Hangingwall Dilution and a Post-Blast Fall in a 60 m Open Stope

Four credible causes, and support is not the most likely leader.

1. Geometry. The stability graph is the strongest signal here. N′ ≈ 12 with a hydraulic radius of 22 m sits in the transition zone, meaning the design itself is on the edge of instability regardless of bolting. A 60 m high, 40 m long hangingwall is a large exposed surface; the graph is telling you the exposure is the problem.
2. Structure. A 4 m altered fault 8 m behind the contact is a classic release surface. If the joint sets daylight into the fault, a wedge is kinematically free, and no pattern of short bolts arrests it. Overbreak will concentrate at one end of the hangingwall, following the fault.
3. Blast damage. Longhole blasts with 5 m burden can generate a damage zone of 1–2 m in a foliated rock mass, which reduces the effective N′ and detaches the first metre of wall.
4. Seismicity. The ML 2.1 event 20 minutes after the blast is not blast vibration - vibration attenuates in milliseconds. Twenty minutes is stress redistribution and slip on the fault, i.e. a rockburst mechanism.

Discriminating data.

• Cavity monitoring survey of each stope: is the overbreak on the hangingwall, the crown, or scattered? Fault-parallel concentration points to structure.
• Mapped joint orientations from scan lines compared with the fallen wedge geometry for kinematic admissibility.
• Blast vibration records and half-barrel counts on the wall to size the blast damage zone.
• Event locations and magnitudes against stope and fault geometry, plus apparent stress, to test whether seismicity tracks extraction.
• Core logging (RQD, alteration) behind the wall and a back-calculation of N′ from the observed stable spans in adjacent stopes.

Without these, the manager's support hypothesis is unfalsifiable and any pattern change is a guess.

Mechanics. A 6 m ungrouted split set anchors at the collar and generates friction along its length. In a hangingwall where the loosening or blast-damaged zone can be 2–4 m and the fault sits 8 m back, a 6 m bolt cannot reach stable ground. It acts as surface restraint and beam-building, which has value against small blocks but cannot hold a 12 m wedge. At 1.5 m spacing it also provides roughly one bolt per 2.25 m², which is a dense, cheap, surface-focused pattern.

A 12 m grouted twin-strand cable anchors beyond the yield zone into intact rock and carries load in tension over its full grouted length, which is the correct mechanism for a wedge releasing on a fault. At 2.5 m × 2.5 m spacing that is one cable per 6.25 m² - roughly 16 cables per 100 m², or about 380 cables over a 2,400 m² hangingwall. Unit cost is several times a split set, but the failure mode it addresses is the one actually occurring.

Judgement. Neither pattern addresses the root cause by itself. The stability graph says hydraulic radius drives stability, so splitting the stope into two 30 m lifts is the highest-leverage change: it roughly halves the exposed hangingwall area per lift and lets paste fill support the wall between lifts. That single change may move the design from the transition zone back into the stable field, after which a lighter cable pattern plus mesh and split sets for surface retention is defensible.

Recommended package. Split the lifts and paste fill as the primary control; install 12 m cables at 2.5 m × 2.5 m on the hangingwall of the upper lift where the fault is closest; retain split sets and mesh for local block control; and trim blast design (reduced charge concentration at the wall, pre-split or trim holes) so the damage zone does not consume the anchor length you paid for. Cost per tonne should be compared against the 10 points of excess dilution, which is worth far more than the support difference.

Interpretation. A 20-minute delay places the event in the stress-redistribution domain rather than the blast-vibration domain. The stope excavation and the preceding blast changed the stress state on the fault, slip occurred, and the hangingwall wedge was released. That has two consequences: larger blasts create larger stress steps, and the wall is most dangerous in the hours after firing, not during it.

Operational changes.

• Fill primary stopes before mining adjacent secondary stopes; the fill is both wall support and a stress buffer.
• Avoid mining two adjacent stopes on the same shift. Sequence so that no two active faces load the same fault segment simultaneously.
• Reduce blast size at the wall and use trim or pre-split holes to lower the stress step and the damage zone.
• Scale and wash the hangingwall after every blast before any work returns to the brow.
• Track the array: event rate, magnitude and apparent stress. A rising b-value or clustering near the fault should trigger a review of the sequence, not a bolting change.

Re-entry rule. Do not tie re-entry to a fixed clock. Require all of: (a) the seismic array quiet for a defined window after the blast, defined from the site's own event-decay statistics rather than a round number; (b) a physical inspection and scaling of the hangingwall by a competent person; (c) confirmation from the cavity monitoring survey that overbreak has not extended since the last survey. If the array shows continuing events near the stope, keep the level closed and reassess the sequence. Write the trigger levels into the ground control management plan and brief the crews on why the 20-minute delay matters - a crew that understands the mechanism will report small rockfalls that would otherwise go unlogged.