When shallow soil won’t cooperate, you have more options than you think — but choosing the wrong one costs time and money.
Not every wind farm sits on good ground. Several of the projects I’ve worked on had some geotechnical issue that made standard gravity foundations impractical without treating the soil first.
Maybe it’s loose sand prone to liquefaction, or soft clay that would settle unevenly under turbine cyclic loads. The question quickly becomes: what kind of ground improvement do we use, and how deep does the problem go?
I’ve seen the right treatment save hundreds of thousands of euros per turbine compared to piled foundations. I’ve also seen the wrong choice lead to months of delay. So let me walk through the six techniques I encounter most, split into shallow and deep, with real numbers and honest limitations.
Shallow treatments
These are what you reach for when the problem is close to the surface — within the first 3 to 4 metres. Less specialised equipment, generally cheaper, but hard depth limits.
Excavation and replacement is the most straightforward: dig out the bad soil and replace it with compacted granular fill. I’ve used it on dozens of projects and it remains the most economical option when conditions allow. Any competent earthworks crew can do it. The practical maximum depth is around 3 to 4 metres (beyond that, volumes become uneconomical and pit stability gets problematic). A high water table makes life difficult — you’re fighting dewatering the whole time.
Soil-cement and lime stabilisation works well for weak cohesive soils. Mix a binding agent into the existing ground: around 5% lime by mass for plastic clays, or a minimum 18% Portland cement by volume for mixed soils, with at least a 20-centimetre treated layer. The cost is low and the equipment is standard (a soil mixer or rotavator, basically). But lime only works in cohesive soils — throw it at a clean sand and you’re wasting your money.
Deep treatments
When the problem extends to 5, 10, or 20 metres depth, you need techniques that reach down without excavating. Specialist equipment and contractors required, but they open up sites that would otherwise need expensive piling.
Dynamic compaction is hard to miss on site. A crane lifts a weight of up to 200 tonnes and drops it from heights of up to 40 metres. The impact densifies soil through shockwaves — dramatic, loud, and surprisingly effective (the neighbours tend to notice). It works very well in granular soils and is one of the few methods that directly addresses liquefaction risk by increasing relative density.
Depth of influence reaches 10 to 12 metres. Less effective in cohesive soils, where the energy dissipates without lasting densification.
Vibroflotation uses a vibrating probe lowered into the ground, assisted by water jetting. The vibration rearranges granular particles into a denser configuration. Very effective in clean sands and gravels with low fines content (less than 10 to 15% fines).
Gravel columns use similar equipment but feed gravel into the hole as the probe is withdrawn. The result is a grid of stiff, free-draining columns that reinforce the soil and — critically in saturated ground — provide drainage paths that reduce excess pore water pressure. This dual function makes them useful for a wider range of soils than vibroflotation, including mixed soils with some cohesive content. The natural ground does need enough lateral consistency to confine the columns (in very soft clay below about 15 kPa undrained shear strength, they bulge and lose effectiveness). A wind farm near Bristol in the UK used gravel columns at some locations, rigid inclusions at others, and a combination at a few — that kind of mixed treatment is more common than textbooks suggest.
Controlled Modulus Columns (CMC), also called rigid inclusions, are low-strength mortar columns installed by a displacement auger. The auger screws into the ground without extracting soil, then mortar is pumped as it withdraws. The result is a semi-rigid column that transfers load to a competent bearing stratum at depth. CMCs are not piles in the traditional sense (mortar strength is typically 5 to 10 MPa versus 25+ MPa for concrete piles), but they reduce settlement effectively. They need a firm layer at the tip — no stratum, no CMC. And unlike gravel columns or dynamic compaction, they do not address liquefaction. A wind farm in central France used CMCs reaching about 8 metres to a stiff clay beneath loose alluvial deposits.
Bonus: Deep Soil Mixing
Worth mentioning: Deep Soil Mixing (DSM) mixes cement slurry directly into the soil in-situ, creating soil-cement columns. Competitive in sensitive clays where other methods struggle. I’ve seen it on a wind farm in Poland with thick glacial clay deposits. Limited soil types and specialised equipment, but excellent when conditions suit it.
Comparison at a glance
| Technique | Best soil type | Max depth | Liquefaction? | Relative cost | Specialist needed? |
|---|---|---|---|---|---|
| Excavation & replacement | Any (shallow) | 3-4 m | No | Low | No |
| Cement/lime stabilisation | Cohesive (lime), mixed (cement) | 1-2 m | No | Low | No |
| Dynamic compaction | Granular | 10-12 m | Yes | Medium | Yes |
| Vibroflotation | Clean granular (<15% fines) | 15-20 m | Yes | Medium-High | Yes |
| Gravel columns | Mixed, saturated | 15-20 m | Yes (drainage) | Medium-High | Yes |
| CMC / rigid inclusions | Any (needs bearing stratum) | 15-25 m | No | High | Yes |
Choosing the right treatment
The decision is never purely technical. You’re balancing soil type, depth of the problem, water table, liquefaction risk, budget, and contractor availability in the region (in some countries, finding a vibro rig is straightforward; in others, you’re importing equipment from abroad, which changes the economics completely).
A mistake I see is jumping to deep treatments when a shallow solution would have worked (or the reverse, trying to save money with a shallow fix when the problem clearly extends to depth).
A decent geotechnical investigation with CPTs or SPTs to the right depth pays for itself many times over. And don’t be surprised if the final solution is a combination of techniques. Real sites have variable ground, and what works at one turbine location might not work 500 metres away. That’s the nature of the job.
Leave a Reply