Post-Disaster Housing: Building a Premium Community from Prefabricated Container House Modules

A post-disaster community built from prefabricated container house modules is not a camp with better tents. It is permanent housing with a fixed address, a wall thickness, a window that opens and a building permit — and the difference between the two is decided in a factory drawing months before anything reaches the site.

Reconstruction after a disaster is usually described as an emergency problem. It is not. The emergency phase lasts weeks and is solved with tarpaulins, tents and whatever halls are standing. The problem that lasts for years is the third phase: permanent housing, at volume, to a standard a family will accept as a home rather than as an interim arrangement. That phase is a manufacturing problem, and it is the phase this article is about.

What follows is a design and procurement study, not a record of a delivered contract. It works through a two-storey coastal lodge built from the same module system, taken from a single unit standing in a factory yard to a completed block with an external gallery, glazed living space and a service strategy that survives an inspection. Along the way it answers the questions that come up in every reconstruction tender — how fast a community can be delivered, what separates a house that lasts thirty years from one that is patched from year six, which specification items a buyer can actually check in a sample, and where a container house stops being the right answer.

Quick answers for buyers preparing a reconstruction programme

What is the difference between transitional and permanent housing? Transitional housing is designed to be occupied for one to three years and is usually a single-skin unit. Permanent housing is designed for a service life measured in decades, with insulation, a proper envelope, opening windows and a fire strategy. They look similar in a photograph and are not the same product.

What size is one module? The flat pack container house module is 6000 mm long, 3000 mm wide and 2840 mm high — 18 m² of floor. Two modules make a one-bedroom home with a kitchen and a bathroom; four make a family house.

How many modules ship in one container? Eighteen flat pack sets fit in a single 40HQ, because the frame, panels and roof travel knocked down and nested rather than as a finished box. That number, not the unit price, is what decides the landed cost of a reconstruction programme.

How fast can a community be occupied? The structure and envelope are weeks of site work per block, not months. Foundations, utilities trenching, access roads and the statutory approvals are the long items, and they run in parallel with production rather than after it.

Can it go two storeys? Yes — two storeys is the standard configuration for this system, and the block in these images is two storeys. Anything above that changes the frame, the foundation and the lifting plan and has to be engineered for the site.

Is a container house cheaper than local construction? Not automatically per square metre. It is usually cheaper in programme, in site labour and in variance, which on a reconstruction programme is where the money actually leaks.

Who builds it? Global Container (Suzhou) Co., Ltd., Zhenze Town, Wujiang District, Suzhou, Jiangsu, China. Frames, panels and fit-out are produced on our own lines with no subcontracting, and factory inspection is welcome before shipment.

Three phases, three different products

The single most expensive mistake in a reconstruction programme is buying phase-one thinking at phase-three volume. Emergency shelter is a logistics exercise: light, cheap, fast, and nobody expects it to be a house. Transitional housing is a bridge: one to three years, minimum specification, thin envelope. Permanent housing is a building. It has to satisfy a code, carry a mortgage or a title, survive a rainy season without being serviced, and be maintainable by an owner who has no crane and no welder.

Modular construction belongs to the third phase, and that is where it is genuinely strong. A factory can produce the same wall panel for the four-hundredth time with the same thickness, the same coating weight and the same joint detail as the first. Site construction cannot do that, because site construction is done by whoever is available, in weather, at speed, after the local supply chain has been disrupted. Repeatability is the whole product.

The consequence for a buyer is that the specification has to be written before the tender, not after. A programme that specifies "container house, 20ft, white" will receive fifteen quotes that are not the same product at all, and the cheapest one will be structurally different in ways that are invisible in a photograph and expensive in year six. The next section is about that gap.

The prototype is the argument

The photograph below is not a render. It is a standard flat pack container house module standing on the concrete apron of our yard in Suzhou, photographed in ordinary daylight with no styling. It is the physical object that every drawing in this article is a variation of: a welded steel frame with ISO corner fittings, sandwich panel walls, a window with a security grille, a steel door with a reinforced lock set, and a lifting point at each corner.

This is the honest starting point for a reconstruction discussion, because everything a buyer is being asked to pay for is inside it. The frame sections are 2.2 mm, 2.45 mm, 2.75 mm or 3.0 mm true thickness, with 4.0 mm available on request. The corner blocks — the weakest point of any container for both load transfer and corrosion — are 4.0 mm galvanised on the flat pack and 3.5 mm to 4.0 mm on the folding version. The galvanised layer is 60 g double-sided hot-dip as standard and can be built up to 300 g for coastal, high-humidity or salt-spray sites. None of those figures appear in a product photograph, and all of them decide whether the unit is still square and sound in year fifteen.

