of skilled personnel required for implementation are the two main hurdles."
The consulting implication is direct and it is uncomfortable: your recommendation's odds are set less by its quality than by whose budget it lands in. A 2.0 kcal/kg item that a shift engineer can do next Tuesday will outperform a beautifully argued 25 kcal/kg project that needs a capital committee, because the second one has roughly a coin-flip chance of ever happening. Lesson X2 is about writing for that reality.
§ 07 The numbers that matter
| Parameter | India average | India best | Your plant | What drift tells you |
|---|---|---|---|---|
| Specific heat consumption, kcal/kg clinker | 740 | 670 | — | The headline. Everything below explains it |
| Burning-zone temperature, °C | ~1,450 | ~1,450 | — | Up = energy waste; down = free lime rising |
| Kiln inlet gas temperature, °C | 1,000–1,100 | — | — | How much heat leaves the kiln uncaptured |
| Preheater exit gas, °C | 300–380 | <320 | — | Rising = stage inefficiency or false air |
| O₂ at kiln inlet, % | 2–3 | 2–3 | — | High = false air; low = CO and fuel waste |
| O₂ at preheater exit, % | 3–5 | 3–5 | — | The gap vs kiln inlet is the leakage |
| Preheater ΔP, mmWG | 600–800 | <650 | — | Creeping up = build-up or dip tube loss |
| Secondary air temperature, °C | 1,000–1,200 | >1,100 | — | Falling = cooler losing recuperation |
| Clinker at cooler exit, °C | 100–150 | <100 | — | Rising = cooler efficiency deteriorating |
| Free lime, % | 0.5–1.5 | 0.8–1.2 | — | Rising = incomplete burn or chemistry swing |
| Kiln section power, kWh/t clinker | — | 16.8 | — | Fan power tracks false air closely |
| Refractory spend, ₹/t clinker | 20–40 | — | — | Rising on a stable fuel mix = a coating problem |
| Fuel cost, ₹ per '000 kcal | 1.53–1.90 | 1.53 | — | Sets what one kcal/kg is worth |
Provenance. SEC and kWh/t benchmarks: CII-GBC Energy Benchmarking v7.0, May 2025. The edition matters: v6.0 (2023) gave 726 average and 675 best, and its questionnaire went to "more than 110 plants" with a majority responding — so the sample is "over 100," not 110. The average moved up to 740 and the best moved down to 670 — the sector did not regress, the sample changed. Always cite the edition. Fuel cost per '000 kcal: JK 1.53, Ambuja 1.66 and UltraTech 1.90, all verified in the companies' own Q1 FY27 investor presentations. A widely repeated Shree figure of 1.95 could not be found in Shree's disclosures and has been dropped. Refractory spend: Indian Cement Review, December 2024. Operating ranges: the kiln tutorial and the EU BREF. Always state the vintage when you quote these.
§ 08 Build the model
First, by hand
The single most useful calculation in this entire course converts a thermal number into a commercial one. Do it once, slowly.
Step 1. A tonne is 1,000 kg, so 1 kcal/kg of clinker is 1,000 kcal per tonne of clinker.
Step 2. Fuel is priced per thousand kcal. So a tonne of clinker at 1 kcal/kg consumes exactly one pricing unit — 1,000 kcal ÷ 1,000 kcal = 1 — and therefore costs ₹1.90 per tonne of clinker.
Step 3. A 6,000 tpd line running 330 days makes 1.98 million tonnes of clinker a year.
Step 4. 1.98 × 10⁶ tonnes × ₹1.90 = ₹3.76 × 10⁶ = ₹37.6 lakh per year.
So: 1 kcal/kg ≈ ₹1.90 per tonne of clinker ≈ ₹0.38 crore a year on a 6,000 tpd line. Memorise the ₹1.90/t — it is the conversion factor you will use in every kiln conversation.
Step 2 looks wrong, and it is worth understanding why it isn't. The rupee figure does not change between Step 1 and Step 2, which makes the step read like a no-op. It isn't: the units transform while the digits stay put, because two thousands cancel. Fuel happens to be quoted per thousand kcal and a tonne happens to be a thousand kilograms. That coincidence is convenient — it is why ₹1.90 is memorable — and it is also exactly the kind of step where a unit error would hide undetected. Which is why you never trust a conversion factor until you have made it reproduce a real cost sheet.
