Sankey Diagrams

The Council on Cempetitieveness has just called on the presidential candidates to come up with a national energy plan, believing that future economic growth and security of the United States depends on energy efficiency across the U.S. economy, sustainable energy solutions, and development of new technologies.

The new president might want to ask John Ziagos to advise him on energy issues. Ziagos is with the Energy & Environment Directorate at Lawrence Livermore National Laboratories, and an expert for energy scenarios. For me as a Sankey aficionado he is simply the “man behind the Sankey diagrams at LLNL”.

LLNL has been publishing energy flow diagrams for the U.S. over the last few years. In his public presentations (link1 – 2.9 MB, link2 – 4.8 MB) Ziagos impressively shows that even when implementing fuel cell technology for all vehicles, switching completely from coal and natural gas to renewable energy sources to generate the nation’s electricity, and building 270 new nuclear power plants … even then it would not be possible to stabilize U.S. carbon emissions between now and 2050. “If we want to move toward a carbonless future, no single technology will do – everything counts.” Ziagos says.

I am reprodrucing two energy flow diagrams for the U.S. from John’s presentations below. The first is for 1976, the second a 2025 projection.

While overall energy produced will rise from 72 Quads in 1976 to an estimated 133 Quads p.a. in 2025 [Note: 1 Quad(rillion) equal to 10E15 BTU, or 1.055 × 10E18 joules (1.055 exajoules or EJ) in SI units.], the ratio of useful energy to rejected energy gets worse.

Statistics Canada in its “Report on Energy Supply-demand in Canada” for 2005 shows two Sankey diagram in the annex (HTML version of the two Sankey diagrams). They show the energy flows for Canada 2005 and 2004 in Petajoules per year.

I have featured similarly structured diagram for other countries here before: Japan, Scotland, Ireland, and the United States.

Out of 21380 PJ of total energy produced in Canada and imported, some 9641 PJ (45%) are exported, while 11739 PJ (55%) are national consumption. If you have the impression that the proportions are not 45:55, you are right, they are more like 39:61! From a graphical perspective this Sankey has more peculiarities worth a mention: the magnitude of the Sankey arrow changes and just before the arrow head they become narrower. The flows labeled “Steam” and “Adjustments” seem to have been added at a later stage as they don’t merge into the other arrow. Steam is represented on the production side as well as in the breakdown of energy carriers with a small, but not unsignificant width, however the quanity is given as zero.

This diagram of sun radiation being absorbed and reflected when hitting earth (from Solar Energy Facts website) is a rather weak remake of the original Nasa diagram.

I find the floating powerpointish arrows kind of disturbing, and with the arrow magnitudes not to scale, would even call it misleading. Took the time to prepare two new versions of it (actually I am beta testing the new version 2.0 of e!Sankey at the moment – so this was a nice little test case).

The first version sticks more to the original idea of the diagram shown above, but the arrow magnitudes are corrected and to scale. The second version is closer to the original ‘Breakdown of the incoming Solar Energy’ diagram by User A1 that can be found on Wikicommons. The latter one has the flow for energy being absorbed by atmosphere (33 PW) branching off as the first arrow horizontally.

ENOVA is a public enterprise owned by the Royal Norwegian Ministry of Petroleum and Energy. It advises the ministry in questions relating to energy efficiency and new renewable energy.

One of their services is the practical development of energy and climate plans (Kommunal energi- og klimaplanlegging) for Norwegian municipalities. On their website they present a sample Sankey diagram to visualize energy flows.

The Sankey diagram shows the energy flows in the Stor-Elvdal municipality in GWh (probably per year, a year is not given). This is very interesting, as the municipality can cover almost all of its energy for industry and private households from renewable energy sources, such as biomass and wind. Energy from fossil sources is consumed through almost entirely transportation. Stor-Elvdal produces 47.5 GWh within the municipality, and imports another 35,8 GWh from wind from outside the municipality.

A small typo can be found in label on the the orange flow (saying 85,8 where it should read 95,8), but this doesn’t spoil the overall quality of this diagram painted with SDraw, I guess.

Download the publication (in Norwegian) as PDF file

Another runner-up in my private “Fancy Sankey Diagram” contest definitely is this Sankey diagram shown on a webpage of the Longford Environmental Alliance (LEA) from Ireland.

It visualizes the “Energy Balance for 2005 as a flow diagram showing our inputs from the left hand side and our outputs or usages on the right hand side.” It is a 3-D image, and kind of floats above the ground, although it doesn’t have a fancy shadow effect as this one does.

I have shown similar diagrams for California, Japan, Sri Lanka, Scotland and the U.S. before. In these national energy balances the various energy sources are shown as entries from the left, while consuming sectors (or the “sinks”) are displayed as output arrows. This Irish Sankey diagram distinguishes ‘Agriculture’ as a separate sector.

