Connection Hierarchy Mapping

Classify each connection by its expected release cycle, value at risk, and performance duty before choosing the joint technology.

Also known as: Connection Schedule for Disassembly; Release-Cycle Mapping; Disassembly Connection Register

Understand This First

• R-Strategies (R0–R9 / 9R Framework) — the value-retention hierarchy this pattern serves.
• Buildings as Material Banks (BAMB) — the asset frame that turns connection information into recoverable value.
• Bolt Don’t Weld — the common rule of thumb this pattern turns into a project-specific schedule.
• Reversible Mechanical Connection — the connection-quality test used for high-recovery interfaces.
• Layered Construction Sequencing — the sequencing discipline that keeps mapped release points reachable.

Context

Design for disassembly often starts with a slogan: make things demountable. That is not enough for a building. A building contains thousands of interfaces, and they don’t all deserve the same treatment. A primary steel splice, a façade cassette bracket, a plant-room skid connection, a raised-floor pedestal, a demountable partition track, and a sacrificial sealant bead have different lives, different risks, and different recovery value.

Connection hierarchy mapping turns that mess into a schedule the design team can use. It asks a practical question during design development: which connections are expected to stay permanent, which may open once at end of first use, which may open several times as layers change, and which must open frequently for maintenance?

The answer shapes details, specifications, access zones, inspection duties, and the handover file. It also prevents the circular brief from spending money everywhere. A project that tries to make every joint equally demountable usually either overruns its budget or ends up with vague notes no contractor can price.

Problem

Without a connection hierarchy, teams default to two bad extremes. One extreme treats every joint as ordinary construction and hopes a future crew will work it out. The other writes a broad requirement that “connections shall be demountable” without saying which connections matter, how often they must open, what performance duties they carry, or what evidence future reuse will require.

Both extremes fail in practice. Contractors can’t price an undefined reversibility duty. Engineers can’t approve a joint whose future release case conflicts with its present load path. Future deconstruction crews can’t infer hidden priorities from a drawing set that records geometry but not release intent. The result is a building with isolated good details and no coherent disassembly logic.

Forces

• Connection value is uneven. Some joints protect high-value reusable components; others connect low-value consumables that will never justify careful recovery.
• Performance duties vary. Structure, fire, water, acoustics, airtightness, security, corrosion, and vibration can make one connection suitable for release and another unsuitable.
• Future cycles differ. A service-access panel may open every year, a tenant partition every lease cycle, a façade bracket once in thirty years, and a structural splice only at deconstruction.
• Documentation ages faster than intent. If the release class is not recorded, the future owner won’t know which joints were designed for opening.
• A hierarchy has to be buildable. The schedule must be clear enough for design coordination, tender pricing, inspections, and later facilities work.

Solution

Create a connection hierarchy during design development and carry it through construction documentation. Treat it as a schedule, not as a paragraph in the sustainability narrative. Each scheduled connection family should state its building layer, component type, release class, performance duty, access requirement, release method, inspection need, replacement-part assumption, and record location.

A useful release-class scale is simple:

The hierarchy should be organized by layer and system. Structure, skin, services, space plan, and fit-out do not need identical rules. A structural splice may need a one-time release class backed by lifting assumptions and inspection records. A service module may need repeat release with isolation valves, labeled connectors, access panels, and replacement seals. A demountable partition system may need repeat release but lower structural evidence. A sealant joint may remain permanent or sacrificial if the component behind it can still be recovered.

Then connect the hierarchy to drawings. Do not leave it as a spreadsheet that floats outside the documents. Tag representative details. Put release classes in schedules. Coordinate access zones with architecture, structure, MEP, fire, acoustics, façade, interiors, and facilities. State the exceptions where the project deliberately chooses permanence because safety, durability, water tightness, cost, or code makes release the wrong priority.

The best maps also include a “future reader” column. That column answers what a person opening the building years later needs to know: where the connection is, what it holds, what must be unloaded first, which tool releases it, what damage is acceptable, what part must be replaced, which inspection is required before reuse, and where the supporting product or material record lives.

