NDT test for concrete as per IS code: methods and reports

Which IS code covers NDT tests for concrete?

IS 516 (Part 5), "Hardened concrete: methods of test, Part 5 Non-destructive testing of concrete". Each test has its own section, and each section says that the results of one method should be complemented by other tests.

Is IS 13311 still valid?

For UPV, no. The BIS portal marks IS 13311 (Part 1):1992 as "Withdrawn", superseded by IS 516 (Part 5/Sec 1):2018.

For the rebound hammer, the picture is less tidy. The foreword of IS 516 (Part 5/Sec 4):2020 says it applies in place of IS 13311 (Part 2):1992, "which shall stand withdrawn after the publication of this standard". But when we checked in October 2026, the BIS portal still listed IS 13311 (Part 2):1992 without a withdrawn mark. Tenders and older specifications still quote it. Test to IS 516 (Part 5/Sec 4) unless your customer's contract says otherwise, and record the standard the customer asked for.

Where do these tests sit in a NABL scope?

NABL 120 (Issue 01, Amendment 06, 22 December 2025) puts them in the Non-Destructive Testing discipline, under "Building Materials – Reinforced Concrete Structures" (1.9.1). That group lists carbonation, cover meter, crack width, deflection (load test), half-cell potential, pile integrity, rate of corrosion by linear polarisation resistance, rebound hammer and UPV. NABL 120 lists cement concrete as a product under the Mechanical discipline (1.8.2), so cube and core strength tests are normally scoped there. A structural audit with both NDT and cores can therefore need two disciplines in your scope.

How is the rebound hammer test done as per IS 516?

What it measures. A spring-loaded mass strikes a plunger held against the concrete. The rebound depends on surface hardness, which is taken to relate to compressive strength (IS 516 (Part 5/Sec 4), clause 3.2). The standard lists four uses: estimating likely strength with a correlation, checking uniformity, comparing concrete with standard requirements, and comparing one element with another.

Equipment.

• A rebound hammer with the right impact energy (Table 1): about 2.25 Nm for normal-weight concrete, 0.75 Nm for lightweight concrete or small, impact-sensitive parts, and 30 Nm for mass concrete such as pavements and hydraulic structures.
• An abrasive stone of medium-grain silicon carbide or equivalent.
• A testing anvil: a steel cylinder 150 mm in diameter and 150 mm high, with an impact face of 64 to 68 HRC. The manufacturer should state the anvil reading range for the hammer.

Hammers work on two principles: some read the rebound distance, others the ratio of hammer speed after and before impact. Clause 3.2 says rebound numbers from the two are not comparable. Note 2 to clause 5.2.2 adds that two hammers of the same type can differ by 1 to 3 units, so compare results from the same hammer, or run comparative tests first.

Checks. Test the hammer on the anvil before starting and after finishing (clause 5.1). The note to the clause warns that a pass on the anvil does not guarantee repeatable numbers at other points on the scale.

Procedure (clause 6).

1. Choose a smooth, clean, dry surface. Rub off loose scale with a grinding wheel or stone. Avoid honeycombed, spalled, tooled or poorly compacted areas.
2. Keep each impact at least 25 mm from any edge or change in shape.
3. Hold the hammer at right angles to the surface. Horizontal impacts on vertical faces are preferred. Upward or downward impacts give different numbers for the same concrete, so compare readings in the same direction or apply established correction factors. Without your own data, the manufacturer's direction correction may be used (clause 5.2.2).
4. Test around every point of observation on all accessible faces of the member.
5. Take six readings around each point. The average of these, after deleting outliers as per IS/ISO 16269 (Part 4), is the rebound index for that point (clause 6.4).

Readings to record. Member and point, face, hammer direction, the six readings, outliers removed, the rebound index, hammer identification and the anvil readings before and after.

Factors that change the number (clause 7).

• Carbonated surfaces overestimate strength, by up to 50 percent in extreme cases. Where the concrete is more than six months old, check the carbonation depth and give it in the report (7.1.5).
• A wet surface can read about 20 percent lower than equivalent dry concrete (7.1.3).
• Trowelled and floated surfaces read harder than moulded ones (7.1.3).
• Readings are generally higher near the bottom of a pour (7.1.6).
• Cement type matters: high alumina cement concrete can show strengths 100 percent higher, and supersulphated cement concrete 50 percent lower, than ordinary Portland cement concrete (7.1.1).
• Age can be ignored for concrete more than 14 days old (7.1.4).
• The method suits only close-textured concrete. Open-textured, honeycombed or no-fines concrete is unsuitable (7.1.3).

