Background<\/h2>\n\n\n\n
On 1st<\/sup> of March 2024, I switched from a traditional electricity provider to one with dynamic day-ahead pricing, in my case, Tibber<\/a>. I wanted to try this contractual model and see if I could successfully manage to shift chunks of high electricity consumption such as:<\/p>\n\n\n\n to those times of the day when the electricity price is lower. I also wanted to see if that makes economic sense for me. And, after all, it is fun to play around with data and gain new insights.<\/p>\n\n\n\n As my electricity supplier, I had chosen Tibber<\/a> because they were the first one I got to know and they offer a device called Pulse<\/a> which can connect a digital electricity meter to their infrastructure for metering and billing purposes. Furthermore, they do have an API [1<\/a>] which allows me to read out my own data; that was very important for me. I understand that meanwhile, there are several providers Tibber<\/a> that have similar models and comparable features.<\/p>\n\n\n\n The graph below shows five curves and is an upgraded version of the respective graph in [6<\/a>] visualizing data over two weeks:<\/p>\n\n\n\n As in [6<\/a>], the green<\/strong> curve is the hourly price of the day-ahead market. It is well recognizable that the price has peaks in the evening and in the morning at breakfast time when residential consumption is high, but little solar energy is available in Germany. The yellow<\/strong> curve is a two-days floating average and shows that the average price is not below a fixed rate electricity contract, an important point that we shall discuss later. The orange<\/strong> curve is the consumption of my house; the higher peaks indicate times when I charged one of the cars using an ICCB<\/a> (one phase, current: 10 A). The red<\/strong> curve shows the consumption of the grid. During the day, when there is sunshine, the red curve lies below the orange curve as a part of the overall consumed electricity comes from the solar panels. At nighttime, the red curve will exactly lie on the orange curve as there is no solar electricity generation. The blue<\/strong> curve shows the average electricity price per kWh based on a mixed calculation of the grid price and zero for my own solar electricity generation. One might argue whether zero<\/em> is an adequate assumption as also solar panels cost money, but as I do have them installed already, I consider them to be sunk cost<\/em> now. The blue curve show an interesting behavior: When my consumption is low, the blue curve shows a small average price. When the consumption of the house is less than the power that is generated by the solar panels, then the blue curve is flat zero. When, however, I consume a lot of power, then, the blue curve approaches the green curve. At nighttime when there is no solar energy generation, the blue curve lies identical with the green curve.<\/p>\n\n\n\n My goal is to consume more electricity in the times when either the green curve points to a low electricity price or when there is enough electricity generated by the solar panels so that my average price that I pay (blue curve) is reasonable low. This also explains why I mostly the ICCB<\/a> (one phase, 10 A) to charge the cars as then, I still can get a good average price (although charging takes a lot more time then). I think that by looking at the curves, I have adapted my consumption pattern well to the varying electricity price.<\/p>\n\n\n\n The graph above shows four curves. The green<\/strong> curve is the cost that I would have incurred if I had bought all the electricity of the respective day in the hour of the cheapest electricity price. This would only be possible if I had a battery that could bridge the remaining 23 hours of the day (and probably some hours more as the cheapest<\/strong> hour of the following day is not necessarily 00:00-01:00). The red<\/strong> curve is the cost that I would have incurred if I had bought all the electricity of the respective day in the hour of the most expensive<\/strong> electricity price. The yellow<\/strong> curve is the average<\/strong> cost of the respective day, the multiplication of the average price per kWh of that day by my consumed energy. The blue<\/strong> curve is the real cost that I have paid. If the blue curve is between the yellow curve and the green curve, then this is very good, and I have succeeded in shifting my consumption versus the hours with cheaper electricity. Without a battery, it is almost impossible to come very close to the green line.<\/p>\n\n\n\n The graph above shows one peculiarity as on 2024-06-26, there were some hours with an extremely high price, but that was due to a communication error that decoupled the auction in Germany from the rest of Europe [7<\/a>], clearly (and hopefully) a one-time event.<\/p>\n\n\n\n The one-time event with unusually high prices [7<\/a>] is well visible in the price heatmap that was already introduced and explained in [6<\/a>]. The one-time event overshadows all price fluctuations in the rest of the week.<\/p>\n\n\n\n The first graph shown below was already introduced and explained in [6<\/a>] and shows the cost of purchased electricity from the grid in 24 rectangles, whereby each rectangle represents an hour. The order of the rectangles starts with the most expensive hour at the top and ends with the least expensive hour at the bottom. The larger the rectangle, the more money has been spent in the respective hour. I already explained in [6<\/a>] that the goal should be – if electricity has to be purchased from the grid at all – the purchases ideally should happen in times when the price is low. In the graph those are the rectangles with green or at maximum yellow color. The rectangles with orange and red color indicate purchases during times of a high electricity price. In reality, one will not be able to avoid purchases at times of a high price completely. My experience is that in summer, I let the air conditioning systems run also in the evening hours when the electricity price is higher, just to keep the house at reasonable temperatures inside. Similarly, in spring and autumn, I opt for leaving the central heating switched off and try to heat the rooms in which I am usually with the air conditioning systems (in heating mode) as in spring and autumn, the difference in temperature between outside and inside is not too high, and the air conditioning systems will have a high efficiency.