With the war in the Ukraine, energy prices in Europe reached new record levels. This initially affected the gas price which does not affect me directly. But it had a big impact on the price for electrical energy too. In my village, the price for electrical energy is now at 0.45 CHF/kWh, starting October 1st 2022. It is twice as much as what it used to be, and three times more what it used to be the price for the energy at night time.

Saving energy always makes a lot of sense, now even more, both for the environment and directly saving money. Luckily, I started thinking about optimizing the electrical energy used in my house back in 2021, and now in 2022 it really pays off: The daily average of 16 kWh/day (including heating and cooling) came down to 7 kWh/day, or from 4.5 MWh/year down to 2.4 MWh/year, or a reduction of 47%.

There were different areas contributing to this very positive result:

The above graph shows the changes in the different categories, from 2021 (blue, 4.5 MWh) to 2022 (orange, 2.4 MWh).

Overview

In this article, I’ll show you what we did to reduce the energy consumed. Maybe this gives you some ideas to reduce your energy bill too.

Note: Because 2022 is not over yet, the 2021 data shown is from October 2020 to October 2021, and the data for 2022 is from October 2021 to October 2022, so not the calendar year, but a year cycle.

The building is from 1999, with the heating system replaced in 2013 using a heat pump. The volume is around 100 m3, with an accessory apartment and a dual-garage.

The building used 4.5 MWh electrical energy in 2021, and this was the baseline to optimize. Optimizations have been started end of 2021 and were implemented during 2022, with the goal to reduce energy consumption by more than 30%.

The following main approaches were used:

1. Standby: Reduce stand-by consumption, turn off devices if not in use.
2. Lights: Replace lights with LED based ones.
3. Devices: Replace older energy hungry devices with newer ones.
4. Building: Reduce energy leaks, optimize heating and cooling system.

As a general rule: The first two items are easier to implement then the second two ones. Standby reduction and replacing lights with LEDs easily pays off after one year. Replacing appliances is a bigger investment and should pay off after 5 years. Changing the building infrastructure should pay off after 10 years.

The increased pricing very much helped to justify the investments. For example my energy cost during the day is at CHF 0.4475/kWh and CHF 0.3732/kWh during the night. This means an average of CHF 0.41/kWh.

1 Watt (e.g. standby) for the full year means 8.76 kWh (24*365=8760h). With a price of 0.41 CHF/kWh this means 1 Watt used all the time costs CHF 3.50 every year (8760*0.41=3.59). As there are usually many devices drawing energy even if not used (‘standby’), this easily sums up to a larger amount of money at end of the year: Every Watt counts!

💡 Of course time for return of investment depends on the local energy prices and many different factors. But many things in life (taking a vacation, buying a new gadget, driving by car compared to taking the public transportation, diner in a restaurant, …) do not ‘pay back’ in the monetary sense. So maybe the question should not be ‘does it pay back?’ but instead ‘is it worth the money?’.

Energy Flow

In a typical building, most energy is used by the heating system. Back in 2013, we replaced our fossil oil heating system with a electrical heat pump, combined with 29 m² solar thermal (ST) panels, see “Hacking the Heating System for Cooling – Geothermal Drilling with extra Benefits“. This system includes a free-flow cooling during hot summer days. So compared to similar systems, the panel size is rather large to cover all warm water needs during the summer and to support the heating system during winter time. Every year, the system collects around 19 MWh thermal energy.

On the electrical side, I installed a Photovoltaic system back in 2012, which produces around 5 MWh every year. Below it shows the energy flow for the year 2020:

The TS system is covering most warm water needs, and the heat pump needs around 1500 kWh electrical energy during the winter months. The PV system covers most of the electrical energy, with a net grid usage is around 1000 kWh.

Putting the usage of electrical energy into different buckets, this showed the following distribution in 2021:

Going forward, the following goals were set:

1. Reduce the 1500 kWh needed for heating. Changing windows from dual to triple glazing. This reduces the need for cooling and needs less heating in winter time. The winter 2022/23 did not start yet, so I have to verify the overall effect on the heating system, but the estimate and current measures indicate a reduction by 30%.
2. Reduce the 4500 kWh needed by the building: turning off devices and replacing old appliances with new ones.
3. Reduce grid usage with a battery system: this does not reduce the energy usage itself, but reduces costs as the price to get a kWh is higher than feeding one back to the grid.

For point 3) a battery system from Tesla has been installed.

Tesla Battery System

Why a battery system? Because it allows to move energy from the day into the night time, to take advantage of the produced solar energy. The calculated cost of a kWh solar energy over 20 years (typical life time of the system) is at 0.11 CHF, or CHF 110/MWh.