Can the rebound number give strength for M25 concrete?

Not on its own. People search for a table of rebound numbers for M20, M25 or M30, but IS 516 (Part 5/Sec 4) gives none. Clause 5.2 says the relationships supplied by hammer manufacturers "shall be used only to provide indications of relative concrete strength". To estimate strength, you need a correlation for that concrete and that hammer, built in one of two ways:

• From cubes (5.2.1). Project-specific cubes on larger projects: at least three cubes for each of three grades. On smaller projects, a regional correlation from cubes with similar materials, repeated every year. The cubes are wet-cured for 27 days, kept in laboratory air for about 24 hours, then held in the compression machine at about 7 N/mm² (for a 2.25 Nm hammer) while you take at least nine readings on each of two cast vertical faces. Then they are crushed to IS 516 (Part 1/Sec 1).
• From cores (5.2.2). At least two replicate cores from at least six locations with different rebound numbers. On smaller projects the number of cores may be limited to six.

Note 1 to clause 5.2.2 allows a curve made for similar concrete in the same region for an approximate estimate, but only where neither route is possible.

Even with a correlation, clause 8.1 says the probable accuracy can be up to ±25 percent, depending on the correlation and how it was built. Rebound indices also reflect only the concrete near the surface.

Why must rebound and UPV be used together?

Clause 8.2 calls the combined use of UPV and rebound hammer "a must" for proper interpretation. A strength figure from the rebound hammer is valid only where the direct UPV is at least 3.50 km/s for grades below M25, or at least 3.75 km/s for higher grades. Where UPV shows the concrete as doubtful, clause 8.3 says no strength assessment shall be made from the rebound hammer test. The same thresholds apply to the locations used to build a core correlation (5.2.2).

The wording of these clauses leaves M25 itself unclear, while the amended UPV grading table in Section 1 puts M25 in the lower band. State on the report which threshold you applied.

A worked example (fictitious numbers). An M30 column. The project correlation from cores is: strength = 1.25 × rebound index − 12, in N/mm².

At C3-2 the arithmetic would give 33 N/mm², but the standard does not allow it to be reported.

How do you interpret UPV test results?

What it measures. The time a longitudinal ultrasonic pulse takes to cross a known path in the concrete. Velocity is V = L/T. Dense, uniform concrete gives higher velocities. Cracks, voids and poor compaction lower them, because the pulse goes around the defect (IS 516 (Part 5/Sec 1), clause 2.1).

Equipment (2.2 and 2.3).

• A pulse generator, a pair of transducers, an amplifier and an electronic timer.
• Transducers of 25 to 100 kHz suit most work. Use 60 to 200 kHz for short paths down to 50 mm, and 10 to 40 kHz for long paths up to 15 m.
• Two reference bars from the supplier, with transit times of about 25 µs and 100 µs, known to ±0.2 µs.

Checks. The short bar sets the zero. The long bar checks accuracy, and the reading must be within ±0.5 percent of its known value. The instrument must measure transit time to ±1 percent between 20 µs and 10 ms. A zero check with the transducer faces pressed together, with couplant, is advised at regular intervals (2.3.2).

Procedure (2.4).

1. Remove plaster or coatings at each point, with a carborundum stone or grinder, without damaging the concrete.
2. Use a couplant such as petroleum jelly, grease, liquid soap or kaolin-glycerol paste.
3. Choose the arrangement:
   • Direct (opposite faces) is the most efficient and shall be used if possible.
   • Semi-direct: adjacent faces, or opposite faces but not directly opposite each other.
   • Indirect (same face) is the least sensitive. It is for when only one face is accessible, or when the surface layer is of interest. Annex A gives the method: move the receiver in steps, plot time against distance, and take the inverse slope as the velocity.
4. Keep the path long enough: at least 100 mm for aggregate up to 20 mm, and 150 mm for 20 to 40 mm aggregate. On an unmoulded surface, use at least 150 mm for direct transmission and 400 mm for surface probing.
5. Mark a grid of 300 × 300 mm or smaller, each junction a point. Up to 500 × 500 mm may be used on large, uniform members with no signs of distress.

