Site icon Connected World

Equitable Home Energy Management

Homes are getting smarter and more energy efficient—but are they within reach for the average homeowner? Not yet, but scientists are now developing a new framework for HEMS (home energy-management systems). Let’s look at how this will impact home construction projects in the future and how this just make energy-efficient homes more equitable in the future.

Dae-Hyun Choi, associate professor, Chung-Ang University, in South Korea, envisions a world where electric cars and energy trade between households.

Today, many energy decisions can be done by a HEMS, which efficiently manages the energy consumption of home appliances, by scheduling when an appliance should start or be turned off. These systems work to minimize electricity bills while also taking the user’s preferences into account. In time, the real hope is they will learn the users’ behavior to maximize energy consumption to its fullest.

Here’s the challenge. Models that use abstract equations to represent appliances and distributed energy resources are not very versatile and give suboptimal solutions. The other option is to use centralized machine learning, where data from thousands of users is collected, sent to a central server, and used to train a model from the ground up. Naturally, this can be costly and complex. Imagine the length of learning that needs to occur here?

Along with PhD student Sangyoon Lee from Chung-Ang University, Choi has proposed a novel data-driven strategy and framework based on F-DRL (federated deep reinforcement learning). This combines the advantages of various machine-learning techniques and a decentralized form of machine learning.

Let’s examine the model even closer. Here each home has a HEMS connected to various appliances and devices. These devices are gathering data. These data collected is all about its users’ energy consumption and tries to optimize a schedule for the appliances by creating a local model. These local models are all uploaded to global server, which averages them to produce a global model.

Afterwards, each HEMS replaces its local model with the global model and proceeds to train it once again using local data. This process is repeated several times, progressively improving the accuracy of both global and local model. In my mind, I liken it to how I trained my puppy to do tricks. I’m certain Professor Choi sees a more complex method here, but that is what I envision happening in the machine-learning process.

This approach was tested through simulations, showcasing its performance when scheduling the operation of various appliances at different homes. Here’s what is really impressive about this. It can manage the energy consumption of multiple smart homes and ensure the comfort of the consumers while taking their preferences into account in a distributed manner, according to Choi.

Would the federated reinforcement learning for energy management of multiple smart homes with distributed energy resources perform effectively here in the United States? It’s intelligent solutions such as this that change the way we look at our smart homes and the way we manage our energy consumption. Perhaps the bigger question is whether builders—and homeowners—will be on board with this model and embrace it? It certainly leads to more secure optimization of energy consumption in the future. Curious what you are thinking?

Want to tweet about this article? Use hashtags #construction #IoT #sustainability #AI #5G #cloud #edge #futureofwork #infrastructure 

Exit mobile version