Watch step 4. ₹3.76 × 10⁶ is 37.6 lakh, not 3.76 crore. The lakh–crore slip is the single most common arithmetic error in Indian plant work, it is wrong by a factor of ten, and it is always wrong in the direction that makes the opportunity look bigger than it is. Say the unit out loud before you say the number.
Now the number that matters: the gap between the Indian average (740) and the Indian best (670) is 70 kcal/kg. That is ₹133 per tonne of clinker, or ₹26.3 crore a year on this line. But read §06 before you present that as a target — it is not one project, it is about thirty items averaging 2 kcal/kg each.
Take a published cost line and see whether the factor reproduces it. UltraTech's Q1 FY27 deck gives two numbers independently: fuel cost ₹915 per tonne of cement, and ₹1.90 per '000 kcal. Work backwards from the first and see whether it lands somewhere believable.
₹915/t cement ÷ 0.67 clinker factor = ₹1,366 per tonne of clinker
₹1,366 ÷ ₹1.90 per '000 kcal = 719 kcal/kg clinker
India's average is 740 and best-in-class is 670 (CII v7.0, 2025). A large, efficient producer landing at 719 is exactly right — and note that nothing in that calculation used the 719; it fell out. The conversion reproduces a published cost line from two unrelated disclosures. That is the test to run whenever someone hands you a ₹-per-kcal number, and it takes thirty seconds.
Run it forward too, so you know the shape: at 670 kcal/kg fuel is ₹853/t cement, at 720 it is ₹917, at 740 it is ₹942. The whole national spread from best to average is about ₹90 a tonne of cement — which is the fuel line's entire addressable range, and worth holding next to the ₹1,005/t sector EBITDA.
Then, with the model
The model has two halves and the comparison between them is the lesson. On the left, the register from §06 — twenty-four documented items from named Indian plants. On the right, the four capital projects a consultant instinctively reaches for. Switch on everything in each half in turn and read the two subtotals against each other.
Select all 24 register items and all 4 capital projects in turn. The register delivers more saving for roughly a seventh of the capital — and about half of it needs no capital at all. That is not an argument against capital projects; the cooler replacement in the right-hand column also unlocked petcoke firing from 40% to 100%, which no register item can do. It is an argument about sequencing: the register is what you do first, because it is cheap, fast, reversible, and it re-baselines the plant so the capital case can be built on a real number instead of an aspirational one.
Note also which register items carry the most weight. The two largest — dip tubes and excess primary air, at 14.3 kcal/kg each — are not improvements at all. They are restorations of design intent. Someone let them slip.
Simple payback flatters capital and punishes maintenance. Sealing false air pays back in weeks — but the leaks come back, so the real question is whether the plant will sustain the discipline. A sixth preheater stage pays back in years — but it is permanent. Never present payback without saying whether the saving decays.
The register is a menu, not a plan. Its items sum to more than the entire average-to-best gap, which is impossible — because several are conditional. "Replace missing dip tubes" is worth 14.3 kcal/kg only at a plant that has lost them. Walk the register against the actual asset before you total it.
The shutdown line is what kills projects. A 30-day outage on a 5,000 tpd line forgoes roughly ₹20 crore of EBITDA — larger than the capex of most of these items. Add up the shutdown days before you add up the rupees. The model treats register items as concurrent — different crews, different parts of the line, one window — so it charges the longest single item, not the sum. Capital projects hold the whole line and are additive. Check that against the plant’s actual shutdown manning before you rely on it.
§ 09 Upgrade paths — what you can actually change
No cement plant abandons a working pyro line. The kiln shell has a 30–50 year physical life, the preheater tower structure 40–50, the foundations effectively forever — and Indian accounting depreciates the whole thing over 25 years as continuous process plant. A line commissioned in 2005 is near-fully depreciated and nowhere near worn out. Its book cost is close to zero; its cash cost per tonne is whatever its heat and power consumption happen to be. Replacing it means writing off nothing and paying roughly ₹6,900 per tonne of capacity for a greenfield line.
