Concrete repair and strengthening

There is a repair that looks perfect on the day it is finished and produces a ring of fresh cracking around itself within three or four years. It is the most common outcome in concrete repair, and it is entirely predictable from the chemistry.

Understanding why is the difference between repairing a structure and buying yourself a shorter interval before the next repair.

A gloved hand uses a ribbed roller to work resin into a sheet of black carbon fibre fabric wrapped around a circular concrete column. Below the sheet, an earlier wrap is already complete and dark with resin. Above it, a band of grey repair mortar has been applied, and above that the original concrete surface is pitted and honeycombed with exposed aggregate. Behind the column, the steel structure and pipework of a process plant recede, with a worker in an orange hi-vis vest and white hard hat walking away. — Three different jobs on one column, in the right order: break out and patch, then wrap. The fabric cannot fix what is still corroding underneath it.

The repair must be designed against a cause

ACI 562 and EN 1504 both begin in the same place: a repair is designed, and the design input is the diagnosed cause. That diagnosis comes from the assessment described in assessing corrosion in reinforced concrete, and without it the repair is a guess.

The practical reason is specific. The two mechanisms demand different responses:

Carbonation damage can usually be repaired by removing the carbonated concrete and replacing it with alkaline repair material. The front is broad, so what you remove is everything the front reached, and what is left behind is sound.
Chloride damage usually cannot, because the chloride is not confined to the visibly damaged area. Patching the spall leaves contaminated concrete all around it, and that is what produces the ring of new cracking.

The incipient anode: the repair that causes the next failure

Break out a spalled area, clean the bar, fill it with fresh alkaline mortar, and the steel inside the patch repassivates. It is now protected.

The steel immediately outside the patch is still sitting in chloride-contaminated concrete and is still depassivated. You have just created a well-defined cathode next to an existing anode, with a potential difference between them and a continuous bar carrying the current.

Corrosion accelerates in a ring around the repair. It is known as the incipient anode, the ring anode or the halo effect, and it typically shows up in three to five years — faster than the original damage took to appear.

The defences are to remove contaminated concrete well beyond the visibly damaged zone, or to stop patching and change the electrochemistry of the whole element instead.

Getting a patch repair right

On the left, two sections through a patch repair in a concrete member, one wrong and one right. In the wrong one, the break-out stops at the face of the reinforcing bar, so contaminated concrete remains behind the bar, the edges of the patch are feathered to a thin wedge, and the repair stops at the edge of the visible damage. Arrows show current flowing from the repassivated steel inside the patch to the still depassivated steel just outside it, with new corrosion marked at the patch edge. In the right one, the break-out is taken about twenty five millimetres behind the bar so the repair material fully encases it, the edges are saw cut square to a defined depth, the bar is cleaned to bare metal all round, and the break-out extends well past the visible damage into uncontaminated concrete. On the right of the figure, a ladder of intervention options runs from doing nothing but monitoring, through surface protection, local patch repair, electrochemical treatment of the whole element, and finally structural strengthening, with the condition that makes each one the right answer. — Left, the same repair done two ways. Right, patching is one rung on a ladder, not the default.

The detailing is not complicated, and nearly all of it exists to avoid the two failure modes — continued corrosion under the repair, and debonding of the patch.

Break out behind the bar. Commonly about 20 to 25 mm of clearance behind the steel, so the repair material encases it completely. Stopping at the face of the bar leaves the back of it in the old concrete, where it carries on corroding under a repair that cannot be inspected.

Extend past the visible damage. Delamination always runs further than spalling. Sound the area with a hammer and break out to where the ring is solid, then go further if the chloride profile says the contamination does.

Cut the edges square. A saw cut 10 to 15 mm deep around the perimeter. No feather edges — a thin wedge of repair material has nothing to hold it and will debond, usually at the first hot summer.

Clean the steel properly. Grit blast to bare metal all round, not just on the exposed face. Any corrosion product left on the bar continues to expand and keeps a contaminated interface against the new material.

Decide about the bar. Measure the remaining diameter. Where section loss is significant, supplementary bars are lapped in — and that is a structural decision requiring a calculation, not a site judgement.

Choose a compatible material, which usually means close to the substrate in elastic modulus and coefficient of thermal expansion, and as low as possible in drying shrinkage. A repair mortar that is far stronger and stiffer than the parent concrete attracts load and restrains movement, which puts tension straight across the bond line.

Cure it. Repair mortars are thin sections with a large exposed surface, which is the worst possible geometry for shrinkage. Most patch repairs that crack were simply not cured.

