Modular Waste Processing Sells a Schedule the Grid Won't Honor

Sumitomo SHI FW publishes four months from delivery to erected plant on its modular line, and fifteen to eighteen months from start to commercial operation. I believe both numbers. I've stood next to a shop-built line that went from flatbed to hot commissioning inside a quarter, and it's a good piece of engineering. But modular waste processing doesn't control the clock that decides whether the plant earns anything, and on the projects I've developed that clock was never the module's.

Two things set it instead. One is a utility interconnection study. The other is a single number buried in 40 CFR Part 60 that most developers don't read until phase two is already quoted.

Four months to erect, five years to plug in

Berkeley Lab has been tracking how long generators actually wait, and the trend is ugly. Median time from interconnection request to commercial operation has roughly doubled: under two years for plants built between 2000 and 2007, more than four years for those built 2018 through 2024, with a typical 2023 project spending close to five years in the pipeline (per Berkeley Lab's Queued Up series). FERC Order 2023 moved the process to first-ready, first-served clustering back in July 2023, and I'd like to tell you it's fixed. Too early to say.

So the brochure schedule and the project schedule are different animals. Erect in four months, then hold a queue position for four years, and you haven't built a fast project. You've built a warehouse that happens to contain a boiler.

And the module's own schedule has soft spots that are rarely engineering. In 2021 we retrofitted a transfer station in Panama and lost five weeks to customs, who held the sorting line on the dock over a documentation dispute. Five weeks. That missed the commissioning window, the crew we'd flown in went home, and the rebooked slot put us into a season we didn't want. Nothing about the equipment was late. The paperwork was.

The year after, on a Caribbean island RDF facility, we pre-positioned spares months before anyone needed them and came in ahead of schedule while running about nine percent over budget. I'd take that trade every time. Regional market entry fails on logistics and offtake credit, not on technology choice, and going modular doesn't change that arithmetic. It concentrates it. Instead of a two-year site build with slack in it, you have a shipping problem with a hard date attached, which is how most of the waste conversion facilities I've worked on actually got into trouble.

Nobody ever adds the second module

The pitch for a scalable waste solution never varies: start with one line, add another when the town grows. I've seen it in every feasibility deck I've touched. I have not once seen the second line built on the original plan's timeline, and the reason isn't only municipal budget drift.

Subpart AAAA of 40 CFR Part 60 covers new municipal waste combustion units with the capacity to burn at least 35 tons per day but no more than 250. A single Sumitomo line, running five to seven tonnes an hour, sits comfortably inside that band. Fine. But the subpart splits those units into two classes, and the split isn't drawn per unit. It's drawn per plant. Class II means the plant's aggregate combustion capacity is 250 tons a day or less. Class I means it's more. Same subpart, different emission limits, different continuous-monitoring obligations.

Read the ordering there carefully, because it's the whole problem. Your first module is Class II. It was permitted as Class II, its CEMS was specified as Class II, its stack-test protocol was written for Class II. Then you park module two next to it, aggregate capacity crosses 250 tons a day, and module one becomes a Class I unit. You didn't touch it. You didn't modify it. It got reclassified because of something you set down beside it.

The second module doesn't just get permitted on its own merits. It re-opens the permit on the first one.

That's what kills the phase-two option in practice. At the top of the vendor's own published range, a pair of lines lands you north of 250 tons a day aggregate, on exactly the wrong side of the threshold. So you either size the pair deliberately under the cliff and surrender the capacity you were expanding for, or you accept Class I on both and re-engineer a plant that's been running fine for four years. A permit in hand beats a signed MOU every time, and a Class II permit in hand is worth defending. Most operators, once they've run it properly, defend it. The second module stays in the deck.

Actually, let me qualify that, because it's a US federal cliff. On the projects I've run in the Caribbean and Central America the binding constraint was almost always the grid connection or the offtake counterparty, and nobody had heard of Class II. But if you're developing stateside and your deck has a phase two in it, 250 is the number that deck is quietly betting against. What's the module for, then, if it isn't for growing into?