So every recommendation you make here is a retrofit, and retrofits sort themselves into five tiers by what they preserve. That is the useful axis — not capex, and not saving.
| Tier | What survives | Typical interventions | Capex | Shutdown | kcal/kg |
|---|---|---|---|---|---|
| 0 · Operating point | Everything | O₂ setpoints, kiln/calciner fuel split, cooler bed depth and grate speed, kiln speed | nil | none | 5–15 |
| 1 · Wear items | Shell, tower, cyclone bodies, cooler casing | False-air sealing, kiln inlet seal, dip tube replacement, refractory upgrade, multi-channel burner | ₹2–25 cr | 5–15 d | 15–30 |
| 2 · Bolt-ons | Kiln, tower structure, cooler casing | APC layer, cooler retrofit reusing the casing, AF feed and dosing system, cross-belt analyser | ₹5–50 cr | 10–30 d | 20–45 |
| 3 · Major retrofit | Kiln shell, foundations, tower structure | Full cooler replacement, sixth preheater stage, calciner modification for AF, chlorine bypass, WHRS | ₹55–225 cr | 25–50 d | 8–30 |
| 4 · Replacement | Land, lease, approvals | New pyro line | ₹6,500–8,000/t | 3–4 yr | — |
The four that matter, with real Indian numbers
- False-air sealing (Tier 1). The best capex-to-saving ratio in the plant, by a wide margin. LBNL puts seal replacement at an operating cost of no more than $0.5 per tonne with a payback of roughly four days. Graphite kiln inlet seals last 4–5 years. The Indian audit cited in §05 got 5.5 kcal/kg from one seal.
- Cooler upgrade (Tier 2 or 3). India has actuals here, not vendor claims. UltraTech's Rajashree line 3 recorded −30.8 kcal/kg and −1.75 kWh/t with clinker exit falling 181 → 120 °C — and critically it took petcoke firing from 40% to 100%. JK Lakshmi recorded −19 kcal/kg, 193 → 100 °C. A 12,000 tpd cooler replacement in the US was completed 48 days after shutdown; retrofits that reuse the existing casing and kiln hood have been done inside normal maintenance windows.
- Advanced process control (Tier 2). This is where you must be careful. CII's Indian figure for automation is −6 to −8 kcal/kg. Vendor and AI claims run −3 to −5%, which on a 720 kcal/kg line is −22 to −37. One documented Indian plant went 750 → 712. The gap is the baseline: vendor numbers are measured against a poorly-controlled manual starting point. For a plant already at 700–720 with a competent control room, budget the lower end and say why.
- Sixth preheater stage (Tier 3). The lever everyone reaches for, and the one where the published evidence genuinely conflicts — so carry both numbers. CII's thermal manual gives −8 to −12 kcal/kg for a 5→6 conversion (4→5 is worth −18 to −25, a different project). But CII's own 2025 benchmark shows the best 5-stage plant at 694 and the best 6-stage at 670 — a 24 kcal/kg gap, matching what is derivable from ECRA. Those measure different things: a retrofit delta on one kiln, versus a cross-section of different plants where the 6-stage ones are newer in every other respect too. The honest reading is that a retrofit lands nearer the low end and the cross-sectional gap flatters it. Either way, at ₹56–89 crore of European capex this is an order of magnitude worse per kcal/kg than sealing or cooler work.
Adding a sixth preheater stage drops the heat available at the preheater exit from about 165 kcal/kg to 140. That is exactly the heat a waste heat recovery plant was going to convert into electricity. The sixth stage and the WHRS are competing for the same ~25 kcal/kg. A plant can fund both projects and get the benefit of one. If a client has both on the capital plan, that is the first thing to raise — and it will be the most valuable thing you say that day.
Indian producers apply an IRR above 20% and a 2–3 year payback to fast-track modernisation. Applied literally, that screen passes burners, seals, mill internals, APC and separators — and fails WHRS, precalciner conversion and cooler replacement on energy grounds alone. Which is precisely why those three are always argued on capacity uplift or fuel flexibility instead. Know which argument your recommendation needs before you make it. The Rajashree cooler was worth 30.8 kcal/kg, but what sold it was going from 40% petcoke to 100%.
§ 10 The frontier, and who is running it
§09 is what this plant can do. This is what the best plants in the world are doing, by name, with their status as at October 2026. The column that matters most is status — a large part of what circulates as cement technology is an announcement, and the gap between an announcement and a running machine is where credibility is lost.
