# Scale, capacity and feedstock: which chemical plants get built: Capacity planning, utilisation and the chemical cycle

Source page: https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p3/

[Understanding Chemicals](https://www.materialinsights.org/learn/chemicals/) / [Module 7: Scale, capacity and feedstock: which chemical plants get built](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/) Part 3 of 5

## Capacity planning, utilisation and the chemical cycle

Capacity arrives in large lumps years after the decision to build it, while demand grows smoothly and unpredictably. Getting the size and timing right decides whether a plant runs full or sits half empty. This part covers the plant-size decision, why utilisation matters so much, how existing plants creep up in capacity, and why the whole industry tends to overbuild at the same time.

### 3.1 Capacity comes in lumps, so choosing a plant size is a bet on future demand

Brennan frames the problem simply: demand is rising towards a company's existing capacity. It can expand the existing plant, build a new plant for the extra demand, or build a new plant large enough to replace the old one too (pdf p. 220). Expansion is cheap but keeps old technology. A replacement plant brings the best technology and scale but needs so much capital that it cannot be done often.

Built too small, a new plant gains little from scale and runs short of capacity after a few years of growth. Built too large, it runs below capacity for years. Brennan's preferred test is net present value (NPV), the value today of all future cash flows discounted at the cost of capital; one looks for the size at which NPV peaks (pdf pp. 228-229). Building bigger means more capital and fixed cost before the plant is full, in return for more sales later.

The best size depends on the demand curve, how costs vary with size, the discount rate, project life, how prices move against costs, and the room for later debottlenecking. A lower discount rate and a longer life favour the larger plant. Falling margins and the chance to expand later favour the smaller one (pdf pp. 229-231). A smaller plant can show a higher percentage return yet a lower NPV, so the return rate alone can mislead.

Brennan concludes that sizing is a strategic judgement, not a mathematical optimisation, because long-term inputs are so uncertain (pdf p. 231). Timing adds to the difficulty: two to three years usually pass from sanction to commissioning, and several more before demand catches up (pdf p. 38).

Data Phased versus single investment in a gas treatment plant

| Item | Scheme A: one plant | Scheme B: two plants |
|---|---|---|
| Plants built | 2020 | 2020 and 2023 |
| Capital ($ million) | 36.2 | 22.3 + 22.2 = 44.5 |
| Annual fixed operating cost, 2024-32 ($ million) | 3.6 | 4.45 |
| NPV at 10% ($ million) | 35.31 | 32.73 |
| Flexibility if demand grows slower | None: capacity already built | Second plant can be deferred or downsized |

 Source: Brennan (2020), Example 6.6 Phased capital investment, pdf pp. 165-168  Worked example

#### One big plant or two smaller ones?

In Brennan's Example 6.6, gas sales grow 8% a year from 2021 to 2032, and capital is 20 x C^0.7 $ million for capacity C. Scheme A builds one $36.2 million plant in 2020 for 2032 demand. Scheme B builds a $22.3 million plant sized for 2023 and a second $22.2 million plant for 2024 on. At 10%, A's NPV is $35.31 million against B's $32.73 million, about 8% higher (pdf pp. 166-168). Brennan still sees value in B: if growth is only 5%, the second plant can be delayed or shrunk, and it can use lessons from the first. That right to wait has a value the base-case NPV leaves out.

- Brennan (2020), Ch 8 Capital investment decisions, 8.6 · pdf pp. 220, 228-231
- Brennan (2020), Ch 2 and Example 6.6 · pdf pp. 38, 164-165

### 3.2 Utilisation decides whether a big plant's cost advantage is real

A plant's annual capacity is its maximum daily output times its equivalent availability, the days a year it can effectively run at that rate after outages and restrictions (Brennan, pdf p. 213). Utilisation is the share of that capacity actually used. Because so much cost is fixed, every tonne not made raises the cost of the tonnes that are.

Brennan's LNG regasification estimate shows the effect. A 100 PJ/year terminal with $300 million of fixed capital costs $50 million a year in total at full use, or $0.50/GJ. At 50% utilisation, even after allowing for lower utilities and selling costs, $42 million is spread over 50 PJ: $0.84/GJ, about two-thirds higher (pdf p. 240). The acid plant in Part 1 behaves the same way. In its first year it sells 100,000 t from a 150,000 t plant, so fixed cash cost rises from $34.70 to 5.2/0.1 = $52/t.