But as you can see from the reference table below, the price I get for a MWh has been always below that point, with the exception starting from 2021/4:

So back in the middle of 2021, we decided to store and use the energy with a battery system, so we don’t have to sell it during the day at low prices and buy it back later at a much higher price. Back in 2020/2 we had to feed back to the grid the 10x amount of PV energy for the price of a kWh from the grid.

The idea is to self-consume the PV energy, and only feedback the surplus, minimizing the grid usage down to zero.

Measure it

So we needed to change things. There is one simple truth about change:

If you can’t measure it, you can’t improve it.

To measure power consumption, an easy way is to use a power monitor plug like the one below.

Additionally, Shelly EM were used to measure the energy used in the house:

I can log the data for example using HomeAssistant:

Beside that, my energy provider EWS has a portal where I can check my grid usage. Data is only provided every 15 minutes, but this is still very helpful. Below the data from 28-Oct-2020:

Yellow is grid usage, and green is what is provided back to the grid. Notice that there is kind of base grid usage around 350 Watt.

Below the same day, one year later (28-Oct-2021). The grid usage is pretty much down to zero, because of the installed battery system:

Below the data from the same day this year (28-Oct-2022), which is similar, except it was not that much of a sunny day:

So instead of feeding back energy to the grid and then get it back later at a higher price, the battery system helped to optimize the grid usage and to self-consume the PV energy. To do this, a Tesla PowerWall 2 system was installed.

Tesla PowerWall 2

The Tesla Powerwall 2 is a system with a battery to store energy, so it can be used later. In Switzerland it is only allowed to store your own PV energy, so you cannot use it to store energy from the grid and use it later.

With the system, it allowed us to have nearly zero energy usage from the grid, since March 2022 until now (mid Nov. 2022). ‘Nearly zero’ means that there are still around 3 kWh/month from the grid, mainly because of the 1 Hz control loop of the system: so for a very short time it might draw current from the grid and will return it later, but depending on the smart meter some little amount of energy still count. So there no 100% independence from the grid, but a very reduced one.

The gateway is the smaller box above the battery with 13.5 kWh capacity:

The gateway manages the flow of energy and has current transformers (CT’s) to measure the energy flow:

The return of investment depends of course of the grid pricing. But assuming a price difference of CHF 0.30 between the PV production costs and the grid costs (which is the case these days) and a grid flow in and out of 5000 kWh (which I would have without the battery), the system pays back in 8 years, which is well within the lifespan of the system (20 years, at least on paper).

The Tesla Smartphone app visualized the energy flow, the energy used with the energy costs. Additionally it provides power backup (one phase only) and lists the grid outage

With the App I can see how the energy is used during the day (green: battery, yellow: solar, grey: grid). Left it shows that the building used 7.2 kWh with 44% from solar, 51% from the battery and 4% from the grid. On the right it shows how the 20.5 kWh from the PV system is used: 16% goes to the house, 29% to the battery and 55% back to the grid:

This provides good information over time, how the energy is used. However, the resolution in the app is 100 Watt, and everything below that is shown in the graph as zero. I believe this is only for reducing the amount of data in the display, because the system is measuring the power with a 1 Watt resolution.

To get a better visibility, I’m using the data from the PowerWall directly with a web interface integration. Below the data for the same day as above (28-Okt-2022) which shows the power used by the house.

You might notice that the 350 Watt ‘standard power usage’ from 2020 has been reduced to about 25 Watt :-).

Later early 2022, the Tesla battery system has been extended with two more batteries (you can easily stack up to three batteries). This not only extended the capacity to 40.5 kWh, but helped reducing the battery cycles and therefore extending the lifetime of the batteries. The extended capacity is helpful if we would switch to an EV (electrical vehicle) in the future. Another reason is better balancing of the 3-phase grid and supply: Each single battery block or PowerWall has an inverter, capable to charge and discharge up to 3.6 kW. The current PV system can deliver up to 5.8 kW, so it means not all energy can flow into the batteries. With 3 batteries, one on each phase, it can charge with up to 10.8 kW peak, so will be helpful as well if we would extend the existing PV system.

Standby

In a typical household there are many devices plugged in, and while ‘doing nothing’ or ‘turned off’, they still draw power.

Here a power strip with a switch are very useful. I like the ones with a ‘remote on/off’ switch which are ideal for a working desk:

Or the ones with the switch integrated:

Don’t used them if they are always turned on, as the switch light can consume up to 1 Watt too!

So the idea is to go around the house and check with a Watt-meter what power all the devices are using: from charging devices, audio/video players or desktop docking stations. And if they can be turned on and off with a power strip.

