Base Plates and Anchor Bolts: Where Strengthening Projects Actually Fail

Strengthening schemes are drawn as members. They fail as connections. A column jacket with the correct longitudinal steel and the correct confinement will still fail if the dowels into the footing were drilled into a void, or if the base plate was sized for the pre-retrofit axial load and the new steel frame delivers twice that.

Anchorages are the least glamorous and most defect-prone part of any retrofit. They are also the one part where a design error cannot be compensated by material quality on site.

Three anchor families, three failure modes

Mechanical expansion anchors transfer load through friction and bearing against the hole wall. They depend on hole diameter, drilling method, concrete strength and correct torque. Their weakness is vibration and crack cycling: in a cracked seismic region, an expansion anchor can lose preload and slip.

Adhesive (chemical) anchors bond a threaded rod into a drilled hole with a two-part resin. Capacity comes from the resin–concrete interface along the embedment length, which makes installation quality decisive. Their weakness is hole cleanliness: dust reduces capacity dramatically, and the reduction is invisible after installation.

Cast-in or through-bolted systems avoid the drilled-hole problem but require access to the far side, which is rarely available in retrofit. Where a new steel frame is bolted to an existing concrete element, adhesive anchors are the default — which is why installation discipline is the governing variable.

Hole cleaning is not a detail

The single most repeated finding in post-failure investigations of adhesive anchors is inadequate hole cleaning. The sequence that produces reliable capacity is specific:

1. Drill with the specified bit to the specified diameter and depth — hammer drilling creates a rougher, dustier hole than diamond coring, and the anchor approval states which method it covers.
2. Blow out with oil-free compressed air from the bottom of the hole, using a nozzle that reaches the base.
3. Brush with a steel brush of the correct diameter, in and out, with a twisting motion.
4. Blow out again.
5. Repeat the brush-and-blow cycle the number of times stated in the approval — often two full cycles.
6. Verify visually with a light and, on critical anchors, with a bore scope.

A hole that looks clean from the surface can hold a plug of dust at the bottom. On a downward-drilled hole, dust falls back in; on an overhead hole, it stays. Both are failure paths, and both are cheap to prevent.

Edge distance governs everything

Anchor capacity in concrete is limited by the concrete cone, not by the steel. Small edge distances and tight spacing produce overlapping cones, and the group capacity falls faster than a linear calculation suggests. This is the reason a base plate that fits the column perfectly can be unbuildable: the plate needs to be wide enough to place anchors at a distance from the concrete edge that the design assumes.

In retrofit, the constraint is often the existing plinth or footing, which is smaller than the new demand requires. Then the options are: enlarge the base (if the geometry allows), use a different load path (a new independent foundation), or reduce the demand (fewer, better-placed anchors with a redesigned connection). Adding anchors at 60 mm centres to a 300 mm wide plinth is not one of the options.

The design checks people skip

• Concrete breakout in tension and shear, with the actual edge distances and spacing on the as-built geometry.
• Pry-out where the anchor is near a free edge under tension.
• Steel failure of the rod, including the reduced section at the thread.
• Splitting during installation, especially in thin or low-strength concrete.
• Seismic qualification — anchors in cracked concrete under cyclic load need a qualification that plain static testing does not provide (ACI 355.2 in the US, ETA with seismic provisions in Europe).
• Grout and shim behaviour: an untamped, non-shrink grout layer under a base plate turns a designed bearing into a point load.

Testing: sample what you cannot see

Because adhesive anchor capacity is installation-dependent, testing is not optional on a strengthening project. A workable regime:

Population	Test	Acceptance
Production anchors	Proof load to 1.0–1.5× design tension, sampled	No movement at hold period
Critical anchors (primary load path)	Proof load on every anchor, or full pull-out testing on a sacrificial set	As above, plus documented
First anchors of each type / crew / day	Pull-out to failure on a test anchor in non-critical location	Capacity ≥ design with margin
Rejected anchor	Replace, relocate, or re-drill at reduced capacity	Engineer’s written instruction

The word that matters is “crew”. Anchor failures cluster by installer. Testing a sample from a different team on the same site proves nothing about the team that installed the base plates in the north bay.

The base plate itself

Three failure modes are routinely missed because attention goes to the bolts:

Plate yielding under compression. A plate that is too thin for the cantilever between the column face and the bolt line will bend, and the load path through the grout becomes non-uniform. Check plate thickness against the actual bolt layout, not the nominal one.

Weld and column interface. The new steel column is often welded to a plate that is bolted to concrete. The weld at the column base carries the full moment and shear; if it was detailed for gravity load, the retrofit adds demand it was never sized for.

Corrosion at the base. Base plates sit at the point where water, condensation and de-icing salts collect. A plate with no drainage path, embedded in a damp plinth, corrodes from the top of the concrete upward — and the section loss is hidden by grout and paint. Specify a free-draining detail, a corrosion protection system with a stated durability, and a periodic inspection point in the maintenance manual.

The as-built problem

Retrofit designs are made on drawings that are frequently wrong. A plinth drawn at 400 mm may be 340 mm on site; a footing drawn at 1.2 m may be 0.9 m. When the anchor layout cannot be achieved, the tempting answer is to reduce the edge distance and hope the calculation margin covers it. The correct answer is to survey before fabrication: measure the existing concrete, identify the reinforcement with a cover meter and, where necessary, a radiograph or local breakout, and confirm that the drill path does not hit a bar.

Drilling through reinforcement is not a trivial choice. Cutting a main bar in a column or a footing reduces its capacity at exactly the location where the new load is delivered. Where a bar must be interrupted, the reinforcement needs a designed replacement — usually additional bars or a local enlargement — not a promise that the bar was probably a stirrup.

A checklist that prevents the usual failures

1. Anchor approval covers the drilling method, the base material and the seismic condition.
2. Edge distance and spacing verify on as-built geometry, including reinforcement location.
3. Hole cleaning procedure written down, with brush and blow cycles counted.
4. Installation temperature within the resin’s approved range, with substrate temperature recorded.
5. Cure time respected before torque or load — including overnight and weekend gaps.
6. Torque applied with a calibrated wrench, logged per anchor.
7. Proof testing by crew and by anchor type, results filed before the next lift.
8. Base plate free-draining, grout tamped, corrosion protection complete.

Anchors are cheap relative to the rest of a retrofit and disproportionately responsible for its failures. The four hours spent proving them are the best-value hours on the project.