Wittcoff explains why this shapes the whole industry. With high fixed and low variable costs, there is a wide gap between the break-even price, where price equals total cost, and the shut-down price, where price falls below cash cost (pdf p. 50). In between, a plant loses money on paper but still covers its cash costs, so it keeps running. Scale can even make it rational to run flat out and cut the price. This underlies what Wittcoff calls the industry's chronic overcapacity.

Utilisation also reflects cost position. The IEA expects low-cost Middle East crackers to keep higher utilisation, while China's coal-based plants have tended to run at low rates, a sign of overcapacity (pdf pp. 34, 79).

Data Cost of running below capacity

| Case | Full use | Reduced use | Change |
|---|---|---|---|
| LNG regasification, 100 PJ/yr | $0.50/GJ at 100% | $0.84/GJ at 50% | +68% |
| Acid plant fixed cash cost, 150 kt/yr | $34.7/t at 100% | $52.0/t at 67% (year 1) | +50% |
| Acid plant total cash cost | $41.5/t at 100% | $50.2/t at 80% | +21% |

Acid plant rows computed from Brennan's Example 6.8 cost data.
 Source: Brennan (2020), Table 8.2 and Example 6.8, pdf pp. 174-176, 240  Common misreading

#### Running at 80% is not 'nearly full'

At 80% utilisation, fixed cost per tonne rises by 1/0.8 = 1.25, a 25% jump. For Brennan's 150,000 t/year acid plant that lifts fixed cash cost from $34.70 to $43.40/t and total cash cost from $41.50 to $50.20/t, about 21% higher. In 1980 US sulphuric acid plants ran at about 80% of capacity (Shreve, pdf p. 346). In a downturn, a penalty of that size can decide which plants close.

- Brennan (2020), Ch 8 · pdf pp. 213, 240
- Wittcoff (2004), Ch 1 · pdf p. 50
- IEA (2018), Ch 2 and Ch 4 · pdf pp. 34, 79
- Shreve (1984), Ch 19 · pdf p. 346

### 3.3 Existing plants creep up in capacity, but rarely catch new world-scale plants

No plant is perfectly balanced. Designers add margins at each step, so a real plant has a capacity distribution. Brennan's example has five steps rated Y, 1.2Y, 1.4Y, 1.2Y and 1.2Y: the first step is the bottleneck, and raising it alone gives 20% more output (pdf pp. 222-223). Operators also learn to push equipment beyond its rating. Debottlenecking, replacing equipment, adding units and intensifying the process all grow an old plant.

How fast it can grow depends on its structure. Plants that can add capacity in parallel at every step grow by up to about 10% a year; plants with at least one step that cannot be duplicated are typically limited to about 4% (pdf p. 223). In Australian plants with constrained expansions, capacity typically rose to about 1.6 times its starting level, with a time constant of about 2 years, while cumulative capital rose with capacity at an average exponent of 0.64 (pdf pp. 224-225).

Expansion is cheap capacity, and per tonne it improved raw materials by 0.6% a year, energy 1.5%, personnel 4.5% and fixed capital 3.0% (pdf p. 226). But new plants of best available technology and scale, which Brennan calls BATS plants, moved faster. In the 1960s new-plant scale grew 13-26% a year, far above what constrained expansions managed, so plants that started small fell further behind (pdf p. 227). Expanded plants also inherit older safety and environmental standards. And firms consistently underestimate how long and how large a plant will eventually run (pdf p. 228).

Data What expansion delivered in Australian plants

| Measure | Typical value |
|---|---|
| Capacity growth, parallel-streamed expansion | up to about 10% a year |
| Capacity growth, constrained expansion | about 4% a year |
| Ultimate/initial capacity in a growth phase | about 1.6 |
| Time constant of capacity growth | about 2 years |
| Exponent of cumulative capital on capacity | 0.64 average |
| Exponent of staff numbers on capacity | 0.18 average, often 0 |
| Annual improvement per tonne: raw materials / energy | 0.6% / 1.5% |
| Annual improvement per tonne: personnel / fixed capital | 4.5% / 3.0% |

 Source: Brennan (2020), Ch 8, 8.6.2-8.6.3, pdf pp. 223-226  Case

#### Nine furnaces against six

By 1983 an Australian ethylene plant had been expanded to 102,000 t/year, partly by adding furnaces, but its columns and main compressors stayed single stream. It had nine cracking furnaces; a new 250,000 t/year plant needed six. An expanded 120,000 t/year aluminium smelter had 544 pots, while new 240,000 t/year smelters had 440 (Brennan, pdf p. 227). More, smaller units per tonne mean more operators and maintenance per tonne.