The spec sheet already told you what to sell

Look at what Sumitomo SHI FW publishes for one combustion line converting 30,000 to 50,000 tonnes a year. Three options, same fuel, same hardware:

3.7 MW of electricity, 2 MW of electricity and 10 MW of heat, or 17 t/h of steam at 400 C and 40 bar.

Take the middle one and you give up most of your electrical output to gain roughly five times as much energy as heat. That's before you count what either product is worth, or what each one costs to sell.

Electricity means an interconnection request. Heat means a pipe. One of those puts you in a Berkeley-Lab-sized queue behind every solar-plus-storage project in the region; the other puts you in a bilateral negotiation with a dairy, a hospital laundry, a greenhouse, a district loop, a lumber dryer. I have never seen a steam pipe require a system impact study.

So the case for modular I'd make to a board isn't the cheap one. Per megawatt it's expensive: the module runs maybe half of installed cost, and quotes land somewhere around four to ten million dollars a megawatt across the size range, with small units sitting at the wrong end of that spread [industry estimate]. It doesn't scale, for the reasons above. What it does is fit. A modular waste plant is small enough, and light enough on civil works, to sit inside a fence line next to somebody who already buys steam, which is how you get renewable energy from waste onto a revenue line without asking a utility for permission. Demand was never the constraint anyway: EPA still books about half the national municipal stream into landfills each year against under twelve percent to combustion with energy recovery.

Which drags the whole thing back to feedstock, because a modular line's tolerance window is published and it's narrow: seven to twenty-two MJ/kg, moisture up to 55 percent. A big mass-burn plant absorbs a bad week. One 130-tonne-a-day line doesn't, and I've argued before that a waste-to-energy solution is only as good as the feedstock nobody measured. On a single modular unit that stops being a philosophical point. It's your only line.

Where this doesn't hold

No thermal host inside pipe distance and the argument collapses, because you're back in the queue with a plant that costs more per megawatt than the one you'd have built at scale. Portable waste to energy has a genuine niche in remote camps, islands, and disaster response, but it's narrower than the brochures suggest and it usually runs on diesel-parity economics rather than gate fees.

The single host cuts both ways, too. One counterparty is one credit exposure, and a dairy that closes in year nine takes your only offtake with it. I'd rather sell to a hospital campus than a food processor for exactly that reason, and I'd rather sell to a district loop than either. Seasonal load is the other trap: a greenhouse buying heat eight months a year leaves you four months of a plant you still have to staff, and a licensed operator costs the same at 130 tonnes a day as at 900.

Two more places this reasoning doesn't reach. If you're buying a brownfield with a live interconnect, you're buying a queue position, and the whole calculation inverts. And not every modular unit is a grate: modular pyrolysis systems exist, and the Subpart AAAA analysis above doesn't map onto them cleanly, which is its own long argument.

The site visit tells you what the data room hides. The data room says there's a hospital 1.2 km away with a year-round steam load. The site visit tells you those 1.2 km cross a rail easement and a river, and that the hospital replaced its boiler last spring and has nineteen good years left on it.

Disclosure: I develop projects for Renewable Waste Energy, which has commercial interests in both modular and conventional builds. If you want this argument stress-tested against a real site, contact RWE, though the site visit is where it survives or doesn't.

Four months, delivery to erected. I've watched it happen and it's worth watching. I've also watched a finished plant sit behind a locked fence for the better part of two years waiting on a utility study, which is an expensive way to store a container.

Sources & Notes

• Erection and commercial-operation timelines, line capacities, the fuel window, and the electricity-versus-heat options come straight off Sumitomo SHI FW's modular waste-to-energy page. That's their published range, not a project of mine.
• Queue durations belong to Berkeley Lab, from the Queued Up series, which tracks interconnection requests through to commercial operation across US regions. The doubling since the mid-2000s is their finding. The sarcasm about it is mine.
• The Class I and Class II split lives at 40 CFR 60.1045. The 35-to-250-ton applicability band sits one door down, in the subpart's coverage section. Read both before you quote anyone a phase two.
• National tonnage context is EPA's, from Facts and Figures about Materials, Waste and Recycling. The 2018 data year is still the latest full accounting EPA publishes, which tells you something on its own.
• Panama, the Caribbean island facility, and the budget overrun are out of my own project files. The dates are exact. I've kept the counterparties out of it.