A Cabin Near Black River Falls Wisconsin: Performance and Logistics Intelligence for Direct-Print Walls
The cabin and the supply-chain gap
This is a sample case. Picture a small cabin near Black River Falls, in Jackson County, in west-central Wisconsin. The owner wants printed walls. The cabin is the kind of regional, owner-financed residential build that the additive construction industry’s central supply chain does not economically reach. The nearest commercial 3DCP-grade bagged-mix terminals are several states away. There is no other regional 3DCP service operating at this scale near Black River Falls. The owner could either build conventionally, with a wood or block shell at the local market’s rate and weather window, or go to a service that closes the freight and operating gap regionally.
That latter option is what this case is about. Sunnyday Technologies is a software, design, and data company. We do not run print jobs. We make the formulation, the sourcing map, the QC plan, and the cost model. The CEMFORGE prediction engine (cemforge.ai) writes the mix design. The customer’s engineer of record signs off on the structure and the code path. A separate 3DCP contractor does the actual printing. Three roles. The deliverable is the data each role needs to do its job.
The result, at the P0 confirmation stage, is a packet of performance and logistics intelligence that turns a vague “we want printed walls” inquiry into a printable spec a contractor can quote against and an engineer of record can sign off on.
This case is about that packet, and about the buyer the engagement serves: residential and small-build owners whose project does not fit the central 3DCP supply chain.
What the engagement actually delivered
CEMFORGE produces a structured deliverable; the structure itself is a Sunnyday work product and the company does not publish it in full. What follows is a summary of the highlights this engagement produced, written to the level of detail appropriate for a public case study.
The deliverable opens with an explicit Engineer-of-Record handoff. The packet’s first page identifies the deliverable as performance and logistics intelligence from Sunnyday Technologies, and names the licensed engineer of record on the project as the responsible party for code applicability and final mix-design approval. Sign-off is required before any structural pour. This framing is load-bearing for how the engagement is scoped: Sunnyday delivers the prediction layer, the sourcing layer, the QC plan, and the test-program logistics; the EOR delivers the structural sizing, the code submittal, and the final mix-design approval. The two are complementary, and neither replaces the other.
Beyond the EOR header, the deliverable is built around four artifacts: a recommended composition window with Monte Carlo robustness analysis, a sensitivity-driven QC plan, a regional sourcing map, and a project test schedule sized to the print application.
The recommended composition window
The recommended formulation is a single robust window rather than a set of alternative candidates. The chemistry uses portland-limestone cement (Type IL) as the primary binder, paired with a supplementary cementitious material chosen for late-age strength contribution under freeze-thaw. The fresh-state package includes a viscosity modifier and a high-range water reducer dosed to the target slump-flow. The aggregate is concrete sand at the gradation the project’s print parameters require. The specific supplementary materials in any individual formulation are not categorical for 3DCP at this scale; they are what the prediction surface settled on for this project’s strength target, this project’s printability window, and the supply chains the project can actually procure from. A different cabin in a different geography would land on a different window.
The performance target is a high-strength wall mix conditioned for the severe-freeze-thaw exposure profile of Climate Zone 6A residential. The Monte Carlo robustness analysis, run at one thousand draws under symmetric ±10% supply variation across the six composition variables, indicates that the design floor is not breached anywhere in the sampled envelope. The hard-floor margin is intentionally conservative: the model places the per-cylinder reject threshold materially below the design target so that compounded supply error has somewhere to go before quality is at risk. That margin is what makes the window deployable on real plants with real batching tolerances rather than only on the laboratory bench.
The window itself is the operating envelope. Inside the window the prediction surface is well-behaved; outside the window the prediction layer’s confidence drops, not because the chemistry stops working, but because the model’s training-coverage thins. The deliverable is honest about that boundary. It says, for each of the six composition variables, what the lower bound is, what the upper bound is, and what the dominant failure mode is on each side.
The sensitivity plan and what it implies for QC
The sensitivity analysis identifies the variables that move the strength outcome the most. The dominant negative driver is water dose: increases in water reduce the predicted strength more than any other axis at the unit-percent variation, which is a chemistry-of-concrete result the bench engineer would expect and which the prediction surface confirms. The dominant positive driver is the portland-limestone cement dose, with about three-quarters the magnitude of the water effect in the opposite direction. The supplementary cementitious material contributes a smaller positive effect on late-age strength. The remaining axes (the viscosity modifier, the superplasticizer, and the sand) carry small sensitivities on strength, which does not mean those variables are unimportant; it means their effects are dominated by fresh-state behavior (viscosity, slump-flow, layer adhesion) rather than by hardened compressive performance.
The QC plan is sensitivity-driven. Water and cement are tightened to ±5% of target with hard rejection rules below the composition window’s lower bound. The remaining four variables are loose at ±10% of target with reasonable QC discipline. The result is a QC plan that asks the plant to be tight where it has to be and lets it be loose where it can. This is the difference between a QC plan that asks a regional plant to operate as a research lab and a QC plan that respects how regional plants actually batch.
