Perhaps the most important innovation in the Matter standard – introduced with version 1.3 – is energy management for devices. All products that make use of it can help to optimize the energy flow in the household. This is done primarily by exchanging information: Electrical consumers provide data on their current or planned energy consumption. Power generators such as a balcony power station or the PV system on the roof communicate how many kilowatts they produce per hour.
External information such as the weather report, statistical consumption forecasts or an energy supplier’s dynamic electricity tariff, whose prices can fluctuate over the course of the day, are added as required. All with the aim of making the use of energy more efficient, saving costs and protecting the environment.
Table of contents
Co-author
This article was written in collaboration with Wido Wirsam. The software architect, consultant and developer has more than 20 years of professional experience with AI, energy management, IoT and cloud-based services. With his company intuisoft (link), he also offers consulting services regarding Matter.
The goal: optimized power consumption
An energy management system, or EMS for short, matches electricity supply and demand in the building. Its task is to shift the operation of electrical appliances over time or adjust their energy consumption to minimize electricity costs. For home producers, this means consuming the free solar power themselves. Feeding it into the grid is usually not economical because kilowatt-hours that are sold to the electricity supplier for a few cents have to be bought back at a higher price when needed. A positive side effect: CO₂ emissions are reduced, and energy grids are less burdened because not all households cover their needs at peak times in the morning or evening.
With EMS, the wallbox at home charges the electric car when a lot of solar power is available. Or the hot water tank heats up to reserve a few degrees for the night. A smart fridge goes the other way round and reduces the temperature in the freezer compartment when energy is cheap so that it can then take a cooling break. Start washing programs so that the energy-intensive heating phase falls at a time when electricity costs are low? Intelligent control of air conditioning systems? Use the display on your smartphone to identify where hidden consumers are lurking? There are many ways in which an EMS can save electricity costs in everyday life.
The cross-manufacturer approach of Matter
So far, so good – and familiar. After all, energy management systems have been around for quite some time. Many companies developed their own proprietary solutions for this purpose. Manufacturer-independent approaches such as the EEBus communication interface (link) or the SunSpec protocol (link) have also been in place for many years.
The charm of Matter lies in its wide range of applications across industries and product groups. The specifications already cover many areas of application, from smart plugs to household appliances and wallboxes. Additional device types that are important for energy management, such as heat pumps, are currently under development. On top of this lies the international approach: hundreds of companies from all over the world have come together within the standardization organization CSA. This gives Matter a good starting position in the competition between solutions.
Measure and display energy consumption
First and foremost, management requires a solid database. For most people, their own energy consumption is a “black box” without transparency that is only opened once a year: when the bill arrives from their energy supplier. With Matter, this previously lacking transparency could now enter every household.
The Matter 1.3 specifications contain an “Electrical Energy Measurement Cluster” (2.12) and an “Electrical Power Measurement Cluster” (2.13.). Clusters are data collections that provide attributes, events and commands for a specific function. Both clusters provide information on electricity consumption. The Power Measurement data tells the current power consumption. For instance: This device is consuming 1500 watts at the moment. The Electrical Energy Measurement provides cumulative values over a certain period of time. For example, it can provide information on how many kilowatt hours (kWh) an appliance consumes per day. How much it needs per month or how much it has used since the beginning of the year.
This enables measuring sockets to transmit the energy consumption of connected devices in a Matter-compatible data format. Possible application: The app of a smart home system displays the power consumption. Programmed actions are also possible. For example, a Matter platform could automatically disconnect all home theater devices from the mains as soon as the TV goes into standby mode – because its energy value falls below a defined threshold.
These scenarios are already possible today – but with proprietary programming interfaces (APIs). In order to build compatible devices, manufacturers have to adapt their software, write drivers or work closely with other companies. The standardization of Matter simplifies the process because the exchange of energy data is already provided for in the specifications.
However, its use is not limited to sockets and power outlets. As these are software clusters, the functions can be integrated into any Matter device with suitable hardware. Domestic appliances would thus be able to provide up-to-date consumption values at any time. A balcony power station would provide information about its output – provided this device type is available in Matter at some point.
According to the specification, special hardware is not even required for measurement. If such components are missing, devices can also estimate the data. The consumption of an LED lamp, for example, is known to the manufacturer. The developers know how this value fluctuates between 0 and 100 percent brightness. This makes it easy to calculate how many watts the light source requires in the respective operating state. A data field in the cluster informs potential recipients whether the values are actually measured or estimated.
Energy Management Systems in Matter
A Matter EMS takes the utilization of energy data to the next level. The display of power data and simple on/off commands become a predictive, intelligent control system. The functions for this are spread across several clusters – from the “Device Energy Management Cluster” to definitions for electric car chargers and wallboxes (Electric Vehicle Supply Equipment, EVSE) to specifications on the extent to which the consumption of a controllable device can or should be optimized.
Where this management takes place technically is left to the ingenuity of developers. It could run as isolated software on its own hardware, for example as a DIN rail component in the electrical distribution board. Alternatively, a computer in the home network can take over the task or another device that performs the functions as a piggyback, so to speak. If the hardware is powerful enough, it would be conceivable to integrate a Matter Controller with EMS into the heating control system – as an additional feature that a heat pump manufacturer offers to its customers.
