Energie-Erneuerungsplanung

Topic: Algorithmic solution for valuing consequences in infrastructure maintenance

This project reflects this challenge

Challenge

Maintenance of company assets of the low-voltage power distribution grid should be prioritized based on asset state and asset importance. While the state of the asset, typically the age since the last maintenance, is well-document, there is currently no measure of importance for these assets. One idea to value the importance of the assets is to analyze the amount of energy that is distributed by them. The amount of consumption is, of course, not an optimal measure for how damage is caused by an asset breakdown, or for what consumer would pay for an insurance against power interupts - if they were allowed to pay for it. This challenge, therefore, leave it open how, for the purpose of optimization, the importance will be measured.

Provided Data

The data that was provided consisted of - List and point coordinates of transformer stations - List and point coordinates of distribution enclosures - List and point coordinates of consumers - List and line coordinates of connection lines - Anonymized consumption data for the last four years

Approach

Since no importance measure was given, but consumption can be considered a first good proxy and consumption is only known per consumer but not per connection line, distribution enclosure or transformer station, the first part of the project was dedicated to enriching the data by calculating these figures.

The second part of the project focussed on the optimization of the prioritization of maintenance work based on a flexible importance measure.

Data enrichment

The basic idea is to calculate the amount of energy that flows through transformer stations, distribution enclosures and connection lines by analyzing the topology of the network, starting from the transformer stations. This topology is constructed from the point and line coordinates. The quality of the coordinates is high enough to do this correctly.

After the topology of the network and the distrance from the transformer stations is created, the yearly power consumption is aggregated for each element of this network starting with the ones that are the furthest away from the transformer stations. The maximum depth of the network is 17 hops.

The data is visualized in an open source software QGIS. This calculated data is added to the network elements such that the visualization can be enhanced by this data. The size of the points and the width of the lines now corresponds to the amount of energy that flows through this element.

This is a larger view of the results. (brown - transformer station, red - distribution enclosure, orange - consumer, red - connection line)

This is a more detailled view of the results.

Optimization

References

http://geopandas.org/reference.html

openData_eniwa

Wir haben uns bei der Erstellung der Blackoutkarte vorerst auf das Medium Elektro beschränkt, wobei der Ansatz grundsätzlich auf andere Medien übertragen werden kann und auch für andere Grundlagedaten angewendet werden kann. Herausforderung war und wird sein passende Grundlagedaten für die Analysen zu finden. Als Tool wurde vorwiegend QGIS verwendet

Als erstes haben wir verschiedene Gefahren definiert.

Als nächstens haben wir die unterschiedlichen Karten vereinheitlicht und normert, damit sie im nächsten Schritt miteinander verrechnet werden können. So wurden die unterschiedlichen Farben in hohes Risiken mit dem Wert 200, mittleres Risiko 100 und geringes Risiko 50 konvertiert. Die normierten Risiken wurden nun zusammen addiert. Die Risiken können beim addieren noch unterschiedlich gewichtet werden. In unserem Beispiel haben wir die beiden verwendeten Risiken gleich gewichtet.

Die bestehenden Leitungen inkl. Einteilung der Wichtigkeit der Leitungen lag als CSV vor und wurde mittels SpatialLite in QGIS hinzugefügt. Die gewichtete Risikokarte wurde mit den Leitungen multipliziert, wodurch die neuralgischen Punkte aufgrund Risikowahrscheinlichkeit und Relevanz der Anlagen/Leitugen ermittelt werden konnten.

Die gewichtete Risikokarte wurde weiter in QGIS als 3D-Map visualisiert und mit den Leitungen überlagert. Das Risiko bestimmt die Höhe des Geländes, das heisst je höher ein Punkt, umso wahrscheinlicher ein Vorfall.

Die Berechnungen wurden in Python umgesetzt und als QGIS Analysen

PV-Potential on SBB-Areas

The goal of the project is to assess the PV-potential of areas owned by SBB. We use a dataset from SBB containing: Jahresverbrauch, Anschlüsse, and Gebäude-shapes, for each address. On this dataset we did a general potential analysis over all addresses. For a more detailed assessment and visualization of potential "Zussammenschlüsse" of adjoining parcels we linked the SBB data with the swiss energy planning app of geoimpact which collects, links, and visualizes several open datasets (e.g., from Sonnendach, Elcom, ...)

