General Edition
1ˢᵗ Place | Team Ctrl.Alt.Elite - Four computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Title: Ctrl.Alt.Elite Team - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:26 minutes
Description: Four computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Ctrl.Alt.Elite Team - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
CTRL.ALT.ELITE TEAM
Yellow wipe transition reveals a frame with a self-shot video of one of the team’s members.
[Team member]
Our name is Ctrl.Alt.Elite and Atishya, Jayant and Mayank are final year students at computer science at IIT Delhi. Rajas is an alumni at IID and will soon pursue his masters at Stanford.
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with the team member on the right of the screen.
[Team member]
As you can see from this image our team had a perfect balance of complementary skills which actually helped gain us insights from different perspectives.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of a 2nd team member.
[2nd Team member]
Let’s begin with the intuitions that we gained over different stages of the competition.
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with the 2nd team member on the right of the screen.
[2nd Team member]
Regarding the wind distribution, a simple average over the training set suffices for good generalisation as variation is minimal. To build intuition, we built a cool GUI to visualise the deficit that one turbine creates on others. Now we will talk about our algorithms. We compared different evolutionary algorithms like PSO and CMA-ES, along with local search techniques. As you can see in the diagram on the right, evolutionary models grow faster but saturate earlier than local search. But none of these methods went beyond a score of 528. Finally let’s discuss the tricks that helped us escape these local optima and grab the first position on the public leaderboard. We allowed our local search model a negative step which helps it evade the local optima and get better to solutions. Now we realised our model moves only one mill at a time so we don’t need to recompute the scores from scratch. Using this insight makes our model 10 times faster. Discontinuity creates sparse rewards making it tough to learn. We eliminate these discontinuities by linearly interpolating the power curve. We constrained the search space by reducing the parameters from 100 to 5, exploiting the symmetry of the problem. This allows us to move multiple turbines using just one variable. Finally different models optimised different aspects of the problem. When one gets stuck the other helps out. Our parallel architecture combines all these models, giving the best solution.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with the 3rd team member on the right of the screen.
[3rd Team member]
All of us come from a computer science background and have actively contributed to AI Research. We hope to bring our experience in tackling novel research problems to drive the much needed energy transition.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
2ⁿᵈ Place | Team DeltaPhiAlpha - Thibaut and his cat Bounty, believe his skills in AI can improve wind farm performance.
Title: Team DeltaPhiAlpha - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 1:56 minutes
Description: Thibaut and his cat Bounty, believe his skills in AI can improve wind farm performance.
Team DeltaPhiAlpha - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM DELTAPHIALPHA
Yellow wipe transition reveals a frame with a self-shot video of one of the team’s members.
[Team member]
Hi guys, it’s DeltaPhiAlpha. I’m alone in my team but I got some help from my cat, Bounty.
She’s sleeping most of the time, but she’s helpful.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of the team member.
[Team member]
So let’s start, perimeter-centered type of layout, that’s what I found was the best after some testing. We’ll place the remaining turbines inside using a genetic algorithm type of optimisation and this will find the best layouts at the end and we can further improve it using a PSO particle swarm of optimisation on each of the turbines starting from the edge. That will move some of the turbines to reach an even better solution. At the end, we can finally try different wind data inputs and that will give the best solution. For the uncertainty, we can have a train and test approach. For layout trained on a given input, we can test on the other wind inputs. And to avoid any bias, we’ll use the AEP of the layout versus the AEP of a random layout. We’ll pick for private leaderboard, we’ll pick the one with the highest mean, which in this case is the red one.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ in the right upper corner and the frame with the team member centered on the screen.
[Team member]
I believe my skills in AI can improve the wind farm performance in the future. For example, wind turbines being more autonomous, using algorithms and robotics, and last but not least,
[Video footage]
Team member puts on a Star Wars mask and holds a Star Wars lightsaber.
[Team member]
May the wind be with you.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
3ʳᵈ Place | Team IIT Madras - This team of two final year dual-degree students know that every single percentage of improvement counts towards increasing annual energy production.
Title: Team IIT Madras - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:23minutes
Description: This team of two final year dual-degree students know that every single percentage of improvement counts towards increasing annual energy production.
Team IIT Madras - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM IIT MADRAS
Yellow wipe transition reveals a frame with a self-shot video of one of the team’s members.
[Team member]
Hi, this is Kishore
[Visual]
Yellow wipe transition reveals a frame with a self-shot video of a 2nd team member.