The QC test schedule pairs with the sensitivity plan: every batch tracks fresh density and slump-flow against design tolerances, with superplasticizer (not water) used to recover slump-flow if it drifts. Compressive cylinders are cast at every fifty cubic meters or daily, whichever comes first, with the seven-day and twenty-eight-day pair sized for QC verification of CEMFORGE’s own predictions. Aggregate moisture is calibrated weekly. Cement deliveries carry mill certificates, which the plant’s QC function reviews and which trigger load rejection if any parameter is outside supplier spec.
Sourcing
The sourcing map identifies more than a thousand suppliers within the project’s five-hundred-kilometre radius across Wisconsin and the Upper Midwest. The map prioritizes the primary supply edges. Concrete sand for the formulation comes from a regional aggregate producer in Outagamie County, Wisconsin. Portland-limestone cement comes from a regional cement plant in northern Illinois (Illinois Cement Company at LaSalle Quarry and Mill), which serves the broader Upper Midwest market through established distribution. The specialty admixtures (the supplementary cementitious material, the viscosity modifier, the superplasticizer, and a conditional air-entraining admixture if the trial pour requires it) ship through national distribution channels. None of these are exotic supply paths. The sourcing detail in the deliverable is about which specific producers, at which freight distances, and with which quality-consistency notes the contractor’s QC function should anchor on.
The cabin is planned as an on-site print. The contractor brings the printer to the cabin site and prints the wall set there over the project schedule. Materials are delivered to the site from the regional supply edges in the sourcing map. There is no precast-and-truck step in the project shape, which keeps the logistics simple and aligns the on-site print time with the cabin’s weather window in west-central Wisconsin.
The cost of the recommended formulation is meaningfully higher per cubic metre than conventional ready-mix at the industry-baseline strength target. The deliverable says so plainly. It also notes that the higher unit performance gives the engineer of record room to size the wall thinner or to reduce reinforcement, which can offset the per-cubic-metre premium at the project level. That tradeoff is the EOR’s call, not Sunnyday’s, and the deliverable presents the cost line item without prescribing the structural offset.
What the Engineer of Record does next
The deliverable is the start of a multi-phase engagement, not the end of one. The next phase is the EOR’s review against the project’s structural specification and the applicable code path, which for residential walls in Jackson County means the Wisconsin Uniform Dwelling Code (Wisconsin Administrative Code chapters SPS 320 through SPS 325, with construction standards in SPS 321) plus whatever ACI and ASTM references the EOR brings to bear on a high-performance printed wall section. Sunnyday does not do that work, does not deliver a code-equivalence packet, and does not represent the printed wall as conforming to any specific code section. The EOR sizes wall thickness, sets form-strip and any post-tension thresholds against the project’s structural requirements, and signs off on the final mix-design approval. The CEMFORGE deliverable is the data input to that engineering judgement, structured to make the EOR’s job tractable rather than to substitute for it.
The EOR’s review checklist in the deliverable’s first page is explicit. The EOR confirms the strength target meets the structural requirement, verifies supplier lots match the composition window, approves QC plan tolerances and the test schedule, sets project-specific form-strip and post-tension thresholds, and stamps the packet for the project record. None of those steps are Sunnyday’s call. The deliverable’s value is that the EOR walks into the review with the prediction surface, the sensitivity analysis, the sourcing map, and the QC plan already in their hand, rather than with a blank spreadsheet.
Cold-climate exposure: what the literature actually supports
The cabin’s wall set is in a severe freeze-thaw exposure climate. The deliverable’s resilience framing, drawing on USDA Plant Hardiness Zone and NOAA freeze-thaw climatology data for the Black River Falls site, sets a Sunnyday durability design target for severe-freeze-thaw exposure on loadbearing exterior walls. The predicted strength window meets that target with margin. Strength is not the only durability axis that matters in cold-climate 3DCP, however, and the deliverable acknowledges this directly.
The state-of-the-art review on the subject, Mousavi and Rangaraju (2025, CivilEng 6(3), Article 47, DOI 10.3390/civileng6030047), aggregates findings across many primary studies and is the field’s most current synthesis on the freeze-thaw durability of 3D-printed concrete in cold climates. The directional finding, which the review treats as well-established, is that the layer-by-layer extrusion process introduces interfacial vulnerabilities (anisotropic pore distribution, cold-joint zones, lower interlayer bond strength than the bulk matrix) that do not exist in cast concrete, and that those interfaces are typically the dominant freeze-thaw failure plane in 3D-printed walls. The review also documents the available mitigation levers: air entrainment at the right dosage, optimized print orientation (horizontally-laid layers rather than vertically-stacked where geometry permits), shorter layer-cycle times to maintain fresh-on-fresh contact at the interlayer, and tighter nozzle-standoff control. None of those levers is a single-axis fix; they couple to the formulation, to the printer’s process parameters, and to the ambient conditions at print time.
