[ad_1]
Artificial intelligence (AI) is currently gaining traction and is becoming more and more central to homes and businesses as they move from simple grid connectivity to revenue streams for private power generation, energy storage, electric vehicle (EV) charging, and load balancing. We are finding new uses for it. As AI becomes ubiquitous, what is the difference between advanced control through simple algorithms and true intelligence?
From magazine PV 12/23-01/24
AI may be a buzzword, but when it comes to energy management, it can capture vast amounts of data and make meaningful predictions to optimize the use of renewable energy and storage, especially as EVs become more prevalent. The only tool that can do that right now is AI.
Mainz, Germany-based energy startup Lade focuses on optimizing renewable energy consumption across EV charging and energy management. AI has already proven to be a useful tool deployed for the benefit of customers.
Dennis Schulmeier, Founder and CEO of Lade, said: PV magazine An in-house team of seven dedicated employees is working on the AI in conjunction with the company’s LADEgenius product, which can handle 200 EV chargers, along with inputs and outputs from PV modules, energy storage systems, and local inputs from EV chargers. To meet grid regulations interfacing with data input. LADEgenius is basically an on-site load manager and connector that can make decisions with the help of cloud intelligence. Its cloud intelligence uses AI and machine learning through a system the company calls Lana.
“Lana is an AI because it can predict energy availability,” Schulmeyer said. “Lana collects data from the German weather service and can make forecasts for up to five days to see how much renewable energy is available.
“It also predicts the availability of local renewable energy for buildings and power generation, reads inverter data and weather values at the installation site, and also predicts consumption. Our main [unique selling point] You can also predict the arrival and departure times of your car, as well as the amount of energy it actually requires, up to five days in advance. [we] Calculate the optimal rate plan at that time. ”
Schulmeyer said AI trains on data and runs on models hosted on cloud servers, all of which comes at a “high cost” and that Lard is fully powered by the use of renewable energy. By paying this fee, you are adding additional costs to your company. server.
“Our internal team has been developing AI over the past three years,” the CEO said. “We initially trained to use open source data, but we also used real data from chargers, solar power data from our customers, for example, and even our own real-world data here in Mainz. We also added a world setup.” Schulmeyer acknowledged that adding additional customer data to Lana’s training data further improved the predictions.
Optimization with AI
Ido Ginodi, vice president of products at SolarEdge, explains how AI is being used to optimize energy management systems and how it can be used even in the home in ways that traditional control algorithms cannot. We explained how to handle fundamentally difficult optimization problems and predictions. Israel-based SolarEdge is well-known in the solar power industry, and as complexities emerge between energy generation and storage, EV charging, data, and predictions, Ginodi said his company is trying to figure out how to leverage the benefits of AI. They expressed considerable enthusiasm about how they are being utilized. “The line between a good solid algorithmic approach and AI is blurred,” Ginodi says. “However, after spending several years researching AI in academic settings, we have found that much of what people, including ourselves, are doing in this field is truly AI-driven, and we believe that providing cutting-edge energy optimization Ginodi says the need for AI goes from a single home with just one EV charger to multiple, or even hundreds, of EV chargers. He explained that this is not the only case when the scale of the application expands to apartment complexes, commercial facilities, and industrial sites with chargers. “Actually, I would argue a little bit differently, but in residential use cases, AI is very important,” Ginodi said. “Energy management problems are fundamentally difficult optimization problems. We started with the concept of power optimization, which is optimizing the amount of energy that can be squeezed out of a solar array. takes it a few steps further and optimizes site-wide performance, which is orders of magnitude more complex.” SolarEdge executives explain that energy management systems can optimize metrics for the benefit of the end customer. Did. This is done by coordinating elements such as PV generation, battery distribution, EV charging, and load orchestration. The system also uses data to make decisions that enable dynamic pricing, market participation, and even power outage preparedness, as well as integrated heating, ventilation, and air conditioning for preheating and cooling. You can also optimize it. “You end up having multiple flexibilities,” Ginodi said. “That’s a lot, and it’s fascinating. In some places, AI-driven solutions can produce results far better than what can be achieved with simple algorithmic approaches. But going further We develop predictive models for consumption, production, import/export tariffs, and grid events based on machine learning regression techniques. Once we have these four models, how do we manage the various resources in the system? For the end user, this means either optimizing profits as is typical, or optimizing for convenience or decarbonization depending on the user’s preferences. It refers to a management system that aims to optimize. Ginodi added that SolarEdge portfolio companies are also working closely to incorporate his AI capabilities into their products. In particular, EV charging management company Wevo is working to cost-effectively scale EV charging through predictive load management and capacity management. While static and dynamic load management technologies are becoming increasingly prolific in the industry, AI in the form of predictive modeling can significantly improve concurrency factors, or the ability to fit more chargers into a given grid connection point. bring improvement. “Say a company wants to offer electric parking in their parking lot,” Ginodi says. “Providing 100 new spots of 11/22 kW each is very expensive, which means an additional 1 MW or 2 MW of power is required. However, you don’t have to charge the vehicles all at once, or even statically connect capacity to each charger. That’s dynamic load management. Going one step further, Wevo The generated predictions can be incorporated to build optimal charging schedules. The model assumes that cars enter the parking lot at a constant rate and what the level of local production and total consumption will be at each point in time. “With these predictions in hand, we can serve more vehicles and drivers – up to 20 times more than a simple implementation.”
Schulmeier said advanced software-based controls may solve some problems in a single-family home situation, but when considering multiple EV chargers, it’s important to have a standard load manager or PV surplus charging system. He said it would struggle to show its true benefits right away. “That’s the big highlight,” he added.
In large commercial and industrial settings, energy management needs to be done across a large number of EV chargers to avoid unnecessarily high demands without coordination, making the task increasingly complex. This is further complicated by adding production and consumption forecasting with weather data, while providing features such as peak shaving. Lade’s founder said it would be impossible to operate this without AI technology.
Improve
“We are doing all this and getting better,” he said. “When you connect to our EV charger for the first time, we say the accuracy of estimating the energy your car will need over time is about 67% from a low starting point. But of course, the more data you have, the more The more you have, the better your data will be, and the advantage of a startup is that you can run and adapt many models and AI technologies.”
Schulmeyer was careful to point out the benefits of the entire ecosystem beyond AI. “It’s not just about the AI algorithms…it’s also about how you think as a company,” he said. “We’re not alone, so we’re going to find ways to involve others. In fact, we plan to use LADEgenius to add third-party chargers to the cloud. But there are only a few that exist in this region. This is important because we are not independent in that we are the only ones, and above all our goal is the energy transition with the help of electric mobility.”
This content is copyrighted and may not be reused. If you would like to collaborate with us and reuse some of our content, please contact us at [email protected].
[ad_2]
Source link