| Status | Technology | Plant, company, country | The number |
|---|---|---|---|
| OPERATING | Oxyfuel kiln, large scale | catch4climate, Mergelstetten — SCHWENK plant, owned by CI4C (Buzzi/Dyckerhoff, Heidelberg, SCHWENK, Vicat), Germany. Process by thyssenkrupp Polysius. | 450 t clinker/day. First clinker end-May 2026, oxygen mid-June, inaugurated 8 July 2026. Over €120 m, entirely privately funded. The first large-scale oxyfuel cement kiln anywhere, and the most consequential pyro event since 2024. |
| OPERATING | Amine capture on a cement kiln | Brevik — Heidelberg Materials, Norway. SLB Capturi. Reboiler runs on process waste heat, no extra steam generator. | Design 400 kt CO₂/yr. Opened 18 June 2025. Actual: ~185 kt over the 13 months to August 2026 — about 43% of design, with two months at zero. Abatement cost NOK 1,150 ≈ €100/t. See the box below. |
| OPERATING | Oxyfuel capture, commercial | Qingzhou — China United Cement (CNBM), Shandong, China. | 200 kt CO₂/yr, kiln exhaust above 80% CO₂, then PSA and low-temperature distillation to >99.9% purity. Running since 9 January 2024. ~RMB 260 m. |
| OPERATING | Chemical absorption, first mover | Baimashan — Anhui Conch, Wuhu, China. | 50 kt CO₂/yr, built 2018. The plant manager told ICR it is loss-making: “if we only considered the economics we wouldn’t have done it.” Worth quoting when a client asks why nobody has copied it. |
| OPERATING | AI kiln control | Siggenthal — Holcim, Switzerland. Aixprocess Kiln AIxperT, hybrid model. | 150–200 kJ/kg off specific thermal energy — that is 36–48 kcal/kg, which is larger than most capital projects in §09. Disclosed at Global CementAI, Brussels, May 2026. |
| OPERATING | AI, quality variance | Els Monjos — Cementos Portland Valderrivas, Spain. Fuller ECS/ProcessExpert PXP 9.1 with Imubit soft sensors. | 17% reduction in free lime standard deviation. The right pairing to look for: one plant discloses heat rate, the other discloses variance, and §04 row 8 says variance is what buys the heat rate. |
| OPERATING | Fourth-generation cooler | Nawalgarh — Shree Cement, Rajasthan, India. FLSmidth Cross-Bar, 325 m² grate, hot-air recirculation. | Rated 11,500 tpd, running at 13,695 tpd average since December 2023. Specific air consumption 1.28 → 0.15 kg/kg, vent temperature 355 → 164 °C. ~8 MW recovered from the cooler alone, 15–18 MW plant WHR. The best-documented large cooler installation in India. |
| OPERATING | Sustained high TSR | Solnhofen — Solnhofer Portlandzementwerke, Germany. | 90–95% thermal substitution, up from 60–65% after a new alternative-fuel handling unit. It already had a chlorine bypass — which is §04 row 6, and is what made the step possible. Rohožník (Danucem, Slovakia) runs >80%. |
| OPERATING | Organic Rankine Cycle | Medcem, Türkiye — largest operating ORC on a cement plant. Turboden now lists ~14 operating cement ORC units. | 11.0 MWe, 2025. ORC handles thermal input swings from 20% to 110% of design load and runs water-free — which is why it appears on coolers rather than preheaters. |
| PILOT, DONE | Plasma-heated kiln | Slite — Heidelberg Materials, Sweden. | 300 kW, 54 hours continuous, 60% CO₂ in the flue gas. Completed February 2025. A 1 MWe furnace at Skövde planned for 2026. The best real operating data on plasma calcination that exists. |
| PILOT | Electrified calcination | Hofors — SaltX with Holcim, Sweden. Electric Arc Calciner with plasma burners. | Portland-quality clinker made by a fully electrified process at industrial scale on Holcim raw meal, August 2026. Stated aim: Europe’s first fully electric cement plant by 2028. No MW or tpd disclosed. |
| PILOT | Direct separation calcination | Leilac-1, Lixhe, Belgium. Leilac-2 is at Ennigerloh, Germany — not Hannover, which closed clinker production in 2024. | Leilac-1 demonstrated direct separation at 95% purity with no air ingress, on ~5% of a plant’s process CO₂, with no added chemicals. Leilac-2 is designed for 100 kt/yr and is not operating; the project’s own published dates are stale. |
| BUILDING | Full-scale oxyfuel | Lägerdorf (Carbon2Business) — Holcim, Germany. | 1.2 Mt CO₂/yr, pure oxyfuel. Ground broken April 2024, €109.8 m EU Innovation Fund. Delayed on German CO₂ storage legislation, noted November 2025. No published commissioning year. |
