Semi Leading Smart Grid Change

June 01, 2022 by Kevin Clemens

Semiconductor manufacturers are leading the charge toward zero-emission grid solutions using the latest AI and communications technologies.

Ever since the US established its electrical power grid in the 1920s and 1930s, the goal has been to provide a reliable one-way flow of power from a small number of industrial-scale power plants to homes, businesses, and industrial plants. Coal, oil, natural gas, hydroelectric, and, eventually, nuclear-generating facilities provided predictable energy output that could easily align with hourly, daily, monthly, and yearly power demand at acceptably low costs.


renewable energy grid

Renewable energy grid. Image used courtesy of Kenueone, CC0, via Wikimedia Commons


Greenhouse Gases

We know now that burning fossil fuels to produce electricity brings a high cost in the form of the emissions of greenhouse gasses (GHGs) that are altering the Earth’s climate. The solution is to transition from fossil fuels to renewable net-zero energy sources like solar and wind. The output of solar arrays and wind farms is both highly distributed into smaller generation facilities and highly variable, however, and this has exposed a crucial weakness in the existing power grid system—an inability to easily combine highly distributed power generation sources, along with existing legacy power plants that are slowly being retired.

The answer to managing a variety of distributed energy resources comes using high-performance computing (HPC) technology and developing software-defined smart grids using a technology called edge AI.


Edge AI

Edge AI combines artificial intelligence (AI) with the adoption of Internet of Things (IoT) enabled devices, and edge computing, a process by which computation is done near the user, close to where the data is located, rather than at a centralized computing facility.

Three innovations have allowed the development of edge AI technology.

First, AI has finally matured to the point where generalized machine learning allows AI models to be trained and “learn” to perform tasks under different conditions, providing optimized outcomes.

Second, this has largely come about thanks to advances in distributed computational power that allows massively parallel computational capability away from a centralized computer system.

Finally, the widespread adoption of IoT means data is collected at every stage of an industrial process. Along with the speed and secure information transfer brought about by 5G communication, this big data can be used to allow AI to revolutionize the world’s largest industries. 

Edge AI is used in the energy industry to combine historical data, weather patterns, demand forecasting, grid health, optimal power flow, and other information to create complex simulations that inform more efficient generation, distribution and management of energy resources to customers.

It turns out that semiconductor and computer chip manufacturers are particularly well-suited to the development of edge AI technology for use in energy management systems and development of smart grids.

US-based Nvidia Corporation is one company at the forefront of edge AI applications for energy and power grid management.  The computer chip and software producer has a variety of software and hardware projects working to create a modern smart grid that seamlessly integrates renewable energy into the infrastructure of existing power grids, making them more resilient and efficient.

Using HPC and edge AI enables optimizing power flow, predicting grid anomalies, preventing unplanned blackouts, and automating maintenance. For example, the Electric Power Research Institute (EPRI), an independent, nonprofit energy research and development organization, is building a grid simulator with NVIDIA that can be used to schedule outages of power systems and minimize downtime. 

According to Nvidia, “Using AI and HPC, utilities can model the electric grid as a connected graph with transformers and substations as nodes and transmission lines as edges. These models are trained on historical data to simulate specific grid outages, and address challenges to accommodate variable renewable energy, distributed energy resources, and shifting flow patterns.”



In addition to zero-emission renewable energy sources like wind and solar, another use for smart grids is on the horizon.

As the electrification of transportation progresses throughout this decade and past 2030, rapid charging of electric vehicles (EVs) will put greater demand on electrical power grids. The majority of EV charging is expected to occur at night when the demand on the power grid from air conditioning and industrial applications is less.

During the day, the power grid can be placed under strain by high afternoon temperatures. A smart grid could be used to extract electrical energy from parked EVs to support the grid while minimizing the need for special generators to handle these peak loads. This is called vehicle-to-grid or V2G.

Integrating EVs into the grid system to allow and manage the two-way transmission of electricity, while also paying for energy used to charge the EV, and crediting the EV owner for energy transferred from a parked EV to support the grid requires a sophisticated powerline communication (PLC) device.


Communication is Key

US-based semiconductor and software company Qualcomm has announced a next-generation PLC device to meet the needs for EV charging station communications.


EV Charging

EV charging. Photo used courtesy of Pixabay

The device, the QCA7006AQ, integrates 4G and 5G communications into a smart grid that allows vehicles to seamlessly authenticate on the network through Plug and Charge automated payments for EV charging, and to coordinate the timing and direction of energy to and from the grid and home.

Smart-grid charging applications will allow users the flexibility to choose optimal charging times and extract energy from their EVs as needed, while also supporting the ability of EVs to support peak grid load requirements and aid in the further adoption of zero-emission energy resources.

By bringing the experience and expertise that semiconductor manufacturers have in AI, IoT, and 4G and 5G communications, creating a smart grid that can seamlessly integrate renewable resources and EVs into a new and resilient zero-emission energy grid is well underway.


Feature image used courtesy of Kenueone, CC0/Wikimedia Commons