General Potential Analysis

PV-potential over all SBB-areas:

**** 153 GWh (LCOE<10 Rp/kWh) --> 0.3% of total swiss electricity consumtion ****

**** 270 GWh (LCOE<15 Rp/kWh) --> 0.5% of total swiss electricity consumtion ****

CO2 reduction potential over all SBB-areas:

**** ~7000 tons of CO2 reduction (LCOE<10 Rp/kWh) ****

**** ~12000 tons of CO2 reduction (LCOE<15 Rp/kWh) ****

Figure: Total electricity generation potential from PV: 130 GWh (LCOE<10 Rp/kWh), 270 GWh (LCOE<15 Rp/kWh). Ranking of Gemeinde by total potential area for PV installation (roof area that intersects and within the parcel):

Figure: Potential vs. PV LCOE for each parcel. With Gemeinde names where the parcel is located in. Parcels on on top left corner (high potential, low LCOE) shall be looked into:

Data and code: https://github.com/magnetilo/max-pv-strom-sbb/blob/master/SBB%20ZEV%20-%20levelized%20cost%20of%20electricity.ipynb

Assessment and Visualization of parcels and potential "Zusammenschlüsse zum Eigenverbrauch (ZEV)"

When you want to find ZEV you need to consider the geographic locations of the SBB-parcels, most importantly if two parcels are linked to each other through parcels that belong all to the same owner. In the case of SBB the ownership of train-track-parcels is especially interesting as they link different parcels over a large range (virtually over the whole country), which allows potentially for huge ZEV (see figure "SBB Areale & Möglichkeiten ZEV"). However, it is not clear yet if the connection across train tracks is physically feasible and makes economically sense.

We approach the assessment of the potential for ZEV with a qualitative model considering the aspects: "Energie", "Wirtschaftlichkeit", and "Realisierbarkeit". We divided our team into three subgroups working on the topics data, model, and visualization (see figure "Teamaufteilung").

The main task of the data team was the embedding of the SBB data into the swiss energy planning app, where several nasty problems arose.

The model team developed a qualitative model based on quantitative data for assessing the potential of each SBB parcel (or potentially any cluster of parcels, see figure "Modell").

Finally, we visualized the potential assessment and all the information of interest for each parcel on a interactive map in the swiss energy planning app (see figure "Swiss Energy Planning").

Videmus

Es geht darum, ein Tool zu entwicklen für die Bürgenpartizipation beim Melden von Zuständen (Schäden) an der Versorgungs-Infrastruktur.

Recherchen haben ergeben, dass es solche Tools bereits gibt, so z.B. von der Stadt Zürich (zueriwieneu.ch), als Teil der SBB Mobile App, von den IBB (ibbrugg.ch) und weitere. Diese erfüllen ihren Zweck offenbar gut, haben aber gewisse defizite. Zum einen sind sie nicht einfach zu bedienen, zum anderen werden die Daten nur isoliert betrachtet und nicht öffentlich verfügbar gemacht.

Ziel unseres Ansatzes ist es, solche Daten im Sinne von Open Data öffentlich verfügbar zu machen. Das Team hat ein Kollaborationsmodell entwickelt, das den Wert der Veröffentlichung solcher Daten aufzeigt. Die Besitzer/Betreiber der Infrastruktur haben grosses Interesse an solchen Meldungen, da es ihnen erlaubt, fehlbare Zustände schneller und mit weniger Personalaufwand zu erkennen. Zudem ist es für sie interessant, solche Daten öffentlich zugänglich zu machen, denn ein Betreiber allein hat nur wenige Daten und mit Zugriff auf die Daten aller Betreiber lassen sich aussagekräftigere Datenanalysen erstellen.

Wir haben geprüft, ob die Bürger für solche Meldungen eine Belohnung erhalten sollten. Solche Belohnungen können z.B. in Form von Migros Cumulus Punkten erfolgen; bezahlt von den Infrastrukturbetreibern. Umfragen haben ergeben, dass es vermutlich besser ist, mit der intrinsischen Motivation der Bürger zu arbeiten. Man will ja auch nicht, dass die Bürger Objekte beschädigen, um dann eine Belohnung für eine Meldung zu bekommen.