| BUILDING | Capture at ~95% of a works | Padeswood — Heidelberg Materials, UK. | 800 kt CO₂/yr from kiln and CHP. FID 25 September 2025, operational 2029, CO₂ to Liverpool Bay via HyNet. |
| DELIVERY ONLY | Electrified heater (RDH) | Boyareddypalli — Adani Cement, Andhra Pradesh, India. Coolbrook RotoDynamic Heater, ABB electricals. | Claimed ~60 kt CO₂/yr abatement. Status as at July 2026 is “industrial delivery phase” — equipment delivery, not operation. No MW rating and no start-up date published. Headlines calling it the world’s first commercial RDH describe an order. |
| ON HOLD | Full-scale capture | Edmonton — Heidelberg Materials (Lehigh), Alberta, Canada. | 1 Mt CO₂/yr, up to C$2 bn, promoted for years as the world’s first full-scale net-zero cement plant. Put on hold August 2026 because carbon credits trade near C$45/t against the ~C$170/t the case assumed. Cement decarbonisation capex is policy-priced, not technology-priced. |
Heidelberg Materials has published no captured tonnage from Brevik. But the CO₂ goes to Northern Lights’ Aurora store, and the Norwegian Offshore Directorate publishes monthly injection volumes per wellbore. Until Yara Sluiskil joined the chain on 7 September 2026, everything injected at Aurora was Brevik’s.
August 2025 to August 2026: about 185,000 tonnes over thirteen months — roughly 43% of the 400 kt/yr design. Two of those months recorded zero injection. Only two came within 10% of the design monthly rate. Injection is not identical to capture in any single month because of ship scheduling, so the annual comparison is sound and the monthly one is not.
Use this carefully and use it often. It is the only cement carbon-capture plant in the world whose real output can be independently audited, the audit is free and updated monthly, and the figure in general circulation is the design figure. When a client’s decarbonisation roadmap assumes capture performs to nameplate, this is the number that re-bases it.
The sixth preheater stage is not the lever it is sold as. Best-in-class worldwide is 2,824 MJ/t clinker on six stages against 2,858 MJ/t on five — a difference of 34 MJ/t, or about 8 kcal/kg. Over 2014–2023 five-stage plants improved 3.4% and six-stage only 1.6%. Against the tower height, the extra pressure drop, the fan power and the capex, the sixth stage has nearly stopped paying, and the six-stage curve has flattened. This matches the CII India figure of −8 to −12 kcal/kg used in §09.
The AI numbers are larger than the capital ones, and nobody will tell you the catch. Siggenthal’s 150–200 kJ/kg is 36–48 kcal/kg from software. But no vendor and no plant discloses closed-loop uptime — what fraction of kiln hours the system actually holds control. A system that runs 30% of the time delivers 30% of the benefit. Ask for the uptime figure, in writing, before the business case.
“Taizhou, 500 kt cement CCUS.” It is a coal-fired power station — CHN Energy Taizhou, a 1,000 MW ultra-supercritical unit. It appears in cement decks regularly. Its figures are a legitimate amine benchmark (2.35 GJ and 51.5 kWh per tonne of CO₂, 90.86% capture) but it is not a cement plant.
“Małogoszcz runs at 100% alternative fuel.” Holcim Polska reached 100% TSR for several days as a trial campaign, reported May 2026, announced on social media. No plant anywhere sustains 100% TSR in normal operation.
“ECRA’s oxyfuel pilots at Colleferro and Retznei.” Announced January 2018; no evidence either was built. ECRA’s own 2025 listing does not include them. The industry’s oxyfuel effort went to Mergelstetten and Lägerdorf instead.
“Riyadh Cement has commissioned its ORC.” Turboden’s own December 2025 reference list still shows 12.6 MWe under construction.
§ 11 What's changing now
Current as of August 2026 · refresh every six monthsFive developments a consultant advising an Indian cement plant would look uninformed for not knowing. Each is dated, and each carries the thing it actually means for your client rather than the headline.
catch4climate at Mergelstetten was inaugurated — the world's first pure second-generation oxyfuel cement kiln at demonstration scale. 450 tpd clinker, thyssenkrupp Polysius process, first clinker in May 2026, over €120 million and entirely privately funded by four producers with no public money. Capture figures have not been published yet.
