Capstone: environmental compliance and carbon reduction plan for the expansion

Nine months after the board's September meeting, Wren Halvorsen and Theo Laurent have the pieces of the plan. Each of the previous lessons produced one: an emission inventory, a dispersion screening, a particulate and VOC control scheme, boiler burner choices, a wastewater pretreatment design, a water reuse scheme, a waste inventory, a life-cycle assessment and a carbon account. Now they must become one document that a board can approve, a regulator can review and an operator can run. This lesson follows how Wren builds that document, the analysis behind its carbon section, and the board meeting where it is decided.

The structure of the plan

Wren uses a structure she has refined over several permit applications. The plan opens with a two-page summary for the board: what the expansion changes, what it will take to comply, what it will cost, what carbon reduction it can deliver and what decisions are needed now. The body has two parts. The compliance part goes medium by medium, air, water, waste, and for each lists the sources, the controls, the limits the plant will request, the monitoring, and the permits or approvals needed with their timing. The carbon part sets out the carbon accounts, the options, their costs and their effect on carbon intensity, and a staged pathway. Appendices hold the inventory, the modelling, the life-cycle assessment and the calculations, each with its basis and its author.

Part 1: compliance

The air section begins with the inventory from Lesson 1. With all the controls chosen in Lessons 3 to 5, the expanded plant's potential to emit is about 104 t of nitrogen oxides, 91 t of carbon monoxide, 59 t of VOC, 43 t of sulfur dioxide, 42 t of PM10 and 28 t of PM2.5 a year.

In Ontario, the plant applies to amend its ECA with an updated ESDM report under O. Reg. 419/05, covering every source, and an acoustic assessment for the new fans and compressors. The screening in Lesson 2 shows the RTO stack far below the point-of-impingement standards; the scrubber stack is acceptable only with bisulfite, so the bisulfite system, its feed rate and the scrubber water flow become permit conditions. The new boilers meet the federal boiler regulations with ultra-low-NOx burners. The plant continues to report to the NPRI and the federal greenhouse gas program, and to the EPS.

The water section sets out the equalization tank, the DAF and the EGSB reactor of Lessons 6 and 7, with the biogas going to the boilers. The plant will negotiate a revised agreement with the city covering the new discharge, which will be steadier, near neutral and below the surcharge strength. The water reuse section commits to holding city water at today's level through the expansion with RO on the cooling tower blowdown and more condensate returned to the cook, a water intensity of about 3.0 L/L. The waste section lists the new streams, their classification and their routes, and commits to registering them before start-up.

Each item has a permit, an owner and a date. The permit amendment is the critical path: Wren plans to file the application within six weeks of the board's approval, using vendor guarantees for the new equipment and conservative estimates elsewhere, and to start construction of the affected units only when the approval is in hand. Ingrid Solheim, as a P.Eng., will seal the engineering in the application.

Part 2: carbon

The carbon section starts from the accounts of Lesson 11: about 100 kt a year of scope 1, scope 2 that depends on the grid, about 180 kt of scope 3 from corn, and about 180 kt of biogenic fermentation carbon dioxide. Its question is which reductions to make, in what order, and at what cost. The standard tool is a marginal abatement cost curve. For each option, the annual cost is the capital cost annualized with a capital recovery factor, plus operating cost, minus savings in fuel or power; the cost per tonne is that annual cost divided by the tonnes avoided each year. The options are then sorted from cheapest to dearest, and plotted as bars whose width is the tonnes avoided and whose height is the cost per tonne.

# Capstone: marginal abatement cost of the carbon options for the expanded plant (Ontario values)
GRID, GAS, POWER = 40.0, 5.0, 0.10     # g/kWh, gas price per GJ, power price per kWh (illustrative)
CRF = 0.10 / (1 - 1.10 ** -15)         # 10 %, 15 years
opts = [  # name, capital (millions), fixed O&M (millions/yr), gas saved TJ/yr, power added GWh/yr, biogenic CO2 stored kt/yr
 ('RTO heat recovery 90 -> 97 %',      3.0, 0.05, 118,   0.0,   0),
 ('Biogas to boiler',                  1.5, 0.05,   6,   0.0,   0),
 ('Evaporator and still MVR',         18.0, 0.30, 288,  12.0,   0),
 ('Fermentation CO2 capture, storage', 60.0, 6.30,  0,  18.9, 172),
]
rows = []
for n, cap, om, gas, pw, st in opts:
    fossil = gas * 56.0 / 1000 - pw * GRID / 1000          # kt fossil CO2 avoided
    abated = fossil + st
    cost = CRF * cap + om - gas * GAS / 1000 + pw * POWER  # millions per year
    rows.append((cost * 1000 / abated, n, abated, fossil, cost))
print(f"capital recovery factor {CRF:.4f}")
for c, n, a, f, cost in sorted(rows):
    print(f"{n:35s}{a:6.1f} kt/yr  {c:7.0f} per t  (fossil {f:5.1f} kt)")
# incentives: EPS price on fossil CO2 and Clean Fuel Regulations credits on every tonne of CI reduction
EPS, CFR = 110, 150
for c, n, a, f, cost in sorted(rows):
    net = cost - f * EPS / 1000 - a * CFR / 1000
    print(f"{n:35s} net after EPS and CFR credits {net:6.2f} million per year")
# carbon intensity pathway (g CO2e/MJ), starting from 56.0 with land-use change
LHV = 21.2 * 240e6; ci = 56.0
for c, n, a, f, cost in sorted(rows, key=lambda r: r[1] != 'RTO heat recovery 90 -> 97 %'):
    ci -= a * 1e9 / LHV - (1.5 if a > 100 else 0)          # capture also pays 1.5 g/MJ for transport and storage
    print(f"after {n:35s} CI {ci:5.1f} g/MJ")

