Process Simulation and Reverse Optimization of Modified Biochar Adsorption of Salt and Minerals in Water: Prediction Model of Adsorption Efficiency Based on BP Neural Network

Imagine the world's water supply as a giant, bustling city. Sometimes, this city gets a little too crowded with invisible guests: salt and mineral ions like sodium, calcium, and magnesium. When there are too of these guests, the water becomes "high-salinity," which is bad news for farming, drinking, and the environment. Cleaning this water is like trying to find a needle in a haystack, but instead of a needle, you are looking for specific invisible ions. Scientists have been trying to build better "sponges" to soak up these unwanted guests. One popular type of sponge is biochar, which is basically charcoal made from plant waste like corn stalks. It's cheap and has a lot of tiny holes to catch things. However, plain biochar isn't always strong enough. So, scientists try to "modify" it—giving it a chemical makeover to make it stickier or more attractive to the salt.

But here is the tricky part: figuring out exactly how to use these sponges is like trying to bake the perfect cake without a recipe. You have to guess how much sponge to use, what temperature the water should be, how long to wait, and what the water's acidity (pH) should be. If you guess wrong, the sponge might not work, or you might waste money. This is where artificial intelligence steps in. Think of AI as a super-smart assistant that can look at thousands of past baking attempts and predict the perfect recipe before you even turn on the oven. In this story, the "baking" is cleaning water, and the "recipe" is the perfect mix of sponge type, amount, and conditions.

---

The Super-Sponge and the Smart Assistant

In this study, a researcher named XiYue Gao decided to tackle the problem of salty water using a three-part strategy: build a better sponge, teach a computer to predict how it works, and then ask that computer to find the perfect recipe.

Step 1: Building the Sponges
The team started with corn stalks from a farm in Heze, China. They turned these stalks into three different types of "sponges" (biochar):

1. Plain Biochar (BC): Just the basic charcoal sponge.
2. Acid-Modified (SBC): The sponge was given a sulfuric acid bath to change its surface.
3. Iron-Modified (FeBC): The sponge was loaded with iron, like giving it a magnetic superpower.

They tested these sponges in real farm drainage water, which was full of salt and minerals. The results were clear: the Iron-Modified sponge (FeBC) was the star of the show. Why? Because it had a much larger surface area (315.2 m²/g, which is like having a massive, crumbly sponge compared to a smooth one) and was loaded with iron and special chemical groups that loved to grab onto salt ions. It was simply the stickiest sponge of the bunch.

Step 2: The Smart Assistant (The BP Neural Network)
Next, the team needed to figure out the best way to use these sponges. Should they use a little bit or a lot? Should the water be hot or cold? Should they wait 10 minutes or 4 hours? Instead of running thousands of messy experiments, they built a BP Neural Network.

Think of this neural network as a very hungry student who has read every single lab report from the past. The student was fed five pieces of information for every experiment:

• Which sponge was used?
• How much sponge?
• How acidic was the water?
• How hot was the water?
• How long did they wait?

The student's job was to guess the adsorption efficiency (how much salt was removed). After studying 216 different experiments, this "student" became incredibly good at guessing. It got the answer right 93.5% of the time during its practice tests and 93.0% of the time on new, unseen tests. The difference between what it guessed and what actually happened was tiny—less than 1.2%. This means the computer model was a reliable crystal ball for this specific type of water.

Step 3: The Reverse Search (Particle Swarm Optimization)
Now came the fun part: Reverse Optimization. Usually, scientists try a condition, see what happens, and try again. But here, they asked the computer: "If we want the absolute best result, what should the recipe be?"

They used an algorithm called Particle Swarm Optimization (PSO). Imagine a flock of birds searching for the highest point in a mountain range. Each bird (or "particle") flies around, checking the height (or in this case, the salt-removal efficiency). If a bird finds a high spot, it tells the others. Eventually, the whole flock converges on the very highest peak.

In this study, the "flock" flew through the computer model and found the perfect combination for the Iron-Modified sponge:

• Amount: 0.8 grams per liter of water.
• Acidity (pH): 7.5 (slightly neutral).
• Temperature: 35°C (a warm day).
• Time: 180 minutes (3 hours).

The Grand Result
When the team actually tested this "perfect recipe" in the lab, it worked almost exactly as the computer predicted. The sponge removed 91.8% of the salt and minerals. Even better, when they looked at the cost, the Iron sponge was the most efficient. It cost about 28 yuan per kilogram to make, but because it worked so well, the cost to clean a ton of water dropped to just 1.6 yuan, and the "cost-benefit ratio" (how much cleaning you get for your money) hit 57.4% per yuan.

What This Means (and What It Doesn't)
The paper suggests that combining a chemically modified sponge with a smart computer model is a powerful way to clean salty water. The Iron sponge is definitely the best performer among the three they tested, and the computer model is accurate enough to predict the results with high confidence.

However, the author is careful not to say this is a magic bullet for the whole world yet. They note that this was done in a lab with a specific type of farm water. They haven't proven exactly how the iron grabs the salt at a microscopic level (they suspect it's a mix of physical trapping and chemical attraction, but they need more advanced tools to be 100% sure). Also, while the cost looks great on paper, they haven't tested it in a giant real-world factory yet.

But for now, this study offers a very promising recipe: take corn stalks, give them an iron makeover, and use a smart computer to find the perfect temperature and timing. It's a playful, data-driven way to turn farm waste into a water-cleaning superhero.