1 — The seven failure families
Think of a hospital triage desk. Every arriving patient is sorted — fracture, fever, chest pain, cut — into a small number of streams before any test is run. The reason is speed, not accuracy: a wrong stream sends the patient for the wrong tests; a right stream sends them for the right ones. A failure investigation works the same way: classify the family first, then confirm the mechanism. The seven families below cover most field problems I see on audits.
Cracking covers seven sub-types (plastic shrinkage, drying shrinkage, thermal, settlement, structural, chemical/ASR, crazing); each has its own time of appearance, pattern and first response. The first three questions are: when did it appear, what pattern does it make, and how deep is it. A 30-minute cracking on a hot afternoon is plastic shrinkage; a 6-month map-pattern cracking in a coastal zone is chemical swell. Surface defects include scaling, dusting, blistering, efflorescence and pop-outs; cause is usually finishing, curing, or reactive aggregate. Placement defects (honeycombing, segregation, cold joints, lift joints, grout loss, voids behind reinforcement) show up at strip-out; cause is in the pre-pour decisions. The companion article Honeycombing covers the five root causes.
Strength / cube shortfall covers a 28-day cube below fck, an in-situ strength below design, a delayed strength gain, or a wide spread within one pour. The first move is to verify the cube represents the in-place concrete, not only the mix that left the chute. Durability / chemical covers carbonation- and chloride-induced corrosion, sulphate attack, ASR, DEF, and acid attack; the visible signature is usually progressive — map-pattern cracking, spalling cover, rust staining — and the investigation requires a chemical/petrographic brief. Movement / dimensional covers excessive deflection, creep, joint failure, settlement, and thermal movement; the concrete is the witness, not the cause. Leakage / waterproofing covers water ingress through the body, joints, penetrations, or a failed membrane; the correct first move is to find the path the water is taking, then decide whether to repair the body, the joint, the penetration, or the membrane.
2 — Investigation workflow: the first 48 hours
The first 48 hours decide everything. Before any chipping, repair or restart: photograph the symptom with scale and metadata; measure crack width, spall area, deflection, and dosage/cover where exposed; label and store every sample with chain of custody; sample with the lab's preparation protocol, not improvised; record the weather, the load history, the mixing, placing and curing records. Information density drops sharply after 48 hours — the site moves on, dust falls, and the broken surface gets patched or painted over.
A patient chart no one kept. A doctor who sees the patient only at week 3 has no first-day observations, no fever curve, no record of what was given at hour 1. The diagnosis is built on inference, not evidence. Concrete is the same — the chart is the first 48 hours, and the discipline is called evidence preservation.
3 — Evidence, the lab, and the repair decision
Evidence is the chain of custody. Every sample is identified by member, location, depth, date, time, operator, and project; every transfer is logged; every storage condition is recorded. The lab brief is the contract — the project specifies, the lab executes, and the report is returned against the brief, not improvised. Common failure modes of the lab brief: requesting the wrong test (petrography vs chemistry), wrong sample size, wrong curing, or no acceptance limits. A lab report without a brief is just a number.
The repair-or-replace decision is structural, economic and contractual. IS 456 Cl 17 routes the engineer toward core testing (85%/75% rule) and load testing (deflection limit + 75% recovery in 24 h) before any surgical decision. ACI 562-21 and EN 1504 give the international language for assessment, repair design and product selection (patching mortars, bonding agents, coatings, FRP strengthening). The contractor proposes; the consultant reviews against the brief; the client signs. The temptation is to skip the lab and inject a resin — but the cost of a wrong repair is greater than the cost of a two-week investigation. I've watched the same basement wall get re-injected three times in one monsoon because the path was never traced.
4 — Prevention: close the cause, not the symptom
Prevention always wins. Most failures trace to one of six root causes: mix (water content, SCM, admixture, aggregate contamination), cover (spacer type, spacing, displacement during placement), vibration (under-vibration, over-vibration, sequencing), curing (premature drying, missed period, hot/cold regimes), joints (location, preparation, water-stop detailing), and workmanship (free-fall height, retempering, finished-without-compaction). Each maps to a QA/QC control — ITP row, method statement, hold point, calibration register — that the contractor and consultant already use on the project. The repair closes the symptom; the CAPA closes the cause. A repair without a CAPA is a recurring cost.
Practical field blocks
A 5-question investigator's questionnaire, a 5-step evidence checklist, and two worked examples that show how the workflow closes into a decision.
Questionnaire — five questions
- What is the failure family? Cracking, surface, placement, strength/durability, movement, or leakage? Read the symptom, then classify.
- What is preserved? Photographs with scale, measurements with instrument, samples with chain of custody, weather and load history, mix and curing records — all within 48 hours.
- Is the cube representative of the in-place concrete? Verified identity, casting, curing, testing — or is the cube a property of the heat box, not the member?