Well done Éire, home of late Mr. Sankey…

The World Business Council on Sustainable Development (WBCSD) in an article on “Making Tomorrow’s Building’s More Energy Efficient” features a great three-dimensional Sankey Diagram, to illustrate that “more than 90% of the energy extracted from the ground is wasted before it becomes useful work”. The article calls for green buildings where energy is produced onsite, and losses are minimized.

The Sankey arrows representing the losses bend down sharply, they remind me of the Iguazu Falls. Neat 3D images of the equipment are placed on the diagram to visualize the process steps where energy is lost. The whole thing hovers over the ground throwing a faint shade. The graphic designer who did this really merits an applause.

If ever I launch a ‘Best Sankey Diagram Award”, this one will have good chances to win it. Any sponsors out there? Any volunteers for the award jury?

The Mexican National Commission on Energy Saving (Comisión Nacional para el Ahorro de Energía (CONAE) present several success stories (casos exitosos) on their website.

One success story dates back to 1997, and describes how an energy efficiency study of fired heaters (i.e. boilers) was carried out in a Nafta producing facility in the Veracruz state of Mexico. As a result of the study, several suggestions for optimization were implemented. Fuel consumption could be reduced by 23-24 %, while the efficiency of the ovens could be raised by 13% (calentador BA-2001 B) and 16% respectively (calentador BA-2001 A).

Para los hispanoparlantes: el título oficial del proyecto fue “estudio técnico económico e ingeniería conceptual realizada a los calentadores a fuego directo BA-2001 A/B de la planta hidrodesulfuradora de naftas, del C.P.Q. “La Cangrejera”, ubicado en Coatzacoalcos, Veracruz” (otro candidato para el concurso Mundial de titulos largos).

The heat losses are shown as Sankey diagrams. The first describes the optimal situation, with an energy efficiency of 82,4 % “as guaranteed” by the maker of the fired heater.

The two other Sankey diagrams show the energy balance of the heaters A and B before the implementation of the measures. They run with an efficiency of 60,6 % and 62,35 %.

The arrows branching off at the top show the heat losses. I like the fancy icons that show how energy is lost through the walls, because of deteriorated or insufficient insulation, and heat energy in the effluent gases. The flows are given in MMBTU/h (millions of BTU per hour).

Unfortunately two of the diagrams are not to scale: The arrow to the right in the second diagram should be roughly 2/3 of the width on the left side. It is about 4/5 (or 80 %) of the width, similar to the width in the first Sankey diagram. This is a visual exaggeration of the inefficiency. However, I refrained from featuring this in my informal “Lying with Sankey diagrams” series.

California Energy Commission in 2005 published the final report of a project called CALEB (California Energy Balance). The CALEB database is compiled by the Lawrence Berkeley National Laboratory (Berkeley Lab) and contains “data for energy production, transformation, and consumption for the State of California for the period 1990 to present.”

The report shows the California energy balance for the year 2000 as a Sankey diagram. (Source: Murtishaw S, Price L, de la Rue du Can S, Masanet E, Worrell E, and Sathaye J, 2005. Development of Energy Balances for the State of California. Sacramento, CA: California Energy Commission CEC-500-2005-068 (LBNL-54923))

A PDF with this diagram explains how to read it:

Reading from left to right, the figure shows all the inputs of primary (and imported secondary) energy into California’s economy in 2000. These are summed by major fuel types in the middle of the figure: petroleum and associated products, natural gas, and inputs to electricity generation. The right-hand side shows how all of the fuels are allocated to the various end uses.

The executive summary explains that the diagram shows some peculiarities for ‘Arnieland’ compared to other states. It also points out that further research is needed to collect more data and to be able to break down aggregated energy data.

This way of presenting the energy balance seems to develop as a kind of standard for representing energy flows for a state or country. Other examples can be seen here: Japan, Scotland, U.S., or Sweden.

The diagram has two special features which distinguish it from other Sankey diagrams:

Firstly, it shows negative flows as dotted Sankey arrows, indicating “instances where consumption is greater than supply” – something which on a physical level is of course impossible, but is explained with statistical differences and “unexplained transformation gain … in petroleum refineries” (NB: I would prefer to read “data inconsistency” here).

Secondly the breakdown of primary energy and imported secondary energy (input from the left) is fanned out leaving gaps between the arrow lines. The merged arrows are then “condensed” to be to scale again. This creates a 45° gradient at one side of an arrow section (or “slope” for those of you who are into winter sports). In these sections the arrow width can not be proportional to the quantities. A tribute to pay for better legibility of the diagram.