# Structure ### Material Quantity Estimation C.Scale estimates quantities of major structural materials using a suite of machine learning models trained on a database of bills of materials from completed buildings (n > 1,200), assembled from internal and public sources. This approach is preferable to a first-principles method — i.e., assuming an optimized structural grid for a given geometry — because it reduces truncation error and better describes the variation present in real buildings. Models are updated regularly as new data becomes available. Each model is trained on a small set of predictors and weighted by data provenance. LCI data from a detailed whole-building LCA is given the highest weight, with data from journal articles, white papers, partial data, and modeled data given progressively lower weights. Training data is split 80/20 between training and test sets, and k-fold cross-validation (n=5) is used to confirm that each model can consistently predict unseen data. There are five ML models in our model pipeline. The least accurate has a Mean Absolute Percentage Error of 16.7% and an r2 value of 0.74. The four other models each achieve an r2 value over 0.95. These methods for calculating a structural bill of materials have been reviewed by colleagues at [MKA](https://www.mka.com/), [Carbon Leadership Forum](https://carbonleadershipforum.org/), and Autodesk with additional comment from colleagues at [Arup](https://www.arup.com/). If you are a structural engineer or data scientist interested in providing further review of our modeling pipeline, [please reach out](mailto:hi@cscale.io). #### Simple Editing View In simple editing view, users select a primary structural system and, where applicable, a secondary structural system. C.Scale uses these inputs to generate a complete estimated bill of materials for the project. Material assignments are generated automatically based on structural system type and are divided between substructure and superstructure. Assignments may not reflect all project-specific conditions and can be reviewed and edited in the Detailed View. #### Detailed Editing View In the detailed editing view, users can see and interact with the full material inventory, organized by substructure and superstructure. Users may edit quantities, remove materials, or add materials from C.Scale's library of over 300 generic structural materials. Each material entry can be assigned to an element type (foundations, columns, beams, framing, etc.), and quantities can be adjusted to match a specific structural design. All entries are currently managed as individual line items. > **Coming soon:** Batch import of bills of materials from external software. The structural systems available in C.Scale are described in the table below. #### Superstructure
Structural SystemDescription
Reinforced ConcreteA structural system comprised of columns, beams, and slabs of cast-in-place concrete reinforced with steel that provides tensile strength.
Steel FrameA structural system comprised of foundations, columns, beams, girders, and decking constructed from steel structural members connected with rigid or pin joints.
Light Wood FrameA platform frame system comprised of dimensional lumber, plywood sheathing, and reinforced concrete cores.
Mass TimberA structural system comprised of massive beams, panels, and columns, often assembled by aggregating many smaller timber elements.
Hybrid Concrete-Steel (High-Rise)A structural system that combines rigid steel frames with concrete columns, beams, and slabs. These hybrid structures are more materially intensive and may be used when there are significant seismic loads, in high-rise buildings, or for programs with very high live or environmental loads.
#### Substructure The substructure includes all below-grade construction, from the lowest-level slab to the underside of the ground floor. This may include foundation slabs, footings, grade beams, basement walls, and below-grade columns. All substructure elements are assumed to be cast-in-place concrete. ### Carbon Emissions of Structural Materials C.Scale calculates carbon emissions for each structural material by applying global warming potential (GWP) values drawn from industry-average EPDs to the estimated material quantities. In all cases, C.Scale uses GWP-100 characterization factors. These sources are documented in the [Reference Data Sources ](/the-c.scale-tm-data-model/reference-data.md)section of this guide. For most structural materials, emissions data reflects North American industry averages at the time of the last model update. Concrete is an exception — it is treated regionally given its local supply chain and the relative abundance of regional EPD data. See the Concrete section below. #### Concrete Concrete is a local material, rarely traveling more than 25 miles between production and use. Across North America, the emissions associated with typical concrete mixes can vary considerably due to variability in local sand and aggregates, availability of supplementary cementitious materials (SCMs), and regional industry practices. C.Scale accounts for this by applying regional benchmarks rather than a single national average. **In the simple editing view**, users set concrete emissions as either **Average** or **Low-Carbon**: * **Average** references the relevant NRMCA or Canadian Regional benchmark mix for the project location. * **Low-Carbon** references the low-carbon mix from the NRMCA national average EPD, corresponding to a 50% SCM mix design. **In the detailed editing view**, concrete entries are organized by strength class. Users can apply national averages or select a specific regional benchmark for each strength class. NRMCA mixes are available across a range of SCM percentages; Canadian mixes are available with and without air-entraining admixtures (AEAs). > **Coming soon:** Support for product- and facility-specific EPDs across all structural materials. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.cscale.io/the-c.scale-tm-data-model/whole-life-carbon/structure.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.