Readings to record. Member and point, transmission mode, path length, transit time, velocity, transducer frequency and the reference bar readings.

Which UPV grading table should a report use?

The one in force after Amendment No. 1 of November 2019, which replaced Table 1 of the 2018 standard with grade-wise criteria. The BIS portal lists two amendments to Section 1, so check both against your copy.

"Doubtful" means further tests are needed. The 2018 table had four bands: above 4.40 Excellent, 3.75 to 4.40 Good, 3.00 to 3.75 Doubtful and below 3.00 Poor. If your report template still has a "Poor" band, it predates the amendment.

The difference matters. Take a fictitious M20 column with a 300 mm path and a transit time of 83.3 µs. V = 300 ÷ 83.3 = 3.60 km/s. Under the amended table that is "Good". Under the 2018 table it was "Doubtful".

Other rules in Section 1.

• The grading is for quality only. It "shall not be used for estimating the concrete grades" (2.5.2).
• A location more than 10 percent below the average of its member or part is treated as having internal flaws, segregation or micro-cracks (2.5.3). If a beam averages 4.10 km/s, a point at 3.60 km/s is 12 percent lower and gets flagged.
• Surface probing usually reads lower than cross probing, often by about 0.5 km/s in good concrete. The standard recommends adding 0.5 km/s to surface-probing values above 3.0 km/s (2.4.3.2.5). Report the mode and any adjustment.
• Strength from UPV alone is "not adequate". A correlation made with the same materials and mix may still be off by ±20 percent (2.5.4).
• Annex C gives a method for the depth of a surface crack, and Annex D for the dynamic modulus of elasticity.

Limits (Annex B). Saturated concrete can read up to 5 percent higher than dry concrete. Temperatures of 5 to 30 °C have no significant effect. Between 30 and 60 °C velocity can fall by up to 5 percent, and frozen concrete can read up to 7.5 percent higher. Steel bars near the path raise the reading, because pulses travel 1.2 to 1.9 times faster in steel, and the standard gives no correction for this. Stress above about 60 percent of ultimate strength lowers the velocity.

Section 1 has no clause listing what the report must contain. The report rules of ISO/IEC 17025 clause 7.8 still apply, and we cover them below.

How is the half-cell potential test done and read?

What it measures. The corrosion potential of uncoated reinforcing steel, as a voltage against a reference electrode on the concrete surface (IS 516 (Part 5/Sec 2), clause 3). The scope says it shows the possibility of corrosion. It cannot confirm that corrosion has occurred or measure how much. The test can be used at any age of the structure, in any climate, as long as the temperature is above +2 °C.

Equipment (clause 4).

• A half-cell, usually copper/copper sulphate. The tube must be at least 25 mm inside diameter, the porous plug at least 13 mm, and the copper rod at least 6 mm in diameter and 50 mm long.
• A junction device, a contact solution and lead wires.
• A battery-operated voltmeter with ±3 percent end-of-scale accuracy and an input impedance of at least 10 MΩ.
• Below about 10 °C, add about 15 percent by volume of isopropyl or denatured alcohol to the contact solution.

Procedure (clause 5).

1. Set up a coordinate system on the structure and record the position of every reading, even in a random pattern. Keep the grid spacing constant if you will plot statistics. The standard notes that 1.2 m has suited bridge decks, and that spacing should be reduced where adjacent readings differ by more than 0.15 V.
2. Connect the steel directly to the positive terminal of the voltmeter, with a compression-type ground clamp or a brazed or welded rod. Connect the half-cell to the negative terminal.
3. Place the cell and watch the reading. If it drifts, pre-wet the concrete until the reading stays within 0.02 V for at least 5 minutes, with no free water left between grid points.
4. Record each reading to the nearest 0.01 V. If the cell temperature is outside 22.2 ± 5.5 °C, apply the temperature correction of about 0.00091 V per °C given in clause 5.6.

Checks (clause 7). The same cell, disconnected and reconnected at one spot, should agree within 0.01 V. Two different cells at the same spot should agree within 0.02 V. Clean the copper rod with dilute hydrochloric acid, never with steel wool.

What do half-cell potential values mean?

Table 1 of Section 2 gives the likely corrosion condition. Readings from other electrodes use the equivalent values in the same table.