capital recovery factor 0.1315
RTO heat recovery 90 -> 97 %          6.6 kt/yr      -22 per t  (fossil   6.6 kt)
Fermentation CO2 capture, storage   171.2 kt/yr       94 per t  (fossil  -0.8 kt)
Evaporator and still MVR             15.6 kt/yr      155 per t  (fossil  15.6 kt)
Biogas to boiler                      0.3 kt/yr      646 per t  (fossil   0.3 kt)
RTO heat recovery 90 -> 97 %        net after EPS and CFR credits  -1.86 million per year
Fermentation CO2 capture, storage   net after EPS and CFR credits  -9.53 million per year
Evaporator and still MVR            net after EPS and CFR credits  -1.64 million per year
Biogas to boiler                    net after EPS and CFR credits   0.13 million per year
after RTO heat recovery 90 -> 97 %        CI  54.7 g/MJ
after Biogas to boiler                    CI  54.6 g/MJ
after Evaporator and still MVR            CI  51.6 g/MJ
after Fermentation CO2 capture, storage   CI  19.4 g/MJ

The curve has a clear shape. Restoring and upgrading the RTO heat recovery costs less than nothing: the gas it saves pays for it. Capture and storage of the fermentation carbon dioxide is by far the largest option, at a cost of about 95 per tonne before any credit. Vapour recompression on the evaporators and stills is smaller and dearer per tonne, because it trades gas for electricity and its capital is large for the tonnes it saves. Burning biogas is very expensive per tonne of carbon, but that is the wrong way to see it: the anaerobic reactor is justified by the wastewater, and the biogas is a by-product whose use is almost free once the reactor exists.

In Ontario, the credits change the picture. The EPS charges the carbon price on fossil emissions above the plant's allocation, so every fossil tonne avoided saves about 110 dollars; the Clean Fuel Regulations reward every tonne of life-cycle reduction, fossil or biogenic. After both, the RTO upgrade, MVR and capture all have a net benefit, and capture's is the largest by far, about 9.5 million dollars a year on the illustrative credit price. The carbon intensity pathway falls from about 56 g/MJ to about 52 with the plant-side measures and to about 19 with capture.

The staged pathway

Wren's recommendation is a pathway in three stages, ordered by certainty and cost. The first stage is built with the expansion and is not optional: the compliance measures, plus every carbon measure that pays for itself or is needed anyway, namely the RTO upgrade, the ultra-low-NOx boilers, the anaerobic reactor with biogas to the boilers, the RO water reuse and the vents to the RTO. The second stage is MVR and heat integration on the evaporators and stills, designed into the expansion so it can be added within two years, with its final approval tied to gas and power prices and to the value of the carbon intensity reduction. The third stage is capture and storage of the fermentation carbon dioxide. It is the largest and most valuable, and also the most uncertain, because it depends on infrastructure outside the plant's control: a storage licence in southwestern Ontario. The plan recommends spending now on what makes capture possible later: space in the plot plan for the compressor building, a tie-in on the scrubber outlet, and a front-end engineering study, with a decision gate when the storage route is firm.

Theo, writing the risk section, lists what could change the answer. Credit prices and tax law can move by half in either direction. Storage could be delayed by years or could arrive sooner. Gas prices move the MVR economics. A new rule could tighten the plant's NOx or PM2.5 limits. Each risk has a trigger, an owner and a response in the plan. Nadia reviews the safety implications of each option; Priya reviews the occupational ones; Dev reviews the sampling and analysis the monitoring needs; and Malik reviews the money.

Presenting to the board

The board meeting in June gives the plan forty minutes. Wren speaks for ten, using four slides: what the expanded plant will emit and how it will comply; what it will cost; the carbon pathway as a single line from today's carbon intensity to the target; and the decisions needed. She does not show the abatement curve until a director asks why capture is not in the first stage, and then uses it to show that capture is the cheapest large option per tonne, and the riskiest. Malik covers the money: the compliance measures add a few percent to the expansion capital and remove the sewer surcharge; the first-stage carbon measures pay for themselves; capture is a separate investment decision for later.

The questions are pointed. One director asks whether the plant could skip the RTO enlargement and run the dryers cooler instead. Wren answers that cooler drying is part of the plan but cannot alone meet the VOC limit for three dryer sections. Another asks whether zero liquid discharge should be reconsidered; Wren refers him to the water study and its conclusion. The investor who asked about water asks whether the carbon intensity claims will be audited; Wren confirms that the life-cycle assessment will have an independent critical review and that the carbon intensity is verified for the fuel program each year. The chair asks the question Wren expected: what happens if the permit takes longer than planned? Ingrid answers that the schedule allows nine months for the permit, that the application will be filed early with conservative estimates, and that the units that do not affect emissions can be built first.

The decision

The board approves the expansion with the plan as a condition of the funding. It approves stage 1 in full, approves the design provisions for stages 2 and 3, and asks for a decision paper on capture when the storage route is firm. It asks for an annual report on compliance, water intensity and carbon intensity against the plan's targets. Wren files the permit application three weeks later.

Theo's last task is the one Wren insists on for every project: a short note for the file of what the team learned. He writes four points. Environmental engineering is process engineering, and the mass balance is the most important tool. Prevention and recovery beat treatment, and they usually also cost less. Compliance and carbon are now two separate problems with two separate logics, one binary and enforced, the other continuous and priced, and a plan has to address both. And every number needs a source, because someone will ask.