Prior 2021 we already used some power strips, but now everything which can be turned off is behind a switch. For example the TV and the Sonos system is turned off, saving 30 Watt standby or 260 kWh/year.

I added a switch to my desktop (docking station, two displays, USB hub) to reduce standby power by 5 Watt, saving around 32 kWh/year.

And it makes a difference if working from home or in the office: with a laptop, a docking station and two displays it needs around 125 Watt while working, so this easily sums up to 1 kWh a day.

Lights

Another step was to replace the lights with new ones based on LED technology. Overall this has cut the energy used by half. For most lights there are simple replacements. For example in the workshop and in the garage the old 36 Watt tubes have been replaced with 15 Watt LED replacements, with the electronic ballast device removed.

Many lights were replaced by ‘smart lights’ from the Philips Hue system:

They communicate with the ZigBee Light protocol and compared to other ‘smart lights’ they need only 0.1 Watt standby.

There old halogen spots have been replaced with LED versions instead, coming down from 45 Watt to 6.1 Watt each

The 12V halogen spots with GU5.3 socket used a electronic switching device, which needed a load of at least 10 Watt, so a simple light replacement did not work. One solution would be to replace the electronics with a new one, but it was more effective to use a new GU10 230V socket instead.

The LED technology uses about half of the energy compared to the lights I had before, so saved around 400 kWh/year.

One downside of the Philips Hue portfolio: they focus only on lights. But they do have as well a ‘Smart Plug’ which can be used to turn on/off devices:

I really like these, because they have much less standby current (around 0.1 Watt) compared to WiFi plugs which draw around 1 Watt or more in standby. The plug does not measure the power, it is just an on/off device, with a button on the side to turn it on/off manually. With the Hue plug I can turn on/off devices like power supplies or non-Hue lamps, or other devices which are not in use.

The other very useful device is the Philips Hue motion sensor:

It not only detects motion, but measures temperature and ambient light level. Compared to other motion sensors (e.g. the one from the Shelly), they are really excellent, and I can configure the motion and ambient light threshold. The motion sensors are use to automatically turn on devices if presence is detected, and to turn them off after a timeout. Depending on the time of the day, the lights are turn on differently, and the lamp light level is adapted to the ambient light level.

Honestly, I did not expect much of savings, but actually it did. The automatic off and the adjusted light level resulted in savings of around 150 kWh/year.

Electronics

Newer electronic parts can reduce the energy bill. For example I have replaced the 5 Watt of two 5 port switches with a single new one which uses only 1.4 Watt: The investment of CHF 25.00 for the new switch pays off in less than a year.

Another topic are wireless devices: they are very useful, but the wireless connectivity needs more energy compared to wired devices.

For example I’m using Shelly WLAN devices which consume around 1.2 Watt in standby mode, so much more than the Philips Zigbee Hue devices (0.1 Watt).

For the WiFi I need an Access Point (AP). So the first thing to do was to replace the old (10.3 Watt) with a new D-Link DAP-2020 (2.1 Watt). Because in the garage I did not had any WiFi coverage, I added a small AP (MikroTik) there which only needs 1.2 Watt. This change alone saved 60 kWh/year.

One thing not easy to change is the internet modem/router. The Sunrise ADSL modem had a setting to turn off the lights, saving 1.1 Watt, or 9.6 kWh/year.

Kitchen

The old microwave oven has not been replaced (yet). But because it uses 5 Watts (!!) in standby, it is now turned on with a power strip to get rid of the standby. In contrast: the coffee machine turns off automatically after a timeout and then draws zero Watt.

Turning the microwave off saves 43 kWh/year.

A big impact had new appliances in the renovated kitchen: the new fridge uses less then half of the energy, and the oven has a better insulation making it 30% more efficient.

All the new devices had the ability to connect to the cloud: this of course has been turned off, saving around 2 Watt on each device. The devices have a ‘night’ mode which turns off displays and other things to save energy. Below the steamer (top) and oven (bottom), the display/clock for the lower device has been turned off. Many dishes cook healthier and need less energy if cooked in the steamer instead of using the oven.

The new dishwasher has a ‘SolarSpar’ program, reducing energy from 1 kWh to less than 0.1 kWh per washing cycle: it does not heat up the water and instead uses the warm water from the solar system. This saves 279 kWh/year.

The new fridge (Miele K7473) needs less than half of the energy (87 kWh) compared to the old one (247 kWh). It has been configured to 7° C instead of the usual 5° C. Result: 160 kWh/year less.

The other big savings with 330 kWh/year was the replacement of the old freezer. It was over 20 years old, and needed a de-icing every month.The new one does an automatic de-icing, is more efficient has been set to -18°C instead of -20°C.

