In our blog we wrote about sustainability issues several times already. Wamisoftware team strongly supports environmental initiatives. Currently, we are collaborating with Israeli company FSIGHT, developer of Energy AI — a platform for distributed grid management based on Artificial Intelligence. Our developers are involved in different projects aiming at the optimization of the operation of the hydroelectric power station; managing tariff evaluation engine; forecasting intraday price etc. Keep on reading to see the insight from Wamisoftware's Java Team Lead.

We now face an urgent challenge to transition from fossil fuels to prevent the worst effects of climate change. Of course, all energy sources have some impact on our environment. But fossil fuels — coal, oil, and natural gas — do substantially more harm than renewable energy sources by most measures, including air and water pollution, damage to public health, wildlife and habitat loss, water use, land use, and global warming emissions. Still, fossil fuels are the largest sources of energy for electricity generation — they account for 80 percent of global energy consumption. Nevertheless, we are slowly moving in the right direction. U.S. Energy Information Administration analysis says that in April of 2019 U.S. monthly electricity generation from renewable sources exceeded coal-fired generation for the first time. Renewable sources provided 23 percent of total electricity generation compared to coal’s 20 percent. EIA also forecasts that renewable generation will continue to outpace coal in 2021. Moving from brown to green energy is imperative to achieve better lives.

Knowing that, what difficulties can we expect from shifting to clean energy?

To begin with, renewables are often highly dependent on the weather and time of year and cannot be switched on and off to produce energy any time people need it. The big sources — sun, wind, and water — come and go on nature’s schedule.

Additionally, the increasing reliance on renewables is expected to generate bigger demand peaks on the electricity market.

Long-term energy storage like batteries and software platforms that will stream real-time aggregation of telemetry from thousands of batteries to measure, forecast, optimize and control — together can solve a problem with the equation of demand and supply.

People can and do take control over their energy consumptions if given the tools to do so.

And as ‘time of use’ tariffs spread globally, thousands of power walls are installed in people’s homes to do peak shaving for an electrical utility. A rechargeable battery stores energy from solar panels during the day or from offpeak hours when grid energy is cheapest and uses that energy during the rest of the day. The remaining problem is that batteries cost approximately one arm and one leg. Some appliances will use more than the output of one battery, so families would probably need a few battery packs. The price of a solar setup including installation, supporting hardware and tax credits sometimes exceed $15,000. Hopefully, as the cost of batteries and solar panels continues to drop, going green and owning a battery-powered home in the future could be available for everyone.

Other than that, behind-the-meter systems like residential solar panel installations can be a great way to save on electricity bills and maximize how much of their own solar production customers are using. They can prioritize charging at-home batteries during low-cost times, arrange backup power for a house for days in the event of a power outage, charge electric vehicles, or even make some money buying and selling from the grid. In the case of wind and solar, excess generation can be sent through the meter and to the grid for credit on the electricity bill. All of this can be done with a corresponding app so residential customers can monitor their solar energy storage and see electric grid costs at different times.

Industrial customers also can eliminate expensive peak demand charges with a combination of solar and AI-powered energy storage. Battery farm stores huge amounts of energy from renewable sources and funnels it out to the grid when usage is high significantly reducing your business’s operating expenses.

Is there even more profitable approach? Yes, community-based power plants.

«At the moment, our team provides the “Gilboa Iris” project with an algorithm for community-based power plants. Closed community is, for example, 300 houses that share one counter and are connected into one net. Based on collected historical data, the system calculates the best way to save on bills. Let’s say that in given time period you paid this much money, but if you have a PV (Solar Photovoltaic) and add a set number of batteries, the new price will be better resulting in 20 percent savings. The algorithm for community-based power plants determines at each point in time whether your neighbour wants to obtain energy, so you can give it to him if you have energy surplus. If the community doesn’t have storage to save energy, the algorithm sells it to power company.

The advantage of local community is that there is a domestic price resulting from managing demands within it and you have an opportunity to buy both on local and external market. Rates in the latter are formed by power company, which makes them more expensive than on local market.

In RD Energy project Wamisoftware team provides forecasts for Israeli power plant, so they know how many raw materials to purchase in order to prevent excess or deficiency. We also apply these forecasts for communities. Knowing the approximate quantity of consumption by every single individual in the community, we can harmlessly take their excessive energy, because we forecasted that after a certain time electricity will be cheap for recharging the batteries.

For Backend Development we use Java, UI is made with React, and in the future we plan to break Monolith application into Microservices, we also plan to use C language for forecasting.»

But how exactly are predictions about peak loads made? The thing is, we can only know if there’s a high likelihood of a peak. Our goal is to discharge home battery packs during the peak grid load and the homeowner will use them for backup power. It is also important to control both when the batteries discharge and when they charge while spreading out the charging over a longer period of time. And once we’ve decided to discharge batteries to avoid the peak, how do we control them?

First, we bring all batteries together into hierarchical aggregations. It’s hard to virtually represent them in cloud software because of the diversity of installations, home loads, and uncertainty in communication.

Second, we need to aggregate local telemetry (power, frequency, and voltage measurements) and the telemetry across the virtual power plant as well. So at the same time, we have to take into account central optimization and local needs.

Third, having aggregated telemetry, we find out the capacity that’s available in the virtual power plant. Smart grid management platform learns and profiles the behavior of each prosumer, and chooses, automatically and in real time, the most cost-effective strategies for using, storing and trading energy. Based on price forecasts local optimization makes a plan for the battery, for example, the algorithm temporarily lowers the capacity of the battery in order to affect the peak. Huge aggregations demand different APIs and data processing patterns. Furthermore, developers are constantly dealing with complex concurrency issues. All these algorithms depend on the software platform.

The rise of renewable energy production creates the need for grid flexibility, distributed energy resources and information and communication technology can help provide that flexibility — if they are enabled and encouraged. Hardware is not on its own enough to replace fossil fuels while maintaining our current standard of on-demand and reliable power.

Software is really the key to enabling diverse components to act in concert.

What are your thoughts about it? 👇🏻