Introduction
The rapid explosion of electric vehicles is making us reframe how we provide power to the grid. Fixed allocation-based conventional charging stations are often not capable of catering to the load requirement during peak time periods. Overheats, fluctuating voltage, and energy loss ensue. To overcome these limitations, the Pingalax research laboratory is investigating the integration of AI and IoT in a single platform for intelligent charging.
Not a product unto itself, the Pingalax Cloud is a platform that unites hardware with software and predictive algorithms. The goal is simple: be more efficient, give chargers longer life, and save both public and private charging networks money.
From Fixed Allocation to Dynamic Charging
Regular EV chargers flow power continuously, whether or not there’s a car connected to one, and how much will depend on demand from the car, with no regard for what others may need. This creates inefficiencies. A car that has an almost full battery is still given power, while one with nearly empty batteries can be waiting needlessly.
The charging of AIoT changes are being made to the distribution in real-time. Data on usage profiles, battery SOC, and grid conditions are obtained in the system.
Machine-learning models then forecast demand and redistribute energy where it is needed. In recent simulations, this scheme resulted in a 22% reduction in peak-hour power consumption over a fixed allocation-based charging scheme, with an 18% reduction in the charging cycle time for high-priority users.
How Pingalax Cloud Operates
At the center of Pingalax Cloud is a multi-layer algorithm. A first layer will observe the chargers themselves. A second layer pools information within a network across stations. This third bank is connected with external elements such as solar panels or an energy storage means. The system could then ultimately decide to use grid power less, for instance, during peak demand periods, tapping energy that has been stored instead.
When it was tested in Jakarta, the predictive diagnostics on board detected problems 35% faster than before. This reduced maintenance downtime and saved operators about 28% of the costs for high-use hubs.
Case Applications
One test location, a public charging hub for fleet vehicles, had been grappling with idle power draw and spikes at night peaks. With the introduction of dynamic load balancing, idle consumption decreased by just under one-third and the overall station efficiency was increased by approximately 12%.
Another initiative targeted a commercial fleet with over fifty chargers. Applying scheduling algorithms that aligned charging sessions with vehicle usage patterns reduced fleet downtime by 25% and the instances of charging conflicts during high-demand hours by 40%.
Future Directions
The team is currently working with more sophisticated forecasting models that employ artificial intelligence to predict both grid supply and vehicle demand more accurately. He also mentioned that “other future application areas of Loligo will be secured billing via blockchain and the 5th generation mobile networks that will greatly decrease latency in cloud communications.”
Role of Semiconductor Technology
While so much attention is paid to software, Pingalax stresses that hardware matters just as well. The company has developed SiC (silicon carbide) MOSFETs that can switch quickly and deliver energy accurately. The components cut wasted energy by as much as 30% counting lab environments and they also made it feasible to produce more compact chargers.
With custom semiconductors, decisions can be made in real-time, and not just the domain of software. Higher switching cycles also cut charging time by almost 50% in some tests, while reduced heat generation makes for a more reliable system overall.
Conclusion
The study indicates that the use of AI and IoT in charging systems results in tactical advantages by diminishing the charging time, reducing operational costs, and improving grid compliance. “Indonesia has limited electric vehicle infrastructure, as well as our grid’s stability issue, and that kind of innovation can push adoption faster,” she added.
Pingalax will keep improving its AI-based chips and predictive models, and scaling trials in commercial and public charging networks. The results until now have given a clear direction: smart charging is not a luxury of the future, but a necessity, to ensure sustainable growth of EVs.
