Application of Particle Swarm Optimisation to achieve simultaneous nitrification and denitrification removal in the aerobic biological treatment process

Imagine a city's wastewater treatment plant as a giant, busy kitchen where dirty water is the main ingredient. The goal is to cook this "soup" until it's clean enough to pour back into the river without hurting anyone. One of the most important chefs in this kitchen is a team of tiny, invisible bacteria. They have a specific job: eating up ammonia (a nasty chemical from human waste) and turning it into something else. But here's the twist: when these bacteria finish their job, they leave behind a new ingredient called nitrate. While ammonia is bad, too much nitrate in the river is also dangerous for people and animals. Traditionally, if a plant made too much nitrate, it had to build a whole second kitchen just to clean it up, which is expensive and complicated. Scientists have been trying to figure out if they can make the bacteria do both jobs—eating the ammonia and cleaning up the nitrate—at the exact same time in the same tank, like a master chef who can chop vegetables and stir the soup simultaneously without dropping a spoon.

This is exactly the puzzle researchers M. Muloiwa and CM Zvinowanda tackled in their study. They wanted to see if they could use a clever computer trick called "Particle Swarm Optimisation" (PSO) to find the perfect recipe for the bacteria. Think of PSO like a flock of birds searching for the best berry bush. Each bird (or "particle") flies around, trying different combinations of temperature, air flow, and time. If one bird finds a tasty spot, the others fly toward it, but they also keep their own ideas. Eventually, the whole flock converges on the absolute best spot. The researchers used this digital flock to test thousands of different settings in a lab to see if they could get the bacteria to remove nitrate while they were still busy removing ammonia.

The team set up a laboratory experiment using real wastewater from the Daspoort treatment plant in South Africa. They created a small, controlled "aeration chamber" (a tank where air is pumped in to help the bacteria breathe) and ran it for 54 days. They collected data on seven different ingredients: how long the water stayed in the tank, how much air was blown in, the temperature, the amount of food (COD), the amount of ammonia, the amount of bacteria (biomass), and the oxygen level. To make sense of all this messy data, they built a digital brain using a "Multilayer Perceptron" (MLP), which is a type of Artificial Neural Network. You can think of this as a super-smart student who looked at 324 different data points and learned the secret patterns of how the bacteria behave.

Once the digital brain understood the rules, they let the "flock of birds" (the PSO algorithm) go to work. The goal was simple: find the specific settings that would result in the lowest possible amount of nitrate left over. The computer simulation was incredibly successful. It found a "global optimum," or the perfect recipe, that resulted in a nitrate concentration of just 0.5744 mg/L. This is a very low number, well within the safe limits set by water authorities.

To achieve this magic number, the computer suggested a very specific set of conditions:

• Aeration Period: The water should stay in the tank for about 1.514 hours (roughly 1 hour and 31 minutes).
• Airflow Rate: The air pump should push 2.6911 L/min of air into the water.
• Temperature: The water needs to be kept warm, at 30.5270°C.
• Dissolved Oxygen (DO): There should be 5.5768 mg/L of oxygen in the water.

Under these perfect conditions, the system also managed to clean up the other pollutants effectively, leaving behind only 35.8187 mg/L of COD (organic waste) and 2.9126 mg/L of ammonia.

The researchers also did a "sensitivity analysis," which is like asking, "What happens if we change just one thing?" They discovered that two factors were the real bosses of the nitrate levels. Temperature was the biggest driver, accounting for 54.2% of the changes, followed by the Aeration Period at 22.5%. This means that if you want to control how much nitrate is left, you mostly just need to watch the thermometer and the clock. Interestingly, the study found that other methods people have tried, like using electricity or special new reactor designs, are often too expensive or require building entirely new plants. This study suggests that instead of tearing down old plants and building new ones, we can just tweak the settings on the ones we already have.

In the end, the paper suggests that by using smart computer optimization, it is indeed possible to get the bacteria to do simultaneous nitrification and denitrification. The results show that with the right temperature and timing, the bacteria can clean up the ammonia and the nitrate at the same time, potentially saving water treatment plants from needing expensive extra equipment. The authors are confident in these findings because their computer model matched real-world data very closely, with a high accuracy score, but they present this as a successful optimization of existing processes rather than a magic cure-all for every situation.