What it means: oxyfuel is no longer a slide. But it changes the combustion environment fundamentally — this is a Tier 4 conversation, not a retrofit, and the absence of published capture data means nobody should be modelling it yet.Coolbrook's RotoDynamic Heater is going into Adani Cement's Boyareddypalli plant in Andhra Pradesh — announced November 2025 as the first commercial-scale deployment, ~1,000 °C, initially ~60,000 t CO₂/yr and scalable tenfold, with ABB as delivery partner and five further deployments planned. In June 2026 Leilac/Calix signed a joint development agreement with Ambuja for a hybrid electric heating and capture retrofit at the 6.6 Mt/yr Sanghi plant. Meanwhile Leilac's European demonstration at Hanover was cancelled in 2024.
What it means: two things. First, the deal structure is new and worth understanding — Ambuja funds 100% of development and construction; Leilac contributes no capital and takes royalties. Expect that offer to reach other Indian majors. Second, no commissioning date has been published for the Coolbrook unit. Do not describe it as operating.Heidelberg's Brevik plant — the world's first industrial-scale cement carbon capture — injected roughly 105,000 tonnes between August 2025 and the end of May 2026, against a 400,000 t/yr design. Quote it that way rather than as a percentage: 105 kt over ten months is a 126 kt/yr run-rate, about 32% of nameplate, while 105 against 400 reads as 26%. Both circulate; they use different denominators. Heidelberg calls it normal ramp-up, though it had earlier said ramp-up was complete by summer 2025.
What it means: this is the most useful contrarian fact available to you. Any 2030 or 2050 pathway, any internal carbon price, any CCUS-dependent roadmap should be stress-tested against 26%. NITI Aayog's own roadmap concedes India's target is unreachable without CCUS — and CCUS just posted a quarter of design in its flagship year.The Solid Waste Management Rules 2026 require cement plants burning solid fuel to progressively replace it with refuse-derived fuel, from a 5% baseline to 15% within six years. India sits at 4–8% TSR against Europe's 40–50%. NITI Aayog targets 20–25% by 2030.
What it means: this lands directly on the preheater. Plastic-rich RDF carries chlorine into the kiln, chlorine condenses in the tower, and build-up follows. Yet no published data on Indian chlorine-bypass adoption could be found at all — not a low number, no number. If a client is planning 5% → 15%, ask about the bypass first. That nobody can answer is itself the finding.Gigaton — a UCL/Cambridge spinout that raised $26 million in June 2026 — replaces the existing control software outright rather than sitting on top of it, autonomously adjusting fuel mix, kiln speed and oxygen. It has been running at JK Cement's Mangrol plant since about October 2025, with Adani's Marwar Mundwa announced. Claims run to $1–3 million a year per plant.
What it means: treat the claim structure carefully. There is no published independent result from Mangrol, and the gap between CII's −6 to −8 kcal/kg for automation and vendor claims of −22 to −37 is entirely about the baseline. A closed heat balance is your only defence in that negotiation — which is lesson C5.Two more worth carrying, briefly. CCTS compliance obligations entered force on 30 March 2026 covering roughly 490 units including cement, with an FY24 baseline, FY26 and FY27 compliance years, and a penalty of twice the average credit price — but as of August 2026 there is no confirmed first trading session and no published price band, so do not tell a client Indian carbon trading has started. And the RHI Magnesita–Khemka MINPRO joint venture completed in 2026 with a refractory recycling plant in Odisha, which turns spent refractory from a waste line into a procurement lever.
§ 12 Check yourself
§ 13 Mini case
| Parameter | January | August | Change |
|---|---|---|---|
| Clinker production, tpd | 6,050 | 5,880 | −170 |
| Specific heat consumption, kcal/kg | 705 | 748 | +43 |
| Preheater exit gas, °C | 320 | 355 | +35 |
| O₂ at preheater exit, % | 3.8 | 5.6 | +1.8 |
| O₂ at kiln inlet, % | 2.4 | 3.9 | +1.5 |
| Secondary air temperature, °C | 1,080 | 990 | −90 |
| Clinker at cooler exit, °C | 105 | 138 | +33 |
| Kiln shell maximum, °C | 285 | 292 | +7 |
| ID fan power, kWh/t clinker | 8.1 | 9.4 | +1.3 |
| Raw meal residue, % R90 | 13.4 | 13.5 | +0.1 |
| LSF standard deviation | 0.8 | 0.9 | +0.1 |
| Free lime, % | 1.2 | 1.3 | +0.1 |
Select the two primary causes
What the data says
Both O₂ readings rose while fuel rate did not. That is the signature of false air. Air entering at ambient temperature and leaving at 355 °C has been heated for nothing, and the ID fan has to move it — which is exactly why fan power rose 1.3 kWh/t at the same time. The rise in preheater exit temperature is the same story from a different instrument.