Zudem haben wir mit einfachen Mitteln ein Online-Formular erstellt, mit welchem solche Meldungen gesammelt werden können. Die gesammelten Daten veröffentlichen wir als CSV.

Siehe auch die Schlusspräsentation: https://github.com/kreba/videmus/blob/master/videmus_4.pptx?raw=true

Data

Online Formular: https://form.jotformeu.com/90455619251357

Gesammelte Daten: https://docs.google.com/spreadsheets/d/e/2PACX-1vQAlHdhghXRPd7bSL8xOZZ2jyKiDd52i6nc4sGKIBZPOAjEgzGrnp94-cFBAqF-LDNQyfHjjBrAJYq1/pub?output=csv

Landkarte: https://kreba.github.io/videmus/

VisiStrom

Visistrom helps individuals to plan using electricity at a good time in order to use sustainable energy as much as possible. In turn, this behaviour helps utility companies to reduce energy purchases from expensive sources like coal and oil plants.

Energy Hack Days

This project emerged from the following Open Energy Hack Days challenge: Visistrom challenge.

16.02.2019, 08:00 - Second Update: A basic version of the prototype is running.

15.02.2019, 18:00 - First Update: We realized that it is not just about consuming as little as you can - it is also important that when you consume the electricity, you want to avoid coal-produced electricity. Some solar energy is being produced, sometimes it is good to use electricity when the sun is shining. We are trying to come up with a good visualization to tell this message. There are three recommendations for the end user: now is a good time to use electricity, now you should save, or just keep going. We are quite advanced with visualization, we have some nice data we can use, but the challenge is how to decide when to recommend using electricity.

Design Challenge

How might we visualize peaks, valleys and provenance of electric power, so that end consumers know when consumption and production is most ecological and/or cheapest?

Business Goals

Minimize the necessity to buy expensive, "dirty" electricity (produced from oil and coal).

• reduce peaks (peak shaving)
• reduce unplanned differences in consumption and production

Data Analytics

To make a recommendation of when to use electricity or not, we need to try to avoid peaks in consumption, at the individual household level, and at the area level.

1. Consumption for a given household

Predict consumption peaks and valleys for a given time based on the household's past consumption: - for a given hour during the day (e.g. from 3pm to 4pm) - on a given day of the week (e.g. Saturday) - in a given month of the year (e.g. July) --> not using that at the moment yet

At the moment, the front-end app prototype uses the non-processed usage data from last year as a proxy prediction.

Data source:

Aggregated load profiles of apartment blocks 01. Aug 2017 - 31. Aug 2018

These are Smartmeter measurments in 15-minute intervalls by blocks of flats (usually about 3-8). It's not by individual households, but this is a good enough apporximation for now.

2. Consumption for Kanton of Aargau, SwissGrid

As a proxy for prediction data from the utility provider Energieübersicht 2018, Swissgrid (Kanton Aargau) Scaled down to match 01.01.2018-31.12.2018

3. Consumption for a given area (TO DO)

Predict consumption peaks and valleys for a given time based on the area's past consumption: - for a given hour during the day (e.g. from 3pm to 4pm) - on a given day of the week (e.g. Saturday) - in a given month of the year (e.g. July)

Data source: - Aggregated load profiles of apartment blocks (summed up) - Or alternatively: load profiles from Trafo stations (not openly available)

4. Solar panel production for a given household

Predict production peaks and valleys for a given time based on sunlight and on the household' past solar panel production: - for a given hour during the day (e.g. from 3pm to 4pm) - on a given day of the week (e.g. Saturday) - in a given month of the year (e.g. July)

Data sources: Solarenergie Prognose from NETSTRANSPERNENZ

Optional data to refine the prediction at household level: Past production of solar electricity for this household

Functionality

The functionality purpose for the consumers is: - to adapt their behaviors to have cleaner energy usage - to see their progression in energy consumption over time

To adapt their behaviors, accessible data of factors impacting energy production will be shown to the user. Such include prevision of solar intensity for the upcoming week, coupled with other data such as past average consumption habits of the Club. Data can also integrate predictable special events susceptible to induce consequent perturbation in the general consumption (e.g. World Cup Football).