Building

The house had double glazed windows which were usual more than 20 years ago. So improve the overall efficiency, most of them have been replaced with modern triple glazed windows:

The new ones not only pass through more light which is important during winter time for passive heating. In terms of energy savings, the new windows are around 30% better than the old ones, reducing not only the need for cooling during hot summer time, but saving as well around 500 kWh/year for the heat pump during winter time.

Other building infrastructure like pumps or motors had to be looked at too. The motors for the garage doors had a light which already had been replaced with a LED one. But the motor still used 12.5 Watt in standby, causing 2×12.5=25 Watt which equals to 219 kWh/year.

The first approach was to turn it off using a Philips Hue Plug (0.1 Watt) plus a Shelly WLAN Switch powered by a 12 V Power supply (2 W):

To open/close the door remotely from the car I used a Shelly WiFi Button:

But there were two problems with that approach: it required an extra WiFi access point in the garage (+ 6 Watt). So I needed 6 + 2*(0.1+2) = 10.2 Watt to save 25 Watt of standby. Still not bad, but the WiFi buttons did not work well enough from the distance. Finally, my wife did not like the buttons. So I ended up replacing the 20 year old garage door motors: the new ones only consume 2 Watt in standby and have LED lights built-in:

From previous 25 Watt down to 4 Watt, saving 184 kWh/year.

Another optimization area is the heating system. In my case, the heat pump runs usually only from end of November to early February, so no sense to feed the electronics during the other months, cutting off the 5 Watt standby, saving around 32 kWh/year.

During the 3 months heating period, the circulation pumps have been put into ‘automatic night and pressure control mode’, reducing wattage to 5 – 7 Watt, depending on the water temperature (lower during the night).

So during the heating months, the previous 18 Watt has been reduced to less than 7 Watt, saving around 23 kWh/year.

To turn on/off 3-phase powered devices (washing machine for example), 3-phase Shelly power switches have been installed.

One point to consider: the Shelly Pro 3 needs 1.5 Watt standby too, so it only pays off if the devices controlled by it have a higher standby wattage.

Automation

The home automation is run by a combination of Philips Hue (Zigbee) LAN Gateway (1.6 Watt), a Raspberry Pi 4 running HomeAssistant (4 Watt) and a D-Link DGS-108 Switch (1.4 Watt), so a total of 7 Watt.

With a local and remote accesible dashboard the energy flow of the house is monitored and controlled.

It nicely shows the energy distribution with an animated graph:

HomeAssistant is great for creating custom automation. For example it can turn off a power supply in the kitchen after 3 minutes if no presence is detected.

That power supply would consume otherwise 2.5 Watt all the time. That way the ‘on’ time (green below) has been reduced to below than 20% a day. Savings: 17.52 kWh/year.

A similar rule is used to turn off the internet modem at night, which otherwise would consume 11 Watt all the time:

Savings: 24 kWh/year.

Summary

With this, savings were in all categories. The washing machine and dryer did not contribute much (except some standby reduction), as the appliances were already recent.

The distribution changed so that heating and cooking are, because they were already big pieces and the other areas have reduced the consumed energy a lot

Looking what contributed most to the savings, the rough numbers are:

• 750 kWh with newer appliances
• 500 kWh with optimized Lights
• 350 kWh with reduced standby
• 500 kWh with improved windows

Most of the savings can be realized with simple things like using LEDs and reducing standby power, which have an immediate effect. Using newer appliances and devices makes sense with a 3 to 5 year time in mind, and updates to the building like new Windows have an even longer time span. The ‘smart home’ part with HomeAssistant is helping, but to much less extended around 10%: because the intelligence needs power too, and every WiFi ‘IoT’ thing consumes energy too. So they are not a game changer, but can add automation and convenience to the mix.

If you want to save energy and money: invest in LED technology and switch off devices with a power strip first to reduce standby usage. Then think about replacing devices or appliances, considering age of devices and how much better newer ones are. Use ‘smart’ devices only in cases where it makes sense and where you master the technology, otherwise it can be a waste of time, money *and* energy.

So what do you think? Do you have different ideas or suggestions, or questions? Post a comment an let me know…

Happy Savings 🙂

Links

• Home Assistant: https://www.home-assistant.io/
• Shelly: https://www.shelly.cloud/
• Tesla Powerwall: https://www.tesla.com/powerwall
• Philips Hue: https://www.philips-hue.com
• 3D Printed Mounting Bridge for Feller Smart Light Control for Philips Hue
• Building a DIY High Density Philips Hue LED Stripe
• Full Bathroom Remodeling (with the Help of a 3D Printer)