How It Plays Out

A project team is designing a steel-framed civic building. The primary frame is expected to outlast several interior cycles, but the owner also wants a credible route for future member reuse. The engineer maps site-bolted beam and column splices as one-time release connections, with member identifiers, access clearances, bolt specifications, corrosion exposure, lifting assumptions, and inspection records. Welded shop assemblies remain where they make engineering sense. The map doesn’t ban welding; it says which interfaces preserve future member value and why.

In a façade replacement program, the hierarchy sits at the edge of several systems. Primary brackets may be one-time release. Cassettes may be repeat release. Gaskets, seals, trims, drainage pieces, and shading hardware may have different cycles again. The façade consultant uses the schedule to stop an interior fit-out package from burying the bracket access. The facilities team receives a record that separates routine maintenance release from major replacement release, so the first service event doesn’t damage components meant for a longer cycle.

An office landlord uses the same pattern for tenant churn. Raised floors, demountable partitions, ceiling grids, luminaires, service drops, and loose furniture all have short cycles compared with the base building. The connection hierarchy tells the fit-out contractor where screws, clips, tracks, plugs, and labels have to remain accessible. It also tells the landlord which parts are worth cleaning and stocking after a tenant leaves, and which parts should be routed to recycling because their recovery value is too low.

During deconstruction, the map becomes a working aid. The contractor can see which components were designed for intact removal, which connections need temporary support, which joints are sacrificial, and which records should be checked before resale or reinstallation. The contractor still needs site investigation and professional judgment. But the building is no longer a blank puzzle.

Consequences

Benefits

• Turns disassembly-design intent into a priced, coordinated, inspectable deliverable.
• Directs reversible-connection effort toward the joints where reuse value, replacement frequency, or maintenance need justifies it.
• Helps prevent circularity overclaiming by recording where permanence is deliberate and why.
• Gives material passports and building resource passports a physical release logic to reference.
• Improves future deconstruction planning because connection duties, access assumptions, and inspection needs survive the original team.

Liabilities

• Adds design coordination work at the point where teams are already resolving structure, façade, fire, services, cost, and procurement.
• Can be reduced to a paperwork exercise if the schedule is not tied to details, specifications, inspections, and handover records.
• Requires discipline after alterations. A later tenant or maintenance project can break the hierarchy by burying access or substituting incompatible fixings.
• May reveal that some circular ambitions are unaffordable or technically weak. That is useful, but it can create friction with the brief.
• Doesn’t make reuse happen by itself. Storage, testing, ownership, certification, insurance, market demand, and deconstruction contracts still decide whether recovered components return to use.

Related Patterns

Sources

• ISO’s ISO 20887:2020 standard page identifies design for disassembly and adaptability as guidance for integrating DfD/A principles into the design process for buildings and civil engineering works.
• BAMB’s Reversible Building Design topic page and Reversible Building Design guidelines and protocol describe reversible design through transformation capacity, reuse potential, component accessibility, interfaces, and connection types.
• Elma Durmisevic’s doctoral thesis, Transformable Building Structures: Design for Disassembly as a Way to Introduce Sustainable Engineering to Building Design and Construction, supplies the decomposable-building and connection-typology lineage behind BAMB’s reversible-design method.
• Philip Crowther’s Design for Disassembly: Themes and Principles collects design principles for disassembly, including mechanical connections, access to parts, realistic tolerances, minimized connector types, repeated-use joints, labeling, and retained information.
• The AIA practice guide Buildings That Last: Design for Adaptability, Deconstruction, and Reuse frames deconstruction and reuse as design responsibilities, with attention to benefits, pitfalls, case studies, and material reuse.
• The U.S. EPA’s deconstruction manuals page links design-for-deconstruction manuals and identifies exposed connection systems and accessible utility raceways as strategies that make future adaptation and disassembly easier.