The same unit is what arrives on a reconstruction site, except knocked down. The frame, the wall panels, the roof sheet, the ceiling sheet, the flooring boards, the door set, the window set, the trim and the hardware are packed as a kit, wrapped, braced and loaded. Eighteen sets go into one 40HQ. That is the number that makes a programme arithmetic instead of optimism: a thousand homes is fifty-six containers of kit, and it can be scheduled against vessel space rather than against the availability of local labour.

A community is not a row of boxes

Houses that are individually adequate can still produce a place nobody wants to live in. The failures are consistent and they are all in the layout: a grid with no streets, no shared space, no shading, and front doors that all face the same direction so that the settlement reads as storage rather than as a neighbourhood.

Four decisions do most of the work, and all four are cheaper to make on paper than on site.

• Cluster rather than grid. Groups of six to twelve homes around a shared court, with pedestrian routes between clusters. A grid at the density a container system allows produces a wind tunnel and a lot of unowned space that will become a dump.
• Orientation before repetition. Long elevations facing the equator get shade; long elevations facing the weather get rain. Rotating a cluster thirty degrees costs nothing in production and changes the thermal load of every unit in it.
• One community building, built once, from the same kit. A clinic, a shop, a schoolroom or a meeting hall assembled from the same modules as the houses. It is the cheapest social infrastructure available, because it uses the same drawings, the same crew and the same container space.
• Leave the infrastructure corridor wide enough for a vehicle. A service lane designed for a handcart cannot be used by a tanker, a fire appliance or an ambulance afterwards. Retrofitting width into a built cluster is not a design change, it is a demolition.

The block in these images is a two-storey version of the same logic at building scale. It is single-loaded: every room opens onto an external gallery with a door and a window, and the gallery doubles as circulation, balcony, drying space and escape route. A double-loaded internal corridor would have carried more floor area on the same footprint, but it would also have needed a lit, ventilated, fire-rated corridor — an entire building inside the building, with its own services and its own failure modes. On a reconstruction site where maintenance capacity is the scarcest resource of all, the arrangement that needs the least maintenance wins.

The specification choices that decide service life

Buyers comparing reconstruction quotes usually compare a price per unit, which is the least informative number on the table. Four items drive the real figure, and each of them is checkable without a laboratory.

Steel and its coating. Frame thickness and zinc weight together decide how long the structure survives. A 40 g zinc layer against our 60 g standard is a 33 per cent reduction in sacrificial protection, and on a coastal site that is the difference between a frame that is still sound at year twenty and one that is being painted at year seven.

The flooring board. This is where low-cost configurations hide their worst defect. Grade B glass-magnesium boards are prone to salt migration, which corrodes the frame from the inside and cannot be reversed. We use Changzhou Bulu Grade A boards and cement boards from Changzhou Bulu and Shandong Beibanqiu, all asbestos-free. The buyer who specifies "18 mm floor" without specifying the grade has specified nothing.

Insulation density. Our glass wool is 11 kg/m³, where 8.8 kg/m³ is the market norm. Denser wool does not sag, holds its thermal performance for longer and deadens sound. It also makes a house feel solid, which matters more than buyers expect in a programme where residents did not choose the product.

Hardware. Doors with a 0.25 mm skin, ordinary self-tapping screws and untreated bolts are the parts that fail first and cost most to fix in a place with no hardware shop. Our door skin is 0.4 mm with three stainless-steel hinges; carriage bolts are Dacromet-treated; flat pack hex-socket and bottom-tube joint screws are thick-galvanised and every other self-tapping screw is stainless.

The last row is the one to argue about in a tender meeting, because it is the only row that can be verified with a weighbridge. Two units of the same nominal size, one at 1.3 t and one at 0.85 t, are not the same product. The extra 450 kg is profile thickness, zinc, floor grade and hardware. It is also the reason a price comparison alone cannot tell a buyer which of the two quotes is cheaper.

Salt, dust and heat: the same drawing, three different envelopes

The block shown here sits behind a beach. That single fact rewrites part of the specification: 300 g zinc instead of 60 g, stainless screws throughout, and the floor built up off the ground on a continuous apron so that salt-laden ground water never reaches the base rail. The client pays perhaps three per cent more for the envelope and buys a decade of service life.

