IS 5816 : 1999Method of Test Splitting Tensile Strength of Concrete

IS 5816 specifies the method for splitting tensile strength of concrete (the 'Brazilian test') — a cylinder is loaded across its diameter until it splits, and the indirect tensile strength is calculated from the failure load.

Unlike compressive cube strength (which dominates RCC design), tensile strength is what governs: - Cracking in slabs, water-tank walls, mass concrete (thermal cracking) - Pavement design (IRC:58 for rigid pavements) - Punching shear capacity in flat slabs around columns - Anchorage / bond length calculations indirectly - Durability assessment — tensile strength correlates with crack-control limits in IS 456 Clause 35.3.2

IS 5816 is run when: - Pavement-quality concrete (PQC) for highways or airports — tensile strength is the design variable - Water-retaining structure design (IS 3370) — crack-control checks - Investigation of cracking in completed structures (cores tested in split tension) - Mix-design qualification for high-strength or fibre-reinforced concrete

For most building structural design, the assumed tensile strength is taken from IS 456 Clause 6.2.2 (f_ct = 0.7 √f_ck). IS 5816 is what you run when you need to *measure* it, not assume it.

1. Specimen: cylinder 150 mm diameter × 300 mm length (standard), or 100 × 200 mm (small-aggregate concrete). 2. Preparation: cured 28 days in water at 27 ± 2 °C. Surface dry before test. Mark two diametrically opposite generators along the length. 3. Loading: cylinder placed horizontally between platens of a compression testing machine, with two plywood strips (3 mm thick × 25 mm wide × 300 mm long) cushioning the line of contact along the marked generators. 4. Rate: load applied continuously at 1.2-2.4 N/mm² per minute on the cylinder cross-section. 5. Failure: cylinder splits along the loaded diameter. Record peak load P (kN). 6. Calculation:

`f_ct = 2P / (π × L × D)`

where P = peak load (N), L = cylinder length (mm), D = cylinder diameter (mm). Result in N/mm².

Report one f_ct per cylinder; mean of three is the batch result.

Typical f_ct vs f_ck (28-day):

| Concrete grade | f_ck (cube) | f_ct typical | f_ct / f_ck ratio | |---|---|---|---| | M20 | 20 MPa | 2.5-3.0 MPa | ~12-15 % | | M25 | 25 MPa | 2.8-3.4 MPa | ~11-14 % | | M30 | 30 MPa | 3.1-3.7 MPa | ~10-12 % | | M40 | 40 MPa | 3.6-4.4 MPa | ~9-11 % | | M50 | 50 MPa | 4.0-4.8 MPa | ~8-10 % | | M60 | 60 MPa | 4.4-5.2 MPa | ~7-9 % |

Comparison with IS 456 Clause 6.2.2 estimate (`f_ct ≈ 0.7 √f_ck`): - M25: 0.7 × √25 = 3.5 MPa → matches measured upper bound - M40: 0.7 × √40 = 4.4 MPa → matches measured upper bound

The formula slightly over-predicts for higher grades (>M50). Always verify by test for high-strength mixes.

Acceptance for PQC (IRC:58): - M40 PQC: 28-day flexural strength ≥ 4.5 MPa (from beam test) — split tensile is the secondary verification, target ~3.5 MPa. - Report at 7-day and 28-day; the 7-day flexural is typically 75-80 % of 28-day.

• IS 456:2000 Clause 6.2.2 — design tensile strength (the assumed value when you don't test).
• IS 516 Part 1:2021 — compressive strength on cubes (the most common companion test).
• IS 516 Part 5 / IS 9013 — flexural strength on beam (the *third* tensile test method; used for PQC).
• IS 10262:2019 — mix design (high-strength mixes warrant tested f_ct).
• IS 1199 Part 1:2018 — fresh concrete sampling for the cylinder casting.
• IS 3370 Part 2:2009 — water-retaining-structure design (crack control depends on tested tensile strength for serviceability).
• IRC:58:2015 — rigid pavement design (PQC tensile + flexural).
• IS 13311 Part 1:1992 — UPV and rebound tests (in-situ correlation with split tension on cores).

1. Using cubes instead of cylinders. The split test is geometry-dependent. Cube split-test results do not equal cylinder split-test results — the formula assumes cylinder geometry. 2. Skipping the plywood cushion strips. Direct steel-on-concrete contact creates stress concentrations that cause premature failure. Strips must be 3 mm × 25 mm × cylinder length, replaced when crushed. 3. Eccentric specimen positioning. Off-axis loading causes a tensile + bending state, not pure split — invalidates the test. Use a marking jig or a centring plate. 4. Loading rate too fast. Above 2.4 N/mm² per minute the apparent tensile strength rises (rate-dependent material response). Calibrate the testing machine for the slow ramp. 5. Comparing test result against IS 456 estimate without realising the formula has 30-40 % scatter. The 0.7 √f_ck rule is a design-side conservative estimate. Real f_ct can be 80-130 % of the formula value. 6. Casting the cylinder vertically and forgetting to invert the generator marking. The two cushion strips must be along generators, not at the cast top/bottom (which has lower strength due to bleed water). Mark before testing. 7. Testing 7-day cylinders and rejecting. PQC and most contracts accept on 28-day. 7-day result is for *prediction*. Don't reject on 7-day alone.

Assume from compressive strength (use IS 456 Clause 6.2.2 f_ct = 0.7 √f_ck) for routine RCC building design. This is the default in column / beam capacity calculations.

Test in split tension via IS 5816 when: - Pavement (PQC) — tensile is the design variable, not compressive - Water-retaining structure — crack-width SLS check needs accurate tensile - High-strength concrete (M50+) — formula increasingly conservative; test gives more economy - Forensic investigation of cracking — tested f_ct gives evidence of mix quality - New aggregate source qualification — split tension verifies aggregate-paste bond - Fibre-reinforced concrete — fibres dramatically alter tensile response

Test in flexure via IS 516 Part 5 / IS 9013 when: - Pavement design (typically the *primary* tensile test for PQC, with split tension as backup) - Modulus of rupture is needed (different value than split tension)

Split tension is faster and uses standard cylinder moulds you already have. Flexure is more representative of slab loading but needs longer beams (700 × 150 × 150 mm).