Shrinkage porosity in a steel plate casting: riser modulus versus feed metal
A foundry casts a plain carbon steel plate 400 × 300 × 40 mm in a sand mould, one plate per mould, fed by a single cylindrical top riser of D = 70 mm and H = 70 mm. The moulding box limits any riser to 100 mm height. Chvorinov's constant for steel in this sand is 3.5 min/cm² with n = 2, and freezing shrinkage is taken as 3% by volume, all of which the riser must supply. After the first batch, ultrasonic testing finds centre-line shrinkage porosity in the plate directly under the riser. When a riser from a scrapped casting is sectioned, it is fully dense with no pipe and no porosity at all. The foundry has exothermic riser sleeves in stores which, for this size, effectively raise the riser's Chvorinov constant from 3.5 to 5.0 min/cm². A dimensional check of the pattern and the moulding sand moisture both came back within specification.
1. Using the modulus method, show numerically whether the existing 70 mm riser can feed the plate, and explain how the two observed symptoms fit your conclusion.
Read the worked answer
Plate: V = 40 × 30 × 4 = 4 800 cm³; A = 2(40×30 + 40×4 + 30×4) = 2(1 200 + 160 + 120) = 2 960 cm². M_c = 4 800/2 960 = 1.622 cm; t_s = 3.5 × 1.622² = 9.20 min.
Riser (H = D, base in contact with the casting, so the curved surface plus the top only): V = πD³/4, A = πD² + πD²/4 = 1.25πD², hence M_r = D/5 = 70/5 = 1.40 cm; t_s = 3.5 × 1.40² = 6.86 min.
The riser freezes 2.3 min, i.e. about 25%, before the plate. The usual requirement is M_r ≥ 1.2 M_c = 1.946 cm here, so M_r/M_c = 0.86 against a minimum of 1.2.
Symptoms: a riser that feeds properly must itself shrink and therefore ends with a pipe; a fully dense riser is proof that it solidified while still full, i.e. before the plate's thermal centre. The plate then solidified with no liquid reservoir, so the last metal to freeze under the riser had nowhere to draw from and tore into centre-line porosity in exactly that location. Both observations are one single fault, not two.
2. Size a replacement riser. Evaluate the bare-riser option and the exothermic-sleeve option, and check the feed-metal volume as well as the modulus.
Read the worked answer
Modulus requirement: M_r ≥ 1.2 × 1.622 = 1.946 cm → with M_r = D/5, D ≥ 9.73 cm, i.e. a 100 mm standard riser (H = D = 100 mm, at the box limit).
Bare 100 mm riser: t_s = 3.5 × (100/5/10)² = 3.5 × 2.0² = 14.0 min > 11.04 min required, so the modulus criterion is met. Volume = π × 10³/4 = 785 cm³; liquid demand = 3% of 4 800 = 144 cm³ for the plate plus 3% of 785 = 23.6 cm³ for the riser's own shrinkage, i.e. 168 cm³ or 21% of the riser. A H = D riser pipes deeply, so 21% is workable but not generous; H/D = 1.5 would be safer if the box allowed.
Sleeve only, modulus route: need 5.0 × (0.2D)² ≥ 11.04 → 0.2D² = 11.04 → D ≥ 7.43 cm, i.e. the existing 70 mm riser almost works. But at 74 mm the riser volume is only 322 cm³, and the demand of 144 + 9.7 = 153.7 cm³ is 48% of it. A riser cannot surrender half its volume as feed metal; the sleeve fixes the heat loss, not the volume deficit.
Recommendation: 100 mm diameter × 100 mm high with the exothermic sleeve (t_s = 5.0 × 2.0² = 20 min, demand 21% of 785 cm³), plus a short neck of at least 60 mm diameter so the neck does not freeze before the riser. Do not accept the sleeve on the old 70 mm riser on the strength of the modulus calculation alone.
3. The moulder argues that the riser was fine because it was 'full of metal' and blames the gating system. What would you check first, and what would you change in the gating and necking practice?
Read the worked answer
'Full of metal' is exactly the failure signature, not a defence: a sound riser should end with a pipe. The sectioned riser confirms the modulus calculation, so the primary defect is riser size and heat loss, not gating.
Things worth checking in order:
- Neck. If the neck is small or long, it freezes early and chokes feeding even when the riser is big enough. Check its cross-section (target ≥ 0.6D, kept short) and its modulus, which should exceed the plate's.
- Where the hot metal enters. The riser must be the last region to freeze; if the ingate feeds the plate directly and the riser is filled through a cold neck, the riser starts cold.
- Top surface condition. An open riser top loses heat by radiation; check whether a topping compound or sleeve is applied, and whether the mould wash or sand moisture at the riser is out of line.
- Feed distance. Confirm the porosity is at the plate's thermal centre, not at a far edge that the riser could never reach; if it is far away, chills or a second riser are needed.
Changes: exothermic sleeve plus topping compound on a 100 mm riser, a short neck of at least 60 mm diameter, and gating designed to bring hot metal in through the riser (or a riser-fed arrangement) so the riser is the last to freeze. Verify with a sectioned trial casting before releasing the batch.