Secondary air fell 90 °C while clinker exit temperature rose 33 °C. The cooler is recovering less heat and passing it to the clinker instead. Less hot secondary air means the burner must supply more of the burning-zone heat itself.
What it is not. Raw meal fineness moved 0.1 points, LSF standard deviation 0.1, free lime 0.1 — all inside noise. The feed is fine. Shell maximum rose 7 °C, which is not a refractory failure; a real coating loss shows 30–60 °C and a localised hot spot. And there is no burner problem: a bad flame raises free lime and creates shell hot spots, and neither happened. The plant head's hypothesis is wrong, and the data says so in three places.
Now size it
| Component | Working | ₹ crore / yr |
|---|---|---|
| Extra fuel | 43 kcal/kg × ₹1.90 = ₹81.70/t; × 5,880 tpd × 330 d = 1.94 Mt | 15.9 |
| Lost production | 170 tpd × 330 = 56,100 t clinker → 83,700 t cement at 0.67 clinker factor, × ₹1,000/t EBITDA | 8.4 |
| Extra ID fan power | 1.3 kWh/t × 1.94 Mt × ₹4.73/kWh | 1.2 |
| Total annual value at risk | Before any capex | 25.5 |
On the ₹1,000/t. There is no citable "sector EBITDA per tonne" — it is a brokerage aggregate whose value depends on which companies are in it. Q1 FY27 disclosed actuals: UltraTech ₹1,214, Dalmia ₹1,055, JK Cement ₹982, Ambuja ₹931. ₹1,000/t is a round working figure inside that range, and it is exactly the sort of number to replace with the client's own before anything reaches a slide.
Now place it in a tier
Both causes are Tier 1 and Tier 2 — the sealing campaign (₹2–3 crore, days inside the annual shutdown) and cooler optimisation covering bed depth, grate speed, undergrate air distribution and the cooler's fixed inlet section (₹3–4 crore, and it may fit the window if the casing is reused). Roughly ₹6 crore of work against ₹25 crore a year. Payback under four months, and no burner upgrade.
Note what that means for the shutdown calendar: this whole programme fits one annual window and does not compete with anything structural. If the plant were also considering a sixth preheater stage — Tier 3, ₹55 crore-plus, 25–50 days, and only 8–12 kcal/kg on CII's Indian figure — the sequencing argument writes itself. Do the ₹6 crore of Tier 1 and 2 work first, then re-measure, then decide whether the Tier 3 project is still needed. Half the time it isn't.
The plant head was not being unreasonable — a burner upgrade is a real, visible, satisfying intervention and the burning zone is where the drama is. Your value here was not superior process knowledge. It was reading two O₂ readings, a secondary air temperature and a fan power number together, and noticing that they all move for the same reason. That is a 20-minute analysis that stopped a capital project aimed at the wrong equipment.
§ 14 Go deeper
Every link below has been checked. The free cement-resource landscape has degraded since 2023 — two of the best previously-free sites have moved behind paywalls — so this list is deliberately short and weighted toward things that will still be there next year.
§ 15 Carry forward
- "On a line this size, one kcal per kg is worth about ₹1.90 a tonne, so roughly ₹38 lakh a year. The gap between you and the Indian best plant is about 70 kcal — call it ₹26 crore. I'm not expecting one project to close it. JK Cement Muddapur closed five kilowatt-hours a tonne with a hundred and forty-nine projects averaging two lakh each. Can I see your encon register?"
- "About three-quarters of your heat input is doing work you cannot avoid — calcination alone is 410. So I'm not going to talk about the burning zone. I want to look at false air, the cooler, and your refractory condition, because that's where the addressable heat is."
- "Your kiln inlet O₂ and preheater exit O₂ have both risen while your fuel rate hasn't. That's not a combustion problem, it's a sealing problem — and your ID fan power should confirm it. Can I see the fan trend?"
- "Before we talk capex: how many days is your annual shutdown, and what's already booked into it? Everything I'm going to recommend has to fit in that window, and the window is worth about ₹65 lakh a day in forgone EBITDA."
§ 16 Where this connects
This lesson deliberately stops at the boundary of the pyro line. Five other lessons pick up the threads it leaves open.