Visualize industrial information flows with Sankey diagrams
A brief start about Sankey
Diagrams come into play and provide a global view of the information we focus on when we work with data related to industrial systems. Without them, identifying problems may become pretty hard. Sankey diagrams in the industrial sector are used to identify inefficiencies, visualizing energy balance and material flows. They draw attention to the largest flows of the diagram. Today, we are going to explore this helpful diagram by rolling back in its history. Then we will see when and how to apply it to various situations
A rollback in history
As most of the diagrams or mathematical formulas change names; graphical heat balances (Minister of Fuel and Power 1944), energy flow charts (Christensen 1990) [1] were what is called Sankey diagrams today according to the article “The Sankey Diagram in Energy and Material Flow Management.”.
As stated in “The Sankey Diagram in Energy and Material Flow Management.”, the diagram was published in 1898 [2] by the Irish naval engineer Riall Sankey and named after him.
One of the most famous Sankey diagrams existed before Riall Sankey, created in 1869 by Charles Joseph Minard, a French civil engineer who covered the map of Napoleon’s Russian Campaign of 1812 cited in “Mapping Time : Illustrated by Minard’s Map of Napoleon’s Russian Campaign of 1812.” [3] to visualize troop losses throughout the Campaign.
Today, it is used in material and energy management systems that we will drill down in some of its use cases.
Applying Sankey diagrams
Before seeing our use cases, we need to know how to prepare these diagrams. One of the crucial steps is to define our industrial context in data visualization. A well-designed diagram points us immediately to the message we wish to share.
“Visualization gives you answers to questions you didn’t know you had.” — Ben Schneiderman
Identifying the main flow
Let’s understand step by step how Sankey diagrams work.
Let us take our example illustrated in Figure 3. “Distribution of heat and power in a factory site”. First, as we have discussed earlier, this diagram represents flow “circulation”. A flow is actually pointing out a value from a node directing it to another one. A process is a flow between two nodes. A node is actually a type of context. In our example, we can check out that there is a directed flow from Natural gas to CHP unit 1. Moreover, our flow represents a quantity or KPI ( Key performance indicator ). In our case, it is the value of energy produced.
Looking at the red rectangle, we can deduce that “ 4.3 kWh of energy was transferred from Natural gas to a CHP unit 1 converter”.
A must-know in Sankey diagrams is that the width of our flows is proportional to the quantity of the value it transfers. This feature marks out the most important flows.
Secondly, the set of bars ( for instance, Power supply and Natural gas) form up a Node sets called “ sources”. This is equivalent to column names "sources" in a spreadsheet or database table.
Coming back to our example, we can display our value in a process using dashboards as shown in the red rectangle. We can also display the value that contains a Node element before flow arrival or departure.
In this example, we immediately notice that our main energy generator is Natural Gas. We can also observe that our converters produce mostly heat. Lastly, we realize that our system loses energy when converting energy into heat or power supply.
The key to interpret Sankey diagrams is to observe the widths of our flows between Node sets.
In our second example above, our system generates 5280kW of compressed air per day. Then the energy is distributed throughout our compressed air system. At the first glance at it, I notice that most of the produced energy is lost to heat waste and other losses. This can be relevant to analyze the efficiency of my compressed air System.
Without getting lost
As we might guess, every diagram has its own downside. When those downsides occur, our diagrams lose efficiency and clarity.
In Sankey Diagrams, the number of nodes decreases our readability. In other words, the longer our process chain is, the more we lose sight of the information we wish to communicate. Moreover, they can make it becomes difficult to compare flows with similar values (widths). We should visualize a main flow distinguished from the others when we use Sankey diagrams. In Figure 5. “Line production of cosmetic products”, notice how hard it is for us to identify the line which produces the most cosmetic products.
Besides, we should consider using a (stacked) bar graph to compare our production lines.
Coming to an end
We can say that Sankey diagrams are very efficient diagrams used in visualizing data. They draw attention to the largest flows of the system. And they also provide us information about the structure of our system.
References
[1] Schmidt, Mario. “The Sankey Diagram in Energy and Material Flow Management.” Journal of Industrial Ecology, vol. 12, no. 1, Feb. 2008, pp. 82–94, 10.1111/j.1530–9290.2008.00004.x.
[2] Schmidt, Mario. “The Sankey Diagram in Energy and Material Flow Management.” Journal of Industrial Ecology, vol. 12, no. 1, Feb. 2008, pp. 82–94, 10.1111/j.1530–9290.2008.00004.x.
[3] Menno-Jan Kraak. “Mapping Time : Illustrated by Minard’s Map of Napoleon’s Russian Campaign of 1812.” CiNii Books, ESRI Press, 2014, ci.nii.ac.jp/ncid/BB19280683.
Comments