- Brennan (2020), Ch 8, 8.6.2-8.6.5 · pdf pp. 222-228
- Kent (2012), Ch 22 · pdf p. 893

### 3.4 Across the industry, firms tend to build at the same time, and the cycle follows

Wittcoff calls the industry's habit of overbuilding, and building at the wrong time, 'The Old Paradigm' (pdf p. 50). Rising demand lifts prices and profits. Profits prompt investment. New plants take years to build, so supply lags and companies overshoot. The new capacity often arrives as business turns down, so prices and profits fall. Investment is cut and plants close, and the closures overshoot until demand runs ahead again. The commonly quoted cycle is four to five years (pdf p. 51).

US ethylene is the textbook case. By the late 1970s capacity far exceeded demand. Restructuring from 1986 cut US chemical employment by 7-8% and ethylene capacity by about 10%. With demand growing and buyers stockpiling, 1987-88 was probably the most profitable period the US industry has had. Then companies debottlenecked and built again, and by 1991 there was overcapacity once more (pdf pp. 48-50). In Western Europe, new crackers took the average operating rate down to about 83% in 1993-94 (pdf p. 50).

Wittcoff gives several reasons for this. High fixed costs keep loss-making plants running. Scale encourages running full at lower prices. Contractors build turnkey plants for anyone with capital, lowering entry barriers. Countries with cheap oil and gas see chemicals as a route to industrialise (pdf pp. 48, 50). Brennan adds that over-optimistic forecasts produced worldwide petrochemical overcapacity in the early 1980s, and that projects often fail when two companies build for an opportunity that can support only one (pdf pp. 38, 241). Ammonia shows the same pattern: world capacity peaked at 141 Mt in 1989, then high-cost plants shut (Kent, pdf p. 886).

Data The ethylene cycle in the United States and Western Europe

| Period | What happened |
|---|---|
| Late 1970s | US ethylene capacity far exceeds demand |
| 1986 | Restructuring: staff cut 7-8%, capacity cut about 10% |
| 1987-88 | Probably the most profitable years for the US industry |
| 1989-91 | Debottlenecking and new build; overcapacity again by 1991 |
| 1993-94 | New European crackers; operating rate down to about 83% |
| 1994-95 | Profit 'fly-up' after cracker accidents cut supply |
| Late 1990s | European operating rate about 92% |

 Source: Wittcoff (2004), Ch 1, pdf pp. 48-52  Common misreading

#### A 'shortage' is only a few per cent

In 1994-95 profits 'flew up' after accidents in plants around the world, including three US crackers that blew up. Wittcoff points out that in the chemical business a shortage is only a 2-3% shortfall (pdf p. 52). The plants were repaired and the fly-up ended. Because margins swing hard on such small imbalances, one large new plant, or two firms building at once, can turn a tight regional market into a glut.

- Wittcoff (2004), Ch 1 · pdf pp. 48-52
- Brennan (2020), Ch 2 and Ch 8 · pdf pp. 38, 241
- Kent (2012), Ch 22 · pdf p. 886

### See it in the table
[Al Aluminium](https://www.materialinsights.org/commodities/aluminum/)[H2SO4 Sulphuric acid](https://www.materialinsights.org/chemicals/sulphuric_acid/)[NH3 Ammonia](https://www.materialinsights.org/chemicals/ammonia/)[C2H4 Ethylene](https://www.materialinsights.org/chemicals/ethylene/)

### Check your understanding

According to Brennan, which change in assumptions favours building the larger of two plant sizes?
 A longer life and a lower discount rate give more weight to the later years of full-output sales. Falling margins and expansion potential favour the smaller plant.

Why do chemical plants often keep running when prices fall below their total cost?
 Wittcoff explains the wide gap between break-even price (total cost) and shut-down price (cash cost) in a high-fixed-cost industry, which keeps capacity online and prolongs gluts.

An existing plant whose distillation section cannot be duplicated is expanding. What capacity growth rate should be expected?
 Brennan reports constrained expansions are typically limited to about 4% a year, against up to about 10% for parallel-streamed expansions.

[← Part 2: How plant sizes grew, and what stops them growing](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p2/)[Mark read · Part 4: Technology, learning and feedstock: where the cost floor is set →](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p4/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/index.md)
- [Economies of scale: why big plants make cheap tonnes](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p1/index.md)
- [How plant sizes grew, and what stops them growing](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p2/index.md)
- [Technology, learning and feedstock: where the cost floor is set](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p4/index.md)
- [Integration, uncertainty, safety and sustainability: what decides what gets built](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/chemicals/scale-capacity-feedstock/review/index.md)