Section 2's foreword says it drew on ASTM C876-15. The first three bands match the ASTM C876 criteria as the US Federal Highway Administration summarises them: more negative than −350 mV, a 90 percent probability of active corrosion; less negative than −200 mV, a 90 percent probability of none; uncertain in between. The IS adds the "severe" band below −500 mV and the equivalents for other electrodes.

Limits. Clause 6.1.1 says these criteria should not normally be used, unless experience or destructive examination supports them:

• where the concrete has carbonated to the level of the steel;
• for indoor concrete that is not frequently wetted, unless it was protected from drying after casting;
• to compare outdoor concrete with highly variable moisture or oxygen at the steel;
• to judge changes with time after a repair that changed the moisture or oxygen at the steel.

A note to clause 6.1.2 adds that a potential should not be read as a corrosion rate. Chloride and other chemistry at the steel can make potentials more negative.

Readings and report (clause 8). Besides the date, member, visible signs of corrosion and concrete age and grade, the report must give:

• the cell type, if not copper/copper sulphate;
• the estimated average cell temperature;
• how the concrete was pre-wetted and how the voltmeter lead was attached to the steel;
• an equipotential contour map showing the steel contact point, or a cumulative frequency plot, or both (Annex A: contour interval no more than 0.10 V; plotting position = rank ÷ (number of readings + 1) × 100);
• the percentage of readings more negative than −0.35 V;
• the percentage of readings less negative than −0.20 V.

For example, if a fictitious slab survey gives 50 readings, with 9 more negative than −0.35 V and 21 less negative than −0.20 V, the report states 18 percent and 42 percent.

How is carbonation depth tested as per IS code?

What it measures. Carbon dioxide from the air diffuses into the cover concrete and reacts with calcium hydroxide. This lowers the pH of the pore solution from 12.5 to 13.5 down to about 8 or 9, where the protective film on the steel is no longer stable and corrosion can start (IS 516 (Part 5/Sec 3), clause 2.1). The test finds how far this has gone.

Equipment. A 1 percent phenolphthalein indicator: 1 g of phenolphthalein powder in 100 ml of a solution of 70 ml ethanol and 30 ml de-ionised water (3.1). You also need a fine sprayer, a scale to measure depth, and a camera, because the report needs photographs.

Procedure (clause 3).

1. Test a freshly exposed surface: a freshly broken face, or a core, preferably split. Without a splitting facility, surface-dry and seal the core to stop further carbonation. Clean off dust immediately.
2. Spray a fine mist of indicator, without letting it run in channels.
3. Measure once the colour is stable, usually after about 5 to 10 minutes. Concrete that turns magenta is not carbonated. The edge of the colourless zone is the carbonation front.
4. Where the front is a straight line parallel to the surface, record its depth, dk. Where it is uneven, record a graphical average dk and the maximum dk,max.
5. Where the front runs into a dense aggregate particle, use a theoretical line across the particle. Where it meets an air void, include voids under 4 mm deep in the mean and exclude deeper ones (clause 4).

Readings to record. Take at least three specimens from each structural member. Record the depth for each specimen face, then the mean depth for the member (clause 5).

Report (5.1). Mix or grade if known, the member and location, the age of the member, the date of test, the mean depth for the member, individual and average results with photographs (including notes on uneven fronts and aggregate or void corrections), and any comments.

Interpretation and limits. Section 3 gives no pass or fail limit. The useful comparison is with the concrete cover at the same spot. In a fictitious column, three specimens read 18, 22 and 26 mm, so the member mean is 22 mm. If the measured cover there is 25 mm, the front is 3 mm short of the steel. If the mean were 28 mm, the front would have passed the steel, and the half-cell criteria would not normally apply at that spot (Section 2, 6.1.1). Either way, carbonated surface concrete makes rebound numbers read high (Section 4, 7.1.5), which is why that standard asks for the carbonation depth in the rebound report.

What should an NDT test report contain?

Everything that ISO/IEC 17025 clause 7.8.2.1 asks of any test report, which our guide to the NABL test report format covers item by item, plus the items the IS 516 sections and NABL add for work on a structure.

A rebound hammer and UPV report format

A results table like this covers what IS 516 asks for at each test point:

Put the individual readings, anvil and reference bar checks, half-cell maps and photographs in annexures. Keep the instrument calibration certificates and before-and-after checks with the job record. Our guide to instrument calibration in a testing lab covers calibration intervals and intermediate checks.

