If over Swabia the sun is shining, it. closely in the power lines of the Wertachau, a settlement 30 km South of Augsburg On many buildings, solar systems installed, the floods of sunshine, the local distribution grid with electricity. The lines are not designed but. When the weather is nice you can run over sometimes: - current flows in the supra-local networks, which makes the control more complicated. In addition, the responsible network operator, the Innogy, a subsidiary of Lechwerke must operate, the more effort required to keep the voltage in the copper cables constant.
The Wertachau is not an isolated case. Especially in rural areas, some of the distribution networks with the energy transition under strong pressure. The biggest Problem is the network operator lies in the fact that the solar systems provide electricity when only a little is needed. Because at noon, to the highest solar radiation, are most of the residents out of the house. Worries you, however, that the consumption party is likely to evening rise significantly in the future, at a cost to the networks. Because building owners are increasingly opting for electricity-powered heat pumps instead of Gas - or Oil-fired boiler. And also the electric cars to be charged when commuters arrive home in the evening.
Artificial intelligence is expected to vote on power consumption and generation
Innogy, one of the large German distribution network operator, wants to deal with the Situation now with artificial intelligence (AI). Together with the Aachen University RWTH Aachen University and other partners, the company has developed an AI-based control system that coordinates the electricity consumption in distribution networks automatically on the local production. As a test field, the Wertachau was elected. In order to simulate future conditions, Innogy, residents there are three electric cars available, and multiple battery storage installed. The heart of the "Smart Operator" baptized controller is an algorithm that decides itself when the batteries and the connected consumption devices such as washing machines or dryer power - within a time window, the participants in the pilot project define in advance.
in order To make the right decisions, analyzes the algorithm continuously generation, demand and network utilisation. "On the basis of weather forecasts, the Smart Operator knows, for example, to anticipate how much feed is," says Joachim Schneider, head of technology & Operations at Innogy. So he can move load operations or the operation of the electrical equipment in the periods in which the solar systems deliver a lot of power. Also heat pumps and night storage heaters, the Software uses the buffers, since the heat generated can be save. It will get better with time: Through the analysis of network data, you can constantly checks how well the scheme works, and learn from it. Around 35 percent more locally generated green electricity areas, the networks in the Wertachau and in two further test, according to Schneider, so. "In this way, a costly Expansion of the networks can be reduced," he says. Now, the company wants to use the KI solution also in other distribution networks.
A promising approach, says Robert span Heimer from the digital Association Bitkom. "KI can be drawn from large amounts of data, conclusions, resulting in decisions and often self-implement. Such solutions do not work, so be stubborn a pre-defined program, but act depending on the Situation," he explains. Thus, KI for a variety of applications, is suitable in the power supply. Span Heimer cites another example: "KI-solutions can also help disturbances in the energy system, such as a hacker attack or to recognize a technical Problem at an early stage by analyzing network data for patterns." You will recognize unexpected deviations, it indicates a Problem. In perspective, you could initiate even against measures, clamping Heimer.
The Software knows how the inhabitants of the houses the use of available space
Also at the level of individual buildings, he sees opportunities for self-learning analysis systems. So Eon has developed together with Microsoft, a KI-based energy management system for households, which should come onto the market in spring. Thus, the provider aims, among other things, to homeowners who have a solar system on the roof and an electric car in the Garage. "The System anticipates when and how much solar energy is available and how much the electricity in the house is required. On this Basis, it is designed for every day an individual charging strategy," explains Eon-expert Hassan Harb. In addition, the Software should allow households to be on the current market. "Who has, for example, a Solar-battery storage, then says in a high exchange current from the memory prices to sell", Harb. If, however, nationwide a lot of Wind or solar energy is generated and buyers in the stock market even more money for the decrease get can be used the batteries to power from the grid. "Whether, when and how much electricity is bought or sold is, the System on the basis of forecasts of stock prices and the energy needs of the households," says Harb.
During KI in the network operation and control of batteries and charging stations is still in its infancy, has kept them in the boiler room already: In some office and retail real estate, and also in some multi-family houses, are used today in forecasting methods, to the generation of the heat better to the needs. How the Heating works with KI, specifically, explains Nico Klecka of Leanheat, a manufacturer of such controls. "First of all, the System must understand the building: what are its thermodynamic properties, how much the heat storage walls and floors, and how it behaves in the sunshine?" This is the Software with a variety of data lined - about the temperatures in the individual rooms, with the operating data of the heating and weather data such as solar radiation or outdoor temperature. In this information, you will recognize patterns that can then be derived, when and how strongly are heated, so that the desired room temperatures are reached. As the outdoor temperature and the solar radiation are no by the window, a number of important factors, it takes into account weather forecasts. In addition, the systems work on a loop with Feedback, so that the control is constantly accurate. "Households can save between five and fifteen percent of the energy," says Klecka.
need to take the users of the building but in terms of purchasing, that the Software producers know when and what rooms will be used. What is in offices or Hotels no Problem at all, it should make some owners shy away. The data give clues to the everyday life of households - many people so much transparency, abdominal pain. The concern for privacy could also be the use of KI for the control of charging stations and battery to the hurdle to save.