Unsaturated research topics in MPPT and renewables include using AI-based algorithms (like reinforcement learning or hybrid AI methods) to improve tracking under fast-changing weather, MPPT for multi-source hybrid systems, low-cost controllers for small-scale renewables, and integration of MPPT with smart grids and storage. In power electronics, future topics include wide-bandgap devices (SiC, GaN), high-efficiency converters for EVs and microgrids, modular designs, and AI-driven fault detection and predictive control.

Saman Farzadnia - I am assuming that this is some kind of academic joke? The amount of academic research on the topic of MPPT is many orders of magnitude higher than what is practically needed. MPPT is a trivial problem, that somehow academia thought was the most important problem for them to solve. There are literally tens of millions of research papers on this topic and when you analyze all these papers there are basically about half a dozen underlying unique concepts. The reality is that if you choose any one of the half dozen practical solutions a graduate level student would spend less than a week coding it and testing it in a lab and then it will work and achieve over 99% MPPT efficiency.

I work in industry and I have a simple rule - if I come across a student looking for employment and they spent their time in academia researching the topic of MPPT then I simply will not even consider them as a potential candidate.

Your suggestion of applying AI to this topic of MPPT seems like some kind of joke based on pursuing futility. If you want a good application for AI, then I suggest you look at investing time creating more AI cute cat videos as there is orders of magnitude need for more cute cat videos than new MPPT algorithms.

I don't know about AI-based MPPT. but in power electronics, I am utilizing AI in the field of power electronics to design model parameters for devices such as transformers and electrosurgical units. For example, I optimize transformer core size and the number of coil turns using machine learning algorithms.

AI-based algorithms include integrating digital twin models for real-time online monitoring and predictive MPPT, developing adaptive AI/ML-based MPPT under highly variable and harsh weather conditions, and hybridizing multiple renewable sources with advanced AI-driven controllers.

In MPPT research, the areas that are still unexplored:

Hybrid AI

Weather-based prediction

New PV technologies

Multi-source systems

EV charging

In Power Electronics research, the unexplored areas are:

AI-controlled converters

GaN/SiC devices

Grid-forming inverters

Cybersecurity

Digital twin

@Saman Farzadnia

I’m not deeply specialized in MPPT, but I was thinking about the potential of integrating short-term weather prediction with AI-based MPPT strategies. For instance, in off-grid PV systems without battery storage, such forecasting could help estimate how long the connected devices can remain operational under rapid irradiance changes, and enable intelligent load prioritization. The same concept could be extended to off-grid systems with battery storage or hybrid systems, where forecasting can optimize battery usage and scheduling between PV and auxiliary sources. Even in grid-tied PV systems, weather-aware prediction could anticipate the additional demand from the grid (e.g., during cloudy or rainy periods) and allow better planning of energy consumption and cost management. I believe this direction remains under-explored and could provide practical value for real-world PV applications.

Dimas christian teguh Utomo - yes there are an infinite number of non relevant factors that you could consider using as inputs to an AI MPPT algorithm - you could consider counting how many cute cat videos have been uploaded to Tic Tok in the last day, however if you stop and think for a moment then you realize that it is highly unlikely to be in anyway correlated to determining an optimum MPPT algorithm. I don't understand why it seems to be acceptable in scientific circles to completely abandon and rational critical thinking when it comes to AI. Surely if you can mathematically prove that an algorithm is the absolute optimum, then why would you even begin to question if you could make it more than what is provable as the absolute optimum?

The MPPT 'challenge' is basically on a similar level as the challenge of how would you create an algorithm that can add two single digit whole numbers - this problem was adequately solved by the first person who could understand the problem and then the rest of the world moved on. Why don't academics understand the level of triviality to the MPPT challenge? This madness in academia really needs to stop! We don't want any more research being invested in how to develop new routines to add two single digit whole numbers or MPPT algorithms.

The concept of MPPT is kindergarten level and it reflects bad on students when they can't understand just how basic the concept is.

I get your point sir. the fundamental MPPT concept is indeed straightforward and well-solved. But maybe the real interest nowadays is not about creating a new MPPT algorithm itself, but how MPPT interacts with forecasting, load management, or integration in hybrid/grid systems.

So rather than seeing it as reinventing something trivial, I think some researchers are trying to place MPPT in a larger context where it can still add value.