<\/p>\n\n\n\n Then, for 13-Oct (10\/13), one can see that in fact I bought substantial electricity at high prices. The reason for that was that I returned home at night and had to re-charge the car for the next day. So, one cannot always avoid purchasing electricity at high prices.<\/p>\n\n\n\n The next chart offers another view on the same topic. Rather than looking at the cost of the purchased electricity, we look at the amount of purchased electricity. This offers an interesting insight. Except for 13-Oct, the amount of electricity that is purchased at hours with a high price, is less than 5 kWh. This means that with an energy-storage system (ESS), it should be possible to charge a battery of around 5 kWh during times of a lower electricity price and to discharge the battery and supply the house during times of a high electricity price so that ideally, there would be no or only minimally sized orange and red rectangles. Of course, this only makes sense if the low prices and the high prices per day differ substantially as the ESS will have losses of 10%…20%. And in fact, this is exactly an idea that I am planning to try out within the next months (so stay tuned).<\/p>\n\n\n\n Another finding which was interesting for me was that the price of electricity is not necessarily cheaper on the weekend (at least not in May 2024). One might assume that because industry consumption is low on the weekend, the price of electricity is lower. And that is true to some parts, as the lowest prices that occur within a whole month tend to happen on the weekend (green<\/strong> histogram). We can also see a peak at around 19 \u00a2\/kWh in the green<\/strong> histogram which is much lower in the yellow<\/strong> histogram. However, there are also many hours with prices that are similar to the prices that we have during the week (yellow<\/strong> histogram).<\/p>\n\n\n\n The following three graphs show the hourly price curves over the hour of the day per month; the selected Day of the Week<\/em> is irrelevant for this graph as only the variable Month<\/em> is used.<\/p>\n\n\n\n The green<\/strong> curve shows the minimum hourly price per kWh that occurred in the selected month. The red<\/strong> curve shows the maximum hourly price per kWh that occurred in the selected month. The yellow<\/strong> curve shows the average hourly price per kWh that occurred in the selected month. The blue<\/strong> curve shows the average hourly price per kWh that I have experienced in the selected month, calculating my own solar electricity generation as “free of cost” (like the blue curve of the graph in chapter Price and consumption patterns<\/strong>).<\/p>\n\n\n\n There are a couple of interesting findings that can easily be seen in the curves:<\/p>\n\n\n\n For me, this is an indication that it might make sense to install a battery and charge it during the day and discharge it especially in the evening hours (18:00-21:00) when the price peaks.<\/p>\n\n\n\n In order to look into the hourly prices from a different perspective, I have grouped the hourly prices into five categories:<\/p>\n\n\n\n For each month (except for Oct 2023, Oct 2024) the sum of the blocks in different colors are the hours of the respective month; depending on the number of days per month, the number of hours vary, of course. Larger blocks correspond to more hours in the respective price category. Do not pay attention to the numbers listed in the graph in white<\/strong> color as some numbers are missing. I will give a numerical statistic below the graph.<\/p>\n\n\n\n\n
Findings<\/h2>\n\n\n\n
Price and consumption patterns<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Electricity-Price-and-Consumption-new-1024x425.jpg\")
Actual cost, minimal cost, maximal cost, average cost per day<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Actual-cost-versus-minimal-cost-maximal-cost-average-cost-per-day-1024x505.jpg\")
Price heatmap of the hourly prices per kWh<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Heatmap-of-the-Tibber-flexible-hourly-prices-per-kWh-1024x505.jpg\")
Cost and amount of purchased electricity from grid with price indication<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Cost-of-purchased-electricity-from-grid-with-price-indication-1024x444.jpg\")
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Purchased-energy-from-grid-with-price-indication-1024x445.jpg\")
Hourly price distribution<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Histograms-1024x503.jpg\")
Monthly hourly price curves<\/h3>\n\n\n\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Hourly-Price-curves-in-2024-05-1024x468.jpg\")
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Hourly-Price-curves-in-2024-07-1024x466.jpg\")
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Hourly-Price-curves-in-2024-09-1024x468.jpg\")
\n
Electricity price levels<\/h3>\n\n\n\n
\n
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Electricity-Price-Levels-1024x466.jpg\")
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Self-sufficiency-and-self-consumption-rate-1024x503.jpg\")
![\"\"](\"https:\/\/caipirinha.spdns.org\/wp\/wp-content\/uploads\/Tibber-Flow-300-1024x406.jpg\")