A novel intelligent control using RBFNN-class topper optimization for stability enhancement in hybrid standalone microgrid

Keeping the Lights On in Remote Places

Many remote villages, islands, and off‑grid communities are starting to power themselves with a mix of solar panels, small wind turbines, and batteries. But keeping the lights steady when clouds pass or the wind drops is harder than it sounds. This paper introduces a new “intelligent” control method that helps such small power networks, called standalone microgrids, deliver smooth, reliable electricity with fewer flickers, less electrical noise, and better use of clean energy.

How a Small Power Network Works

The study looks at a compact power system that combines three key pieces: a solar array, a wind turbine, and a large battery pack, all tied together on a common direct‑current (DC) line. From this shared backbone, an electronic device called an inverter produces the regular three‑phase alternating current (AC) that homes and equipment use. Solar and wind power are naturally unsteady—sun and wind rise and fall quickly—so the battery steps in to store extra energy or fill gaps. The entire system must constantly juggle these flows so that the DC line stays near its target voltage and the AC delivered to the load remains clean and stable.

Why Traditional Control Falls Short

Conventional controllers in such systems use fixed settings tuned by hand. They can work well around one operating point, but they struggle when conditions change quickly: a sudden jump in demand, clouds racing across the sky, or gusty winds. These methods often allow noticeable swings in voltage and produce distorted current waveforms that can stress devices and waste energy. More advanced approaches based on fuzzy logic or complex predictive models can adapt better, but they tend to require heavy computation and intricate rule sets, making them hard to run in real time on affordable hardware.

A Two‑Layer Intelligent Control Approach

The authors propose a two‑layer control scheme that combines a fast‑acting neural network with a slower supervisory optimizer. The lower layer uses a radial basis function neural network, a kind of machine‑learning model that excels at handling nonlinear behavior. It watches how far the DC voltage and inverter currents drift from their desired values, then quickly adjusts converter and inverter commands to correct these errors. On top of this, a “Class Topper Optimization Algorithm” plays the role of a coach. Inspired by how students learn from the top performer in a class, it periodically reviews how well the system is doing—looking at voltage deviations and how sinusoidal the current is—and gently tunes key control gains and learning rates within safe limits. This separation of fast learning and slower optimization keeps the system responsive without becoming unstable.

Putting the Smart Controller to the Test

To see how well the new method performs, the researchers simulated a microgrid with a 10‑kilowatt solar array, a 20‑kilowatt wind turbine, and a 60‑kilowatt‑hour battery. They subjected it to sharp changes in sunlight, wind speed, and load, including step changes from low to full demand and nonlinear loads that normally introduce strong electrical noise. Compared with a standard proportional–integral controller and a more basic neural‑network controller, the new scheme kept the DC voltage much closer to its target, recovered from disturbances about twice as fast, and reduced unwanted ripples. The electrical current delivered to the load became almost perfectly smooth, with harmonic distortion held well below widely used IEEE power‑quality limits.

Why This Matters for Future Microgrids

For non‑specialists, the main takeaway is that this intelligent control framework helps small renewable‑based power systems behave more like a sturdy conventional grid: lights stay steady, motors run smoothly, and sensitive electronics are better protected, even when the weather and demand are changing quickly. Because the method is computationally light and modular, it can be implemented on practical digital controllers and extended to larger or multi‑area microgrids without retraining from scratch. In short, the work shows a promising path toward more reliable, cleaner local power systems that can bring high‑quality electricity to places far from traditional power lines.

Citation: Myla, A., Gorantla, S. A novel intelligent control using RBFNN-class topper optimization for stability enhancement in hybrid standalone microgrid. Sci Rep 16, 14008 (2026). https://doi.org/10.1038/s41598-026-44798-6

Keywords: microgrid, renewable energy control, neural network controller, battery energy storage, power quality