The integrated data is presented to the user in the form of a table showing predicted energy conditions for the upcoming week, for every hour. The table will highlight the day and hours when clean energy will be the more available among the week. This will enable users to plan their energy most consuming tasks (usually cooking and housekeeping) according to external energy availabiity conditions.

The user will be able to click on the case of the table to access to more details explaining the reasons why he/she should stay to his current task planification or move it to anothe time: - sunlight availability - saturation of Club - etc.

To see their progression over time, consumers will have access to the profile curve of their past days, weeks and months consumptions. They will also be able to see the difference between their actual daily consumption of a specific weekday with average of past consumption for the same weekday.

Home Screen

The Home Screen is split in half. In the upper part, an icon is shown as a quick and easy-to-understand indicator for the user - a call to action. The icon shows one of three states:

• the lightbulb is turned on, which means that the club produces more energy than expected and the user is welcomed to increase his or her energy consumption (invest more energy)
• the lightbulb is turned of, which means that the club consumes more energy than expected and the user is welcomed to decrease his or her energy consumption (save some energy for later)
• the lightbulb is cut in half showing a balance of a turned on and turned off light, which means that the club is on track and the user does not need to change anything.

A click on the icon will show the user a forecast of the lightbulb so he or she can plan their energy consumption ahead.

Forecast Screen

As an end consumer, I want to know when is the best time to use electricity in the coming hours and days, so that I can plan energy-intensive tasks accordingly (e.g. doing the laundry, charging my electric car,...)

Implementation: The Lightbulb Forecast Screen shows a series of lightbulbs from top to bottom. Each lightbulb belongs to a specific time period (e.g. 1 hour, 1 pm - 2pm). Similar to the lightbulb on the Home Screen, the lightbulbs indicate whether the specific period is more likely to be a period to spend or save energy.

Backlog (not implemented yet)

Notification / Reminder:

User Story: As an end consumer, I want to get notified of the best time to use electricity, so I don't have to check the app actively.

Implementation: On the "Forecast" screen, a click on one of the lightbulbs (= time period) gives the user the option to set a reminder (e.g. to vacuum clean).

Drill-Down

As an end consumer, I want to see the forecasted consumption and production for a particular time period (e.g. today 3-4pm), so that I understand why I get a recommendation to use or save electricity at that time.

My Realm Screen (not yet implemented, coming soon)

The My Realm Screen shows the energy consumption of the user's home or flat using a simple graph. The data is limited to the last 24 hours. If possible, different energy consumers will be shown separately as well as aggregated. As an example, the bas load (typical household appliances), the usage of the battery, the pv production as well as the consumption of the heatpump could be shown. The user can switch between the aggregated view, showing only "what flows into / out of the house" or the detailed view.

My Club Screen (not yet implemented, coming soon)

Club is the term for a group of customers connected to the same network node Similar to the My Realm Screen, the My Club Screen shows a simplified overview of the electricity network area the user is connected to (automatically determined, login required). On the one hand, the graph shows the performance of the network within the last 24 hours. On the other hand, the graph shows the number of participants or rather their aggregated energy consumption and production in order to highlight the ratio and improve that.

My History Screen (not yet implemented, coming soon)

The My History Screen provides a filter mechanism to specify a period of interest. As an example, the user can select from last day, last week, specific week, last month, specific month, last season, specific season, last year and custom from/to dates. The specified period is then being displayed as a graph whereas the scaling is determined by the period itself. Additionally, the user can add a second period of the same kind in order to start comparing time series. As an example, the user could compare the last week with the same week a year ago.

Compare Appliances Screen (not yet implemented, coming soon)

The Compare Appliances Screen is meant to support the user in acquisition decisions. It focuses on the operational costs based on electricity consumption of the specified appliances. In the first version, the user can input 5 values into the form: - electricity consumption when turned on in Watt - hours of operation per year in hours - electricity consumption when in stand by in Watt - hours in stand by per year in hours - electricity cost (default value provided by operator)

After putting the values in, the estimated electricity costs of that appliance is being shown. The user can add multiple appliances in order to compare them.

Additionally, a list of typical values is being shown, as for example the number of hours a tv is being operated in general (as a rule of thumb).

What can I do Screen (not yet implemented, coming soon)

In the first versions of the What can I do Screen, a list of typical electrical appliances will be shown. The list should help the user to decide which devices can be turned on or turned off, depending on the current situation. In later versions, these options can be improved by - showing only appliances the user actually has - highlighting appliances that can actually be turned on or turned off (e.g. not showing a boiler that is already beeing running) - sorting the list in a smart way (taking the power consumption, ramp up time, saving potential, etc. into account)

Supportive Messages (not yet implemented, coming soon)

The system motivates the user by displaying a supportive message every now and then. A message can be displayed when the user simply opens the app/website (e.g. 'It's nice to see you here', 'Welcome back!'). More importantly, a message should be shown if the user has seen a call to action (e.g. lightbulb is turned on) and the system notices that the energy consumption of the user actually dropped (e.g. 'Thank you for your support, you helped the Club!').

What is watt?

The goal of this project is to make the energy usage of different activities comparable. It is composed of two parts:

1. Collecting data and creating parametrizable models
2. Providing a gamified interface

A first Demo of the interface.

Hackathon

What is watt was evolved during the Energy Data Hackdays 2019 in the Hightech Zentrum in Brugg. Many thanks to: Anna, Germaine, Katharina, Madlaina, Sinan and Stephen.

Some collected results:

• First Django prototype on Github
• First Django prototype on Heroku
• Collected Data Google Drive
• Frontend Github

Pitch

Ziel von What is Watt ist es, alles mit allem zu vergleichen. Wie viel Energie verbraucht meine tägliche Pendler-Zugfahrt? (12.8 kWh) Und wie oft kann ich mit dieser Energie heiss duschen? (9 Mal) Oder einen Stock mit dem Lift fahren? (1200 Mal) Es bleibt am Ende die Entscheidung der einzelnen Person, ob sie nun näher zu ihrem Arbeitsort zieht oder weniger lang/heiss duscht, um Energie zu sparen.

Natürlich hängen die obigen Werte von der Pendlerstrecke, der Duschdauer und der Liftart ab. Und auch was verglichen werden soll, ist oft sehr individuell. What is Watt ist deshalb auf allen Ebenen als offene Plattform konzipiert:

1. Offene Daten dienen als Grundlage zur Berechnung der Szenarien.
2. Die Berechnungsmodelle sind offen und dokumentiert. Man kann sie somit nachvollziehen, kritisieren und verbessern.
3. Das Resultat jedes Szenarios ist ein Wert in Kilowattstunden. Neue Szenarien lassen sich so leicht hinzufügen und mit allen bestehenden vergleichen.
4. What is Watt stellt die Szenarien über eine offene API zur Verfügung. Damit können unterschiedliche Interfaces und Visualisierungen die Daten auf ganz verschiedenste Art zugänglich machen.

Da alle vier Ebenen offen sind, ist es jedem möglich zu What is Watt beizutragen. Typischerweise ist es ein neues Szenario, das jemanden persönlich interessiert. Wie viel Energie zum Beispiel eine Bitcoin-Transaktion verbraucht. Oder die eine Modelleisenbahn. Oder die E-Gitarre.

Die Teilnehmer der Energy Data Hackdays können sich auf jeder der vier Ebenen betätigen und einen Beitrag zur Plattform liefern.

Und wer sich nun fragt, wie viel Energie eine Bitcoin-Transaktion verbraucht, hier die Antwort:

A single bitcoin transaction uses the same amount of electricity as 176 hot showers. So enjoy one more month of showering this year instead of shopping in bitcoin ;-) pic.twitter.com/iVKOGbEems

— Jonas Oesch (@jonasoesch) May 21, 2018

Running with Docker

You need to have Docker installed and running.

docker build -t whatiswatt . 

Creates a new docker image with all the necessary requirements

To run the application:

On macOS docker run -p 8888:8888 -v ~/Desktop/What-is-What:/app whatiswatt

On Windows

docker run -p 8888:8888 -v /c/Users/jonas/energyconverter:/app whatiswatt

More information