In any case, the focus is on dynamic control of power consumption. This is demonstrated not least by the new product category EVSE (Electrical Vehicle Supply Equipment) in Matter 1.3, which is the first device class to be included in the standard with a focus on energy. However, even the basic “Device Energy Management Cluster” already contains tools for optimizing, as a glance at the specifications shows:
| Name | Function | Task |
|---|---|---|
| PA | Power adjustment | Allows the energy manager to influence the current power consumption of an appliance within predefined limits. |
| PFR | Power forecast reporting | The appliance creates a forecast of its expected energy consumption for the near future. |
| SFR | State forecast reporting | The device announces a planned status change. |
| STA | StartTime adjustment | Allows the energy manager to influence the start time of a device. |
| PAU | Pausable | Allows the energy manager to interrupt the operation of a device. |
| FA | Forecast Adjustment | Allows the energy manager to adjust the future energy consumption. |
| CON | Constraint-based adjustment | The device responds to external commands to change or adjust its energy consumption. |
Controlling consumption with rules and AI
Optimizing the power consumption of appliances and shifting it over time is easier said than done. Because energy consumption does not adhere to rigid rules. Unlike a heating schedule that raises and lowers the room temperature at fixed times of the day, electricity generation and consumption are subject to a whole range of variables.
When is the sun going to shine or the wind to blow? When can the EV be charged at home because working from home is on the agenda – and when will the car be at the office during the day, possibly returning with a full battery because there are charging stations in the parking lot? Another, increasingly common scenario: the energy supplier’s dynamic electricity tariff offers the prospect of cheap kilowatt-hours for the following day, which makes a boil wash seem advisable. For private reasons, however, it would be better to bring the washing day forward.
Mapping such eventualities with classic smart home rules is difficult and becomes a lot of work. Also, nobody likes to be patronized by an energy manager at home. Instead of dictating daily routines to residents, EMS systems should learn from them and adapt to new situations on their own. This is possible with the use of AI or Machine Learning (ML), as this type of artificial intelligence is called.
Initial approaches can be observed at the Matter pioneer SmartThings, even if they are currently being implemented outside the standard. The Samsung subsidiary has its own energy management system for selected devices. In the future, more and more Samsung products will optimize their consumption themselves: A refrigerator, for example, “learns” how long and how often its door is opened, and at what times of day. Software adjusts the compressor output accordingly and ensures that the refrigerator does not cool down unnecessarily at peak times to prevent the temperature from rising (link). Possible energy savings according to Samsung: up to 15 percent.
Another example: Scientists from the University of Oldenburg, the Fraunhofer Institute IFAM and the German Aerospace Centre DLR have calculated that the efficiency of heat pumps can be increased by more than a third with the help of AI (link). To do so, they simulated the private heating network of a residential complex with a 100-kilowatt heat pump and hot water tank and then had the operation optimized by software using a special form of machine learning. The result: the total energy requirement of the heat pump fell by 15 percent, and in combination with a variable electricity tariff, the electricity cost savings amounted to 35 percent.
Example: wallbox with AI management
The strength of solutions with machine learning: they can recognize and analyze recurring patterns. If correlations are known – or can be well estimated – the system also makes complex decisions and finds a compromise between conflicting goals; faster and better than a human could ever do in a comparable situation. Manual intervention is still possible, for example to set priorities or override the AI’s plans.
The following example from everyday life shows how AI-controlled energy management works. It is currently still theoretical because there are no corresponding Matter products on the market. However, every commuter with an EV and a dynamic electricity tariff faces a practical challenge: how do I charge my electric car at home in the best possible way? There are two competing goals:
- In the morning, the battery should be as full as possible so that the vehicle can also cover longer distances beyond the journey to work.
- The costs should be kept as low as possible because charging is preferably carried out when the dynamic electricity tariff promises a low price per kilowatt.
It is therefore important to find a balance between the two requirements every day. The tariff is not always the same price and the battery level varies when you get home, depending on the distance travelled. What does AI do? It first analyzes the data from the (Matter-compatible) wallbox. Let’s assume that the typical charging process takes place on weekdays between 5 or 6 pm and 8 am. On each journey to work, the battery loses between 20 and 25 percent of its charge.
The system draws up a usage profile from the information and compares it with the energy supplier’s price forecasts. Users can now set how risk-averse they are: do they value the security of a well-filled battery and accept higher charging costs? Or do they allow the energy management system to skip charging during periods of high prices in the hope that the electricity will be cheaper the following day? The AI reacts accordingly, but in any case ensures that the battery charge is sufficient for the journey to work – if necessary, at a higher price than desired.
A key characteristic of AI is its ability to deal with uncertainties. The kilowatt price in dynamic electricity tariffs is only known for the next one or two days. It can therefore happen in individual cases that the algorithms “gamble” and make a bad because more expensive, decision. On a long-term average, however, the calculation works out in favor of the user – because the positive effects outweigh the negative ones.
For this to happen, however, the specifications and technical possibilities first have to find their way into purchasable products – which experience has shown takes time with a comprehensive standard like Matter.
Share this information:
