Subsidence affecting 15 terraced homes in Dublin

Solution

An expansive geopolymer resin was injected through small‑diameter tubes at multiple levels within the made ground. The resin permeates and compacts the surrounding fill as it expands, closing voids and densifying loose material, then cures to a rigid polyurethane.

Seven distinct element types were identified across the terrace, and each was designed separately against its own target bearing pressure. A rear wall foundation carrying 125 kPa and a garden area carrying 25 kPa are different engineering problems; treating them to a common specification would mean over‑treating most of the site and still leaving open the question of whether the critical elements were adequate.

Treatment was placed at several discrete levels through the fill so that the layer was improved as a body rather than grouted at a single horizon. The number of levels, their spacing and the quantity placed at each are set by Geobear's own design method and are specific to the loading and ground conditions at each element.

Sequence of works

Each element followed the same controlled sequence:

1. Pre‑test. Every element probed before any resin was placed, establishing the baseline profile.
2. Design review against the pre‑test. Engineering confirmed whether treatment depth could be reduced in light of the actual profile. Where the fill was shallower than assumed, the design was cut back rather than delivered as drawn.
3. Install injection tubes block by block, in line with the programme.
4. Inject under continuous level monitoring.
5. Cure for 24 hours before any post‑treatment testing.
6. Post‑test and sign off, or re‑inject and repeat.

Step two matters more than it looks. Treatment depth was set by measured ground conditions at each property, not by a single assumption applied across the terrace — which is what stops a design like this from becoming an expensive blanket application.

Results

Verification used medium‑weight dynamic probing (DPM): a 30 kg hammer dropped 200 mm, counting the blows required to advance the rods each 100 mm, logged against depth. Every treated location was probed before injection and again afterwards, so each property carries a matched pair of profiles rather than a single post‑works reading.

A Material Quality Test was carried out on the resin prior to injection at each property, and passed.

Below: Dynamic probe profile at the main terrace rear wall, before and after treatment. Before injection the probe recorded 0–1 blows per 100 mm through almost five metres of fill — effectively free‑falling — until it met boulder clay at 4.75 m. After treatment the same location returned 8–22 blows per 100 mm throughout, refusing at 3.5 m.

Verification across the terrace

Validation testing was not sampled. Every treated element at every property carried its own matched pre‑ and post‑treatment probe, giving 94 test locations across the works — between two and seven per property depending on how many element types that property contained. Each was tested against the specific bearing pressure required for that element, and each had to be signed off individually before the property was released for follow‑on above‑ground repairs.

A worked example

At the property set out below, six locations were probed, one for each structural element in the treatment scope. Every location met or exceeded its target allowable bearing pressure at every depth within the design horizon.

Outcome

• Every property stabilised across 50 shifts in ten weeks, delivered by a probing team, one to two drilling teams and an injection crew working to a fixed 7am–5pm weekday and Saturday‑morning window agreed as a planning condition.
• 94 validation tests across the terrace, every one signed off against its own design requirement.
• No resident relocated. Households retained full use of their homes throughout.
• No excavation, no spoil, no heavy plant in a dense residential street.
• Every treated element proof‑tested against a stated bearing requirement and signed off by a chartered engineer, providing the evidence base for follow‑on above‑ground repairs.
• Up to 34% lower embodied carbon than excavation‑based alternatives.

Cured geopolymer requires no cleaning or maintenance, carries no residual hazard, and if ever excavated is classified as inert waste under EWC 17 06 04.