A few months ago I knew nothing about energy distribution. I could only say that there is a power plant, there is an electricity line, a transformer and finally the energy meter and my end devices. I also wasn’t interested in this subject until there was a problem to solve.
I love my job mainly because each day is different and presents new challenges for me. This time I had a pleasure to participate in a meeting where Objectivity accepted the challenge of supporting Microsoft in providing architecture and solution for energy distributors to effectively manage power limitation exceedances.
From zero to hero in 3 weeks
At that time I had no clue what it would be all about. As always, the key thing is to understand the problem in order to provide a solution! I was really worried that we would not be able to fulfil this gap so quickly. Jaroslaw Zarychta (Business Development Executive at Microsoft) had already a vast understanding of the business problem and after several hours we were in a position to accept the challenge. We knew what should be done and we suspected how it could be done however, we encountered a problem with a significant part of domain knowledge that was missing. What is the data involved and algorithms used in calculations? To solve it, Microsoft has invited another partner that could provide expertise in this domain. So, we formed a team and started immediately as the deadline was in 3 weeks… I was really excited about it as we were talking about data from hundreds of energy meters collected each 15 minutes, hence we moved to BigData aspect! At the end of the day it turned out that to solve the problem we do not have to incorporate the whole BigData stack from Azure, instead SQL Server Reporting Services 2016 would be enough. But to the point…
Objectivity’s experts have aggregated data and visualized as required. It is possible to easily identify the moment in time where consumers exceeded power consumption limits (once restrictions applied).
Energy distributors have to report the exceedances of power consumption in specific cases. This is to ensure the country’s energetic safety. To put it more simply – to secure the energy delivery to priority consumers such as hospitals, core industry factories, airports, military facilities, etc. Believe it or not we walk on the edge. Consumption of energy is getting bigger and there are moments in time when it is not possible to distribute the energy to end users to satisfy their demand. Due to the weather conditions (heat) between 11th and 31st August 2015 restrictions were set for entities consuming more than 300kW of power so the reason to report exceedances is extremely important. Upon this report the consumers who do not meet the limits will pay some financial penalty.
To make it even better Objectivity has used PowerBI with SSRS2016 to provide a better data visualisation. Aggregating the exceedances and presenting them on the map gives a valuable insight into the subject. Instead of table, graphical representation including geolocation with a possible drill down to the value of exceedance it gives much better control and visibility to the subject of exceedances awareness and control.
But interestingly not really… I have questioned myself: “Does it really solve the problem?”.
Well in my opinion not. Knowing that the exceedances occurred is one thing but making sure the consumer pays for it quite another. But none of them prevents it happening, also none of them allows us to react immediately to stop it occurring! – hence, it doesn’t answer the real problem: to ensure the country’s energetic safety!!
All together we started to imagine how the system could look to solve the issue. Assuming that:
- we have the metering data from big consumers (<300kW) available each 15 minutes,
- we know their business safety level, so minimum energy that has to be delivered to not introduce a serious injury,
- we know their contractual power,
- there are appointed representatives on consumer side that we could communicate with,
We could collect data about consumption for some time and build a profile of consumer. This would create a power consumption model per consumer. Relying on it, we could propose a plan of power reduction specific for that consumer, adjusted to their power reduction capabilities.
Step one – to make planned reduction achievable – done.
Next thing is awareness. Firstly, I really do not expect all serious companies to listen to the radio each day, waiting for announcements about introducing the power consumption limits! We live in 21st century, it must be organized better. What I would expect is the message, sent directly to appointed contact/representative, informing that the limitations apply and there is a consumption limit set to specific level in accordance to defined plan (which by the way is achievable).
Step two – make people aware – done.
Even now it is not perfect. We assume that people react accordingly but we do not have a mechanism to check if they did decrease the consumption to meet the planned levels. Let’s introduce monitoring with IoT. It doesn’t need to be enabled all the time – only when limitations apply. When we collect the data each 15 minutes we do have a chance to recognize the fact that the exceedance of the limit takes place. Moreover, I strongly believe that no one likes to put the country’s energetic safety at risk, therefore, when we notify the consumer that the limits are exceeded, they will, in majority, act immediately. To make it work they need to know that it is happening now and the reason why they should shut down part of their machine park – to ensure energetic safety of the country.
Step three – monitor and inform when exceedances occur – done.
There is a huge potential for BigData and IoT in the area of energy and its distribution and I would love to explore that more. The proof of concept we did with Microsoft is just a small step on the way to a better management of our main resources – which in this case is energy. As in most cases applying basic pattern makes sense: collect data – learn (machine learning)/predict, monitor and react. Without data we are blind.
If you would like to read more about our experience from this PoC, read the following Case Study. I hope there will be enough energy for you to do that.