The same module goes to sites with completely different dominant loads, and the specification has to follow the site rather than the catalogue. In a hot-arid context the enemy is not the average temperature but solar gain through glass and dust ingress through every opening, and the answer is external shade, a smaller glazed ratio, and a sealed and filtered supply of ventilation air rather than more air conditioning. In a savannah context the dominant loads are water, insects and access: storage sized for the dry season, every opening screened, and a floor level set high enough that a vehicle can reach the building in the wet season without a four-wheel drive.

That comparison is set out in detail separately, including what changes in the panel build-up and what does not change at all. The point to carry into a tender is that a modular system is not a fixed product: the frame is a constant, and the envelope, the glazing, the shading and the fixings are variables that a factory can set per project.

What a factory can promise, and what it cannot

This is the part of the study that has to be stated plainly, because it is where suppliers mislead buyers and buyers mislead themselves.

A factory in Suzhou can guarantee that the frame it ships is the thickness on the drawing, that the zinc weight is the weight on the test report, that the floor board is the grade on the certificate, and that the unit it loaded matches the approved drawing. It can open the doors for an inspection, provide material samples and test documents, and produce the same unit five hundred times.

It cannot certify a building against a code it does not administer. Structural calculations for a specific site, the fire strategy, means of escape, accessibility, foundation design, utility connections and the planning permission itself belong to the buyer's engineer and the local authority. A supplier who offers to handle all of that for a jurisdiction they have never built in is telling you something about their sales process, not their engineering.

The practical form of that boundary is a document set. A serious order runs on an approved drawing, a material schedule, a loading plan and a packing list. If a supplier cannot produce those four before taking a deposit, the rest of the conversation is a price negotiation with no technical floor under it.

Cost and programme: where the saving actually is

A reconstruction programme does not fail on unit price. It fails on variance: the school that was finished eight months late because roof trusses arrived twice wrong, the clinic with a leaking roof at handover, the street that had to be re-dug because the drainage was designed after the houses. Modular construction attacks those, and the arithmetic is unglamorous.

• Labour moves from site to plant. Site labour is the item that is scarcest, least predictable and most expensive after a disaster. Factory labour is scheduled, supervised and repeatable.
• Programme compresses because phases overlap. Foundations are cast while units are being produced, not after. That is weeks saved on every block.
• Freight is a design variable. Eighteen sets per 40HQ is a design outcome, not a shipping accident. Every millimetre of packed height that can be removed is money off every unit in the programme.
• Maintenance drops. A galvanised corner block and a stainless screw cost more on day one and nothing for a decade, which in a donor-funded programme with a fixed defect-and-liability window is the cheapest money on the table.

What does not come for free is the design work. A modular community needs a layout resolved before production, and a client who cannot make decisions early will pay for it twice: once in the drawing office and again in a site full of units that had to be adapted afterwards.

A checklist for the buyer writing the tender

1. State the phase. Emergency, transitional or permanent — in writing, in the brief. Everything downstream follows from it.
2. Specify the frame by true thickness and the coating by zinc weight in grams, not by "heavy duty".
3. Specify the floor board by grade and manufacturer, not by thickness alone.
4. Ask for the unit weight, then verify one unit on a weighbridge at pre-shipment inspection.
5. Require an approved drawing, a material schedule, a loading plan and a packing list before the deposit.
6. Fix the foundation type and the floor level against a flood level or a wet-season access requirement, whichever is the higher constraint.
7. Decide who is responsible for the structural calculation, the fire strategy and the permits, and write the names into the contract.
8. Reserve the maintenance question: who supplies the spare door hinges, screws and window seals in year three, and at what price.

Answers to the questions this raises most often

A finished building is the only argument that survives a site visit

Everything above is a specification, and specifications are easy to write and hard to prove. What settles a reconstruction decision is a building that a buyer can walk into: a door that closes squarely, a window that opens, a floor that does not flex underfoot, a wall that is straight along eight metres, and a corner joint where the paint has not begun to blister after a season of salt air.

The block in these photographs is a two-storey coastal lodge produced from the same module family, in the same factory, on the same lines as the unit standing on the yard apron. It is not a flagship product built for a photograph. It is what the system produces when the client specifies properly and the factory is held to the drawing — which is the whole of the argument for modular construction in reconstruction, and the reason the specification is worth more attention than the price.

Global Container (Suzhou) Co., Ltd. builds prefabricated container house, flat pack container house and detachable container house systems from Zhenze Town, Wujiang District, Suzhou, Jiangsu, China, and has exported to more than thirty countries since 2012. Configuration lists, material samples and test documents are available on request, and a tailored configuration can be prepared for a specific project.