Experimental Study and Wavelet Neural Network Prediction of Unconfined Compressive Strength of Acrylate Copolymer-Emulsion-Cured Saline Soil

Imagine you're trying to build a super-strong foundation for a road, but the ground you're standing on is a salty, soggy mess. In coastal areas like Jiangsu, China, this "saline soil" is a troublemaker. It swells up when it gets wet, shrinks when it dries, and can even freeze and heave, cracking roads and buildings like a bad case of the shivers.

For years, engineers tried to fix this with old-school materials like cement or lime. But the authors of this study say those methods are a bit like using a sledgehammer to crack a nut: they can be messy, bad for the environment, and sometimes just don't stick well enough. So, they decided to try something new: a special liquid glue made from an acrylate copolymer emulsion (let's call it ZM). Think of ZM as a magical, stretchy polymer blanket that wraps around soil particles, holding them together.

The Big Experiment: Mixing and Matching

The researchers took soil from a construction site in Nantong and mixed it with salt to simulate the real-world problem. Then, they played a game of "mix and match" to see what made the soil strongest. They changed four things:

1. How much ZM glue they added (from 0% up to 1.5%).
2. How long they let it cure (from day 0 all the way to day 28).
3. How salty the soil was (from fresh water levels up to super-salty 9%).
4. How wet the soil was (testing moisture levels around the "sweet spot").

After mixing, they squished the soil samples in a machine to measure their Unconfined Compressive Strength (UCS). Think of UCS as the "squeeze test": how much pressure can the soil take before it crumbles?

What They Found (The Good, The Bad, and The Salty)

The Glue Works Wonders:
When they added the ZM glue, the soil got significantly stronger. The more glue they used (up to 1.5%) and the longer they waited (up to 28 days), the harder the soil became. By day 28, even the samples with the least glue were strong enough to pass the official test for "effective stabilization." The soil that didn't get any glue? It stayed weak, no matter how long they waited.

The Salt Spoils the Party:
Here's the tricky part: salt is the enemy. As the salt content went up, the strength went down. If the salt was low (under 0.4%), the glue still worked great. But once the salt hit 3% or higher, the strength started to drop fast. At super-high salt levels (6–9%), the soil got so weak that the glue struggled to hold it together. The salt crystals act like tiny wedges, prying the soil apart and creating cracks.

The Moisture Sweet Spot:
The soil also needed the right amount of water. If it was too dry or too wet, the strength dropped. The magic happened when the moisture was just slightly above the "optimal" level (about 0–5% higher). It's like baking a cake: too little water makes it crumbly; too much makes it a soggy mess. You need that perfect, slightly damp consistency for the glue to work its best.

The Failure Mode:
When the soil finally broke under pressure, it didn't just crumble slowly; it acted like a "semi-brittle" material. It held strong, then suddenly gave way, showing a "strain-softening" failure. This means once it started cracking, it got weaker and weaker very quickly.

The Microscopic Mystery: No Magic Spells, Just Physics

The researchers looked at the soil under powerful microscopes and used infrared spectroscopy (a way of checking the chemical "fingerprints" of the material). They were looking for a chemical reaction that created new minerals or magical new bonds.

The Verdict: Nope. No new minerals were born. No new chemical groups appeared.
Instead, the ZM glue worked like a physical net. The long polymer chains in the glue wrapped around the soil particles, filling in the empty holes (pores) and sticking them together. It's like using a sticky tape to bind a pile of loose marbles into a solid block. The glue didn't change the marbles themselves; it just held them tight.

The Crystal Ball: Predicting the Future with AI

Here's where it gets really cool. The strength of this soil is messy and unpredictable. It doesn't follow a simple straight line. If you tried to guess the strength using old-school math formulas, you'd probably get it wrong because the factors (salt, water, time, glue) interact in a complicated, chaotic way.

So, the authors built a Wavelet Neural Network.

• What is it? Imagine a super-smart student who is really good at spotting patterns in messy data. Unlike a regular calculator that follows strict rules, this AI "student" uses something called "wavelets" (which are like tiny, flexible waves) to zoom in on the details of the data.
• The Upgrade: The authors didn't just use the standard version; they tweaked it. They gave the AI two different "weights" to pay attention to and taught it to adjust its own focus (using a method called gradient descent) to learn faster and better.
• The Result: They fed the AI all their data (glue amount, time, salt, water) and asked it to predict the strength. The AI got it right! When they tested it on real-world data from a road project, the AI's predictions were within 10% of the actual measured strength.

The Bottom Line

This study shows that this special polymer glue (ZM) is a great way to fix salty soil, provided you keep the salt levels manageable and the moisture just right. It works by physically wrapping and binding the soil, not by changing its chemistry. And because the soil's behavior is so tricky, using a smart, upgraded AI model is the best way to predict how strong the ground will be before you even start building.

The authors are confident that this AI tool can help engineers in coastal areas make safer, smarter decisions, but they emphasize that this is based on their specific experiments and simulations. It's a powerful new tool, but it's not a magic wand that works for every single situation without testing.