When patching is the wrong answer

Where chloride is widespread, the economic and technical answer changes. EN 1504 Part 9 organises this well: its eleven principles cover both defects in the concrete and corrosion of the reinforcement, and the ones that matter here are preserving or restoring passivity, increasing resistivity, cathodic control and cathodic protection.

Option	What it does	When it fits
Monitor only	Nothing, deliberately, with a defined review	Slow rate, high resistivity, low consequence
Hydrophobic impregnation or coating	Keeps water and chloride out of sound concrete	Preventive, or after the cause is removed
Local patch repair	Removes the damaged and contaminated concrete	Carbonation, or genuinely local chloride
Corrosion inhibitor	Raises the chloride threshold at the steel	Supplementary; penetration is hard to verify
Chloride extraction	Pulls chloride out electrochemically, as a temporary installation	Contaminated cover, bars still largely sound
Realkalisation	Restores alkalinity to carbonated concrete	Carbonation too deep to break out economically
Cathodic protection	Makes the whole steel cathodic, permanently	Widespread chloride contamination
Strengthening	Replaces lost capacity rather than lost durability	Section loss, or a changed load

Cathodic protection deserves a line of its own. It is the only option in that list that works regardless of how much chloride is present, because it addresses the electrochemistry rather than the contamination. The price is a permanent installation — anodes, cabling, a power supply for impressed-current systems, reference electrodes, and a monitoring regime that has to be funded and staffed for the life of the structure. Systems that are installed and then forgotten are worse than no system, because everyone believes the problem is solved.

Strengthening

Strengthening answers a different question from repair. Repair restores durability; strengthening restores or adds capacity, either because section has been lost or because the load has changed.

Section enlargement. Add concrete and reinforcement. Crude, heavy, and completely reliable. It needs the interface roughened and dowelled, and it changes the stiffness of the member, which can redistribute load elsewhere in the structure.

Bonded steel plate. Epoxy-bonded steel plating, usually with mechanical anchors. Effective, but the plate needs protection from corrosion and fire and it is heavy to handle on site.

Externally bonded FRP — the carbon fabric in the photograph. The attraction is obvious: very high tensile strength, negligible weight, no change to the member size, and installation with hand tools. The cautions are specific:

It fails by debonding, not by fibre rupture. The fibre is far stronger than the concrete cover it is bonded to, so the governing failure is in the substrate just under the adhesive. Substrate pull-off strength is a design input, verified by test.
Surface preparation is the job. Grinding back to sound concrete, rounding corners on rectangular columns so the fabric is not cut by the arris, filling blowholes, and controlling dust, moisture and temperature during application.
Design strain is limited well below the fibre’s rupture strain, precisely to keep the bond line in control.
Fire and impact. The resin loses its properties at modest temperatures, so most codes require the unstrengthened member to retain enough capacity to carry a reduced load if the FRP is lost — the strengthening cannot be the only thing holding the structure up.
Confinement is the best use of it. A full wrap around a circular column, as in the photograph, puts the fibre in pure hoop tension and confines the concrete, which raises both strength and ductility. There is no debonding problem in a closed wrap, which is why column jacketing is the most reliable FRP application there is.
It does nothing about corrosion. Wrapping a corroding column seals the problem in and hides it. Repair first, wrap second — which is the order visible on the column in the photograph.

External post-tensioning. For adding flexural or shear capacity to beams and slabs, and closely related to the techniques covered in post-tensioned containment walls. Powerful, and it demands attention to anchorage, deviator details and corrosion protection of the new tendons.

Recording the repair

A repair becomes part of the structure’s history, and the next inspection has to know what it is looking at. The record needs the extent and location of every break-out, the materials and batches used, the condition of the steel found and what was done about it, the test results, and the diagnosed cause the repair was designed against.

Without that last item, the engineer who finds cracking around the patches in five years cannot tell whether the repair failed or whether it was never the right repair.

What to take away

Diagnose first. EN 1504 and ACI 562 both make the cause an input to the design, not an afterthought.
The incipient anode is the default outcome of patching chloride-contaminated concrete. Design against it deliberately.
Break out behind the bar, square the edges, extend past the visible damage, and cure the repair.
Compatibility beats strength in choosing a repair material.
Widespread chloride means electrochemistry, not patching — most often cathodic protection, with the monitoring commitment that implies.
FRP fails in the concrete, not the fibre. Prepare the substrate, respect the strain limit, and never wrap over an active corrosion problem.