- What is the lab brief? Test, sample size, curing, acceptance limits, report format — and is the lab NABL-accredited for the requested test?
- Is the repair linked to a CAPA? The cause is named, the system finding is closed, and the prevention is in the project quality plan — not a one-off patch.
Evidence checklist — five records
- Photographs: overall, mid-range, close-up with scale and timestamp; saved in a project-evident folder.
- Measurements: crack width, spall area, deflection, exposure; instrument and operator recorded.
- Samples: identified by member, location, depth, date, time, operator; chain of custody log started.
- Records: mix ticket, fresh-concrete log, cube register, calibration certificates, weather log, load history.
- Lab brief: test, sample size, curing, acceptance limits, report format, accreditation, turnaround.
What happens if… two worked examples
What happens. The cube is below fck − 3 = 27 N/mm², so the individual-result gate has failed. The contractor and client both want to move on; the consultant recommends demolition without core testing. Skipping Cl 17 means the engineer is not following the code route.
Likely outcome. IS 456 Cl 17 route: extract at least three cores from the affected pour, test per IS 516, accept if the average equivalent cube strength is ≥ 0.85 × 30 = 25.5 N/mm² and no core is below 0.75 × 30 = 22.5 N/mm². If cores pass, the pour is accepted with documentation; if not, the engineer orders load test, strengthening, or removal. The demolition is reversed, the report cites IS 456 Cl 16 and Cl 17.
Preventive correction. Add an "IS 456 Cl 17 read" step to the NCR template; every low-cube NCR is reviewed against the group-of-4 and individual-result rules before any demolition decision.
What happens. Water ingress appears at the construction joint between the wall lift and the base slab. The contractor injects polyurethane resin around the visible wet patch; the leak reappears at the next joint, then the next. The investigation never finds the path.
Likely outcome. Document the joint locations, observe the leak during a wet period, trace the water path. Common causes: missing or displaced water-stop, cold joint from late placement, defective surface preparation, or a failed membrane behind the wall. The repair depends on the path: re-prep and re-bond the joint, replace the water-stop, or fix the membrane. The resin is then a secondary seal, not a substitute for the structural fix.
Preventive correction. Treat every successful repair as a CAPA — record the cause, update the method statement, train the crew, audit the next pour.
References & further reading
- IS 456:2000 (reaffirmed 2021) — Plain and Reinforced Concrete — Code of Practice. Bureau of Indian Standards. Cl 16 (acceptance) and Cl 17 (procedure for unsatisfactory concrete) are the Indian backbone.
- IS 516 — Hardened Concrete — Methods of Test. Bureau of Indian Standards. Part 5 covers rebound, UPV and core extraction for in-situ evaluation.
- IS 13311 — Non-Destructive Testing of Concrete (Parts 1 and 2). Bureau of Indian Standards. UPV and rebound classification. This is the legacy reference; these methods have since been consolidated into the IS 516 series — check the current BIS catalogue before citing.
- IS 13896 — Repair and Rehabilitation of Concrete Structures — Guidelines. Bureau of Indian Standards. Indian companion for repair product selection and procedure.
- ACI 562-21 — Code Requirements for Assessment, Repair, and Rehabilitation of Existing Concrete Structures. American Concrete Institute. International companion for assessment and repair design.
- EN 1504 (Parts 1–10) — Products and Systems for the Protection and Repair of Concrete Structures. CEN. International product and system reference for patch mortars, bonding agents, coatings, FRP strengthening.
- ACI 546R — Guide to Concrete Repair. American Concrete Institute. International practical guide.
- ACI 214R — Guide to Evaluation of Strength Test Results of Concrete. Statistical reading used to interpret cube data.
Year notes: IS 456:2000 reaffirmed 2021; verify current amendment against the BIS catalogue. IS 13896 is the Indian repair guide; verify the current edition. ACI 562-21 is the current edition; EN 1504 spans multiple parts with the latest editions cited where relevant. Always verify the controlled copy before contractual application.
Frequently Asked Questions
Related articles
Why Concrete Cracks — A Complete Guide
Seven cracking sub-types, time of appearance, pattern, mechanism, and first response.
Honeycombing in Concrete — Causes, Prevention, Repair
The placement-defect case: five root causes, three repair thresholds, IS 456 Cl 7.4 reading.
Acceptance Criteria — IS 456 Table 11
The group-of-4 mean and individual-result gates, and the Cl 17 investigation route.
Non-Destructive Testing of Concrete
Rebound, UPV and cores under IS 516 Part 5 and IS 13311, with a defensible QA workflow.
Amit Haridas
Founder & Proprietor, ConcreteInfo. 25+ years of experience in concrete technology, RMC plant operations, construction quality, consulting and technical training across India. NRMCA CTI certified, ISO Lead Auditor.