The Challenge
The Assessment That Standard Software Cannot Do
Heritage building load assessment sits at the intersection of structural engineering's two most demanding requirements: technical rigour and practical judgment. It is technically demanding because the material properties, construction techniques, and load-bearing systems in pre-modern buildings do not conform to the assumptions embedded in contemporary structural codes. It is judgment-intensive because the deviations from those assumptions are not systematic — they are building-specific, shaped by the construction practices of a particular era and region, the material quality available at the time of construction, and the modifications and degradation that have accumulated over decades or centuries of use.
The problem for practitioners is that standard structural software is built on contemporary code assumptions. It handles the load cases, material properties, and failure modes relevant to modern construction competently. For heritage buildings, it produces outputs that are nominally valid but practically misleading — because the code values it uses are wrong for the materials, and the failure modes it models are not the failure modes that matter in a 19th-century load-bearing masonry structure.
The gap is not computational. It is epistemological: what the software needs to know about heritage materials, failure modes, and assessment strategies cannot be derived from contemporary codes. It has to be learned from experience — from the accumulated knowledge of engineers who have spent careers assessing these structures, making conservative judgments where the standard tables do not apply, and refining their assessment strategies based on what they have seen fail and what has held.
Heritage building structural assessment is, in practice, one of the most expertise-intensive assessments in the field — and one of the most under-served by existing tools. Experienced practitioners carry the relevant knowledge. Junior engineers assessing heritage structures do so largely without access to it, relying on code values that do not apply and software outputs that do not account for the failure modes that experienced engineers know to look for.
The Approach
Encoding Failure-Mode Expertise Across 40 Projects
Identifying the Right Knowledge Base
Dr Priya Sharma is a structural engineer with 20 years of practice in infrastructure and heritage assessment, including 40 major projects across heritage building types ranging from colonial-era stone construction to early twentieth-century reinforced concrete. Her value to this project was not her knowledge of structural codes — that knowledge is published and teachable. Her value was her failure-mode library: the accumulated record of what she had seen go wrong in heritage structures, under what conditions, with what precursors, and how those failure modes differed from the patterns that standard codes model. This is knowledge that exists only because she has been present for failures, near-failures, and the assessments that preceded them — and has been systematically reflective about what she observed.
Mapping the Decision Process for Heritage Assessment
Knowledge-encoding sessions with Dr Sharma were structured around her project history, working through the major heritage building types she had assessed and mapping her decision process for each. The focus was not on the standard elements of structural assessment — those are documented — but on the deviation points: where she stops using standard code values and substitutes her own judgment, what signals she looks for to identify problematic material degradation, how she chooses between conservative and progressive assessment strategies when the building history is uncertain, and what the specific failure modes are that standard assessment approaches systematically miss in heritage masonry, heritage timber, and early reinforced concrete.
A particular focus was the "judgment trigger" — the specific conditions under which Dr Sharma decides that a standard approach is not adequate and a more detailed investigation is required. These triggers are not documented anywhere. They are the product of experience: of having applied the standard approach, obtained an apparently acceptable result, and then watched the building fail for reasons the standard approach did not capture. They are the most valuable elements of her knowledge base, and the hardest to elicit, because they are defined precisely by the cases where standard approaches were insufficient.
Heritage-Specific Assessment Decision Nodes
The encoding process produced 62 heritage-specific assessment decision nodes — structured points in the assessment process where Dr Sharma's judgment diverges from standard approaches, with documented reasoning for each divergence. These nodes cover load case selection for heritage floor systems, material strength estimation for degraded masonry, assessment strategies for structures with unknown or incomplete construction records, failure mode identification in heritage timber connections, and the specific conditions under which each type of heritage intervention — strengthening, repair, monitoring — is the appropriate response.
Each decision node is documented with the class of conditions that triggers it, the reasoning behind the recommended approach, and the cases from Dr Sharma's project history that informed it. The documentation is specific: not "consider material degradation" but "for lime mortar masonry with visible efflorescence above the damp-proof course level, reduce the characteristic compressive strength estimate by this factor and apply this modified load case, for these reasons, based on these observed failure modes."
Validation Through Project Review
The encoded assessment framework was validated against a sample of Dr Sharma's completed projects — applying the system's decision logic to the assessment parameters of each project and comparing the output against her documented assessment decisions. Projects where the system's recommendations diverged from her actual decisions were examined in detail to identify encoding errors, gaps in the decision logic, or cases where her judgment had evolved since the project was completed. The validation process also identified several areas where the encoded logic required additional nuance for specific building types — refinements that were incorporated before the system was finalised.
The Product
A Guided Assessment Framework Built on Expert Judgment
The resulting product guides structural engineers through heritage building load assessment using Dr Sharma's judgment framework as the decision logic. It is not a replacement for structural engineering expertise — it is a structured guide that makes Dr Sharma's 20 years of heritage-specific experience available to engineers who are applying standard assessment approaches to non-standard structures.
The product works by identifying, at each stage of the assessment process, whether the current conditions match the scenarios where standard approaches are adequate or where practitioner judgment is required and why. When a standard approach is appropriate, the system confirms it and moves on. When the conditions match a pattern in Dr Sharma's encoded case base that requires a different approach, the system flags the issue, explains why the standard approach is insufficient for this specific condition, and recommends the modified approach that her experience has validated — with the reasoning made explicit.
This is a fundamentally different kind of tool from standard structural software. Standard software optimises for the cases where code assumptions hold. This product is designed for the cases where they do not — which is precisely the case that heritage building assessment represents. It makes no claims about modern construction. It makes specific, validated claims about the failure modes, material properties, and assessment strategies relevant to the heritage building types in Dr Sharma's documented project base.
The output of the system is an annotated assessment pathway: a record of the decisions made, the conditions that triggered each decision, the recommended approach for each, and the reasoning behind each recommendation. This record is directly useful for engineering quality review, for client reporting, and for the professional documentation requirements that heritage building projects typically carry.
20
Years of structural failure expertise — encoded and operational
62
Heritage-specific assessment decision nodes mapped across building types and failure modes
40
Major infrastructure projects underpinning the framework — each load recommendation traceable to documented engineering rationale
"The gap in engineering AI isn't compute power. It's the judgment that comes from being wrong once and learning why."— Dr Priya Sharma, Structural Engineer
Engineering knowledge that took decades to build deserves to scale.
If your practice contains technical expertise that standard software cannot encode, Praxa builds the product that makes it accessible.
Work with Praxa