Energy use optimization allows consumers to minimize their electricity cost. Today, smart energy management applications enable to shift the consumption of individual devices to hours with low electricity prices.
However, the optimization potential is far greater, since several energy devices (electric cars, home batteries) can act both as consumers and producers of electricity. The simultaneous and multidirectional optimization of these assets allows households to achieve very large cost savings.
As a result of two years of research and development, I recently introduced an energy optimizer that finds the mathematically best energy usage pattern for an unlimited number of simultaneously operating multidirectional assets with respect to one or more energy markets, taking into account asset-specific limitations (required battery charge level, connection of the electric car to the charger, total capacity, power, charging efficiency, etc.).
This article shows, on the basis of a sample household, how much can be saved at different levels of energy optimization. In order to compare the levels, the optimal behavior of the household was simulated during a one-year period (10/24/2021 – 10/23/2022) and the total cost of electricity was estimated. To illustrate the energy behavior, I will highlight the energy usage pattern in the one-day period 08/17/2022 08:00 – 08/18/2022 07:00.
Disclaimer: data that has been used in this analysis is published in ENTSOE-E Transparancy Platform under a Creative Commons Attribution 4.0 International License (CC-BY 4.0).
Sample household
The profile of the sample household used in the analysis is as follows:
- Location: Estonia.
- Electricity consumption: The household consumes 9654 kWh of electricity per year, and the size of the household’s main fuse is 20 Amps.
- Solar energy solution: Solar panels with a maximum production capacity of 7.5 kW, the angle of the roof is 40 degrees, the roof is facing south, and the total annual production of solar panels is 6981 kWh.
- Electricity package: Variable (based on stock market prices) (link to electricity packages).
- Network package: Network 4 (link to Elektrilevi packages).
In addition, our sample household owns an electric vehicle (EV):
- electricity consumption: 20 kWh/100 km.
- Annual mileage: 15,000 km.
- Electric car charger: power 7.4 kW.
- The car is connected to the charger on weekdays between 18:00 and 08:00 and on weekends between 15:00 and 10:00.
- The round-trip charging efficiency is 90%.
I also implement an additional requirement, that every morning when the EV leaves home, the battery must be fully charged.
Base scenario – no optimization
To evaluate the gain from optimization, we first start with a scenario with no energy optimization. Considering the baseload of the household, the production of solar panels and the consumption of the electric vehicle, the annual electricity cost of the household would be €1830 (using electricity prices of Estonian market from 10/24/2021 to 10/23/2022).
Taking a closer look at electricity consumption over a one-day period in August 2022 (table below), our household would have sold electricity to the grid between 08:00 and 16:00 as the solar panels produced more than the household needed.
At 18:00, when the vehicle is plugged in, it automatically starts charging, despite the extremely high price of electricity. As a general pattern, it can be highlighted that without optimization of energy use, our household would have sold electricity to the grid during the hours with the cheapest price and consumed from the grid when electricity was expensive.
Level 1 – Optimizing the consumption of a single asset by the electricity prices
The simplest level of optimization that I consider in this article is to control the consumption of a single device (e.g. electric vehicle) according to electricity market prices. This can be done either by integrated software (e.g. modern heat pumps) or by external software that communicates with the device (e.g. Gridio, which selects the cheapest hours for the charging).
If our household had optimized the charging of the electric car according to market prices, it would have saved approx. 400€ per year. Taking a closer look at the simulation of 17–18 of August, we can see that the optimization system has shifted the charging of the electric vehicle from 18:00 (when the car was connected) to 04:00 in the morning, when the electricity price is the cheapest.
Level 2 – optimizing the charging and discharging of a single asset
Today, some models of Electric Vehicles allow bidirectional charging, which means that the electric car can act both as a consumer and a producer.
Bidirectional optimization can be performed both "naively" and "non-naively". The former means optimizing purely according to market prices without considering the baseload of the household or the production of solar panels. In this analysis, both the market prices and the conditions of the household were given as input to the optimizer with a goal to find a solution that minimizes the overall electricity cost of the household. The results reveal that the bidirectional charging would have enables to decrease the energy cost of a household by nearly €1380 compared to the base scenario.
The analysis of a single day (image below) shows a significantly more complex behavior pattern than before: the EV sends electricity to the grid when the price of electricity is expensive, charges when the price of electricity is cheap and decides to minimize the total consumption of the household at certain hours (22:00–23:00). At the same time, the optimization system also considers the constraint that the EV must be fully charged by 08:00 in the morning.
Level 3 – Bidirectional optimization of multiple assets
Let’s say that our household also buys a Huawei Luna2000–10-S0 home battery with a capacity of 10 kWh and a charging speed of 5 kW. This means that we have the opportunity to perform simultaneous bidirectional optimization of several devices (EV and home battery).
With this setup, our example household would have achieved a total annual cost of -€191, which means that the income from the sale of electricity would have been greater than the cost of consumption.
The single day analysis (image below) reveals a relatively complex energy behavior pattern. I would like to highlight the relationship between the column "Consumption" and "Electricity price": in many hours, the energy consumption of the household has been reduced to almost zero (10:00–18:00). The reason for this is the network fee and taxes that the household has to pay when consuming electricity, but which cannot be recovered by producing electricity. This in turn means that during many hours it is optimal for the household to be "off the grid".
A fully optimized household is virtually disconnected from the grid for many hours and interacts with the grid only when the arbitrage opportunity is high!
During hours with high electricity prices (08:00–09:00 and 18:00–21:00), the household sends electricity to the grid (consumption is negative) and the EV is fully charged again in the early morning hours of August 18.
Anyone who wants to see the optimized battery and EV pattern over a whole year can access the corresponding Google spreadsheet from this link.
Level 4 – multi-asset and multi-market optimization
In addition to the electricity price market, there are also electricity frequency markets, which aim to ensure a constant frequency (either 50 or 60 Hz) in the electricity grid. Participating in both price and frequency markets would mean the simultaneous optimization of multiple devices for multiple markets. The energy optimizer can handle this complexity, but since the frequency markets in the Baltics are still in the test phase, I leave this analysis for the future. The multi-asset and multi-market situation can be graphically illustrated as follows:
Summary
Let’s take another look at the household electricity consumption pattern on August 17–18, 2022 for all optimization levels. We can highlight the rule that the more complex optimization we allow to be performed, the more the household consumes in the opposite direction to the electricity price, and in hours where the possibility of price arbitrage is small, the optimization system decides to zero the total consumption of the household (as if the household was disconnected from the grid).
A comparison of the household’s annual electricity cost across optimization levels shows that the biggest win comes from the addition of bidirectional optimization capability. In case of simultaneous optimization of several devices, it is possible to make the electricity cost negative (the net income from the sale of electricity exceeds the cost of consumption).
In sum, smart energy control can be very simple (charge on the cheapest hour) or highly complex and optimized (multi-directional simultaneous charging of multiple assets). Sometimes, too complex solutions are not worth it, but not in this case. When we use the full potential of the Energy Optimization, we can potentially achieve the net-negative energy cost.