The deliverable’s response is to specify all of those levers as a system rather than to optimize any one of them in isolation. The composition window includes the late-age strength gain mechanism that the literature supports for densified microstructure of cold-climate cementitious systems. The QC plan flags fresh air content as a per-batch test during the trial pour, on the explicit recognition that the high-shear pumping inherent to 3DCP can entrain a meaningful fraction of air without an air-entraining admixture, which means whether to dose AEA at all is a question to resolve at the trial pour rather than to assume up front. Print orientation, layer-cycle time, and nozzle-standoff distance are specified in the QC plan as targets the contractor’s printer needs to hit; how those targets get hit depends on the printer the contractor brings to the project, and the deliverable surfaces the targets so the contractor’s setup can be evaluated against them rather than negotiated after the fact.
This is the operative reason CEMFORGE is not just a mix-design tool but a logistics tool. The cold-climate durability of a 3DCP wall is determined by the formulation, the air content, the print orientation, the layer-cycle time, the moisture protocol, and the cure conditions, simultaneously. Decoupling those is how published 3DCP walls degrade faster than the bench cubes implied. Coupling them is what the engagement actually does.
A single primary experimental study (for instance, Skripkiunas et al. 2025 in Advances in Civil Engineering, DOI 10.1155/adce/8592029) makes the directional argument vividly: at one tested mix and one printing setup, 3D-printed specimens accumulated substantially more freeze-thaw surface scaling than cast specimens of the same nominal mix, and the authors call for a 3DCP-specific durability methodology because conventional cast-concrete protocols do not predict the printed result. The magnitude in any single primary study is a function of that study’s specific printing method and is not directly transferable to other printers or other mixes. The Mousavi and Rangaraju review’s directional finding is the cleaner anchor for what the field actually knows, and it is the anchor the deliverable leans on.
The 3DCP-specific test program
The deliverable’s test plan separates predictions that CEMFORGE covers at no extra cost from physical lab tests that the project must run regardless. The compressive cylinder set, the fresh-state slump-flow and density panel, the initial-set time, and the per-batch fresh air content are predicted by the model and verified at the bench. The physical lab tests required for the project include the static modulus of elasticity (for the EOR’s deflection and serviceability calculations), the drying and restrained shrinkage tests (for cracking-tendency assessment on a high-paste 3DCP mix), and, specific to layered extrusion, the interlayer bond strength via printed-pair pull-off and the buildability and open-time bench test sized to the project’s stack-height target. The cold-climate residual on the test program is a freeze-thaw durability test at the standard three-hundred-cycle protocol, which the EOR uses as the durability acceptance evidence on top of the strength curve.
The total project test budget, expressed per cubic metre of wall pour, is a small fraction of the cost of the printed concrete itself. It is also the part of the project that converts a model prediction into engineering evidence, and the deliverable presents it as line items so the cabin owner sees what they are buying.
What this case is for
This case is for residential and outbuilding owners, regional contractors, and small-scale project sponsors in geographies the central 3DCP supply chain does not economically reach. The argument is not that printed walls are universally cheaper or universally faster than conventional construction at this scale. They are not, as of 2026, on either axis. The argument is that the regional 3DCP service shape, when paired with mix-design intelligence sized to the project and with an explicit Engineer-of-Record handoff, opens a category of work that today is not built with 3DCP at all because the supply chain that has been built around the technology was not designed for it.
This case is also a description of what CEMFORGE actually is. It is a performance and logistics intelligence service, not a code-clearance service. It produces a deliverable an engineer of record can review, not a deliverable a contractor can pour from. The bench validation, the print trials, the structural sign-off, and the code submittal remain where they belong, in the hands of the licensed engineering and construction professionals on the project. CEMFORGE’s role is to make that work tractable for projects whose economics would otherwise not justify the engineering effort.
How to engage
CEMFORGE engagements for residential and outbuilding projects are scoped to the project’s location, geometry, and target performance. A typical P0 confirmation pass for a project of this scale includes a recommended composition window, a Monte Carlo robustness analysis, a sensitivity-driven QC plan, a regional sourcing map, a 3DCP-specific test schedule sized to the application, and an Engineer-of-Record handoff packet. Subsequent phases extend to bench validation, trial print on a representative section, and full-design support against the EOR’s specification.
The CEMFORGE platform itself (cemforge.ai) is the prediction engine that powers the engagement and is also available as a self-service subscription for engineering teams who want to do the formulation work themselves. The engineering substrate the engagement draws on is documented in the 3D concrete printing mix design review.
For project inquiries: info@sunn3d.com.
This is a sample case. The cabin and owner are not real. We made them up to show how our service works. The mix design we describe is real, though. We ran our CEMFORGE software for this sample (session #170, 2026-04-28) and the design came from that. We share the software’s findings in plain words, not as exact numbers.