NABL symbol rules for NDT reports

If you hold NABL accreditation for these tests, NABL 133 (Issue 09, Amendment 03, 3 September 2024) applies:

• the NABL symbol is mandatory on reports for the tests in your accredited scope, and a narrative reference is not acceptable in its place (5.1);
• tests outside your scope, for example a cover meter survey you are not accredited for, go on a separate report;
• an asterisk or any other mark to flag non-accredited results is not permitted (5.1, 6.2);
• the report must be authorised by a person declared to NABL for review, reporting and authorisation of results (6.3).

Opinions about the structure, such as "column C3 needs jacketing", are not test results. ISO/IEC 17025 clause 7.8.7 requires them to be released only by authorised people, with the basis documented and clearly identified as opinions.

Why does concrete NDT come under NABL 130?

Because the test happens where the structure is. NABL 130 (Issue 06, 6 February 2020) covers tests done at the customer's premises or away from the permanent lab. A lab seeking accreditation for site NDT must identify its site tests separately when it applies (Appendix A), and meet NABL 130 as well as ISO/IEC 17025.

What NABL 130 adds for an NDT team:

• A separate list of site tests, with documented site procedures and arrangements for supervising site work (4.1.1 to 4.1.4).
• Equipment checks before and after. Verify each instrument before it leaves the lab and again when it comes back (5.3.2). Keep the records. If an instrument is found unfit or out of calibration, withdraw it and examine its effect on earlier tests under your nonconforming work procedure (5.3.3).
• More frequent calibration and intermediate checks for site equipment, because it is moved and used in unstable conditions (5.3.4). Plan for safe packing and transport (5.3.5, 5.3.7).
• Your own people and equipment. Site testing is done by lab personnel only, and the lab must not use major equipment other than its own (5.1.2, 5.3.1). Name who is authorised for each site test and instrument (5.3.6).
• Environment. Record site conditions that may affect results, consider them in measurement uncertainty, and state them on the report (5.2.1, 5.2.2).
• Reports. Identify site results clearly, with the location where the test was done (5.5.2).
• Audits. Internal auditors must visit sites as part of the audit, and data held at site must be kept secure (4.1.9, 4.1.11).

The checks in the IS sections give you ready-made evidence for this: the anvil test before and after a rebound hammer job (Sec 4, clause 5.1) and the reference bar check for UPV (Sec 1, clause 2.3.1).

How do NDT labs lose time with an Excel sheet per job?

Most NDT teams run each job from a workbook copied from the last one. It works until the jobs get bigger or an assessor asks a question.

Readings typed twice. Site readings go into a notebook, then into Excel back at the office. One structural column tested at a dozen points, with six impacts at each, is 72 readings before any UPV. ISO/IEC 17025 clause 7.5.1 asks for original observations to be recorded at the time they are made, and clause 7.11.6 asks for calculations and data transfers to be checked. Every retyping step adds checking work.

Old tables carried forward. A copied workbook brings its UPV grading table with it. If that table was set up before Amendment No. 1, every job since has been graded against the wrong limits. Concrete up to M25 reading 3.50 to 3.75 km/s has been called doubtful when the current table says good. Nobody notices until a customer or assessor compares it with the standard.

Strength printed without a basis. The core correlation sits in another file, so it is quicker to read strength off the hammer's own curve. IS 516 (Part 5/Sec 4) clause 5.2 allows that curve only for relative strength, and clause 8.3 rules out any estimate where UPV is doubtful. A spreadsheet will not stop either mistake.

Equipment checks that are not linked to jobs. Anvil and reference bar readings live on paper or in a separate log. When a hammer fails its check at the end of a job, NABL 130 clause 5.3.3 requires you to look at earlier results. With one workbook per job, finding every job that hammer touched since its last good check means opening folder after folder.

Reports assembled by hand. The engineer copies tables into Word, adds the half-cell percentages worked out on a calculator, pastes photographs, and splits out the tests not in the accredited scope. Each step is a chance for a member ID or a percentage to go wrong.

Each of these looks small until an assessor or a customer finds it. What a lab needs is a fast tool that takes the readings once, calculates what the standard defines, and keeps every job's data in one place. For the wider picture, see the hidden cost of Excel in labs.