[2nd Team member]
And this is Pradeep Gopolalakrishnan. We are final year dual-degree students from IIT Madras.
This is our official submission for the Shell India 2020 Hackathon. And we are team IIT Madras.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with Pradeep on the right of the screen.
[Pradeep]
Hey guys, this is our problem statement. Our task is to place fifty wind turbines on a square plot of land, that will be a wind farm. We have to satisfy proximity and boundary constraints and the task is this, to maximize our annual energy output for the farm. This comes obviously with a lot of challenges. The challenges being it’s a very high dimension search problem. We have to include constraints in our search and the fact that our objective surface is highly non-convex.
As you see in the figure on the right, we have a lot of local minimas and maximas and reaching a global maxima is really difficult. Gradient based approaches usually fail here, so we resort to a non-gradient based method. Our approach goes like this. We have a core search module that uses generic algorithms. This allows us to search on scales greater than a hundred meters. And after that, we have a fine search module that uses a motivation based greedy method that searches on scales less than a hundred meters. So now we’ll see a GA in action. This is how it goes. As the generations progress, we see the particles moving the particles, the turbines moving around, and they all tend to converge around the edge of the plot. And in addition to this, the interiors display formation of clusters, turbine clusters. This ensures that the distances between the turbines are maximal.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Blue animated background with Kishore centered in frame on the screen.
[Kishore]
These days of data and optimisation, where every single percentage of improvement counts.
Over the past month, we increased the annual energy production by twenty Gigawatt hours. That is equivalent to powering twenty thousand Indian homes over a year. This is the way forward and this is how we are going to do it.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
Special University Edition
North Zone | Suraj K Team - Suraj is a chemical engineering student, his research work is focussed on simulating nanoscale systems using molecular dynamics.
Title: Suraj-K-Team - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:41 minutes
Description: Suraj is a chemical engineering student, his research work is focussed on simulating nanoscale systems using molecular dynamics.
Suraj-K-Team - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
SURAJ-K-TEAM
Yellow wipe transition reveals a frame with a self-shot video of one of the team’s members.
[Team member]
Hi everyone! I am Suraj K. I am currently in my second year of MS by research at IIT-Kanpur in chemical engineering and my research work is on simulating nanoscale systems using molecular dynamics.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with Suraj K. on the right of the screen.
[Suraj K.]
So in the optimization pipeline we generate a population of ‘n’ random initial configurations, as shown here. And we use the hybrid genetic algorithm on a grid to generate the optimum configuration. Next we do simulated annealing on this new optimum configuration to get the global optimum configuration.
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with Suraj K. centered on screen.
[Suraj K.]
And after the grade optimisation, we use the simulated annealing to get he global optimum configuration.
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with Suraj K. on the right of the screen.
[Suraj K.]
And in the hybrid genetic algorithm on grid it is a iterative optimisation procedure based on the natural selection and it selects the best performing ones in each generation and crossovers them and adds mutation to them to explore the configuration space and in the hybrid part is due to the fact that there is a local search in effect to the, in addition to the global search going on and here we can see that in around one hundred steps, we’ve reached the optimum value and in the simulated annealing optimisation we used the previous grid optimum to calculate the global optimum and it is based on a virtual temperature which we decreased like in the simulated annealing of metals. And we perturbed the positions used and we accept the new position based on the metropolis criterion where the p-acceptance is given here.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Blue animated background with a corresponding slide on the left and the text ‘Shell ai HACKATHON’ and the frame with Suraj K. centered on screen.
[Suraj K.]
As renewable resources are becoming more common and are replacing the conventional energy sources, I think we can focus more on the technologies that are available, like AI and optimisation techniques, to better get the yield out of our available natural resources.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
South Zone | Team 7 - This team of four computer science and mechanical engineers are fans of using neural networks for predicting wind speed.
Title: Team-7 - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:27 minutes
Description: Four computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Team-7 - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM-7
Yellow wipe transition reveals the text ‘Shell ai HACKATHON’ and a frame with one of the team members on the right of the screen accompanied by a corresponding slide on the left.
[1st Team member]
We are Team-7 and I am Awanit Ranjan.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a frame with one of the team members on the right of the screen accompanied by a corresponding slide on the left.
[2nd Team member]
Hi I’m Abhishek Choudhary.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a frame with one of the team members on the right of the screen accompanied by a corresponding slide on the left.
[3rd Team member]
Hi I’m Akash Waitage.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a frame with one of the team members on the right of the screen accompanied by a corresponding slide on the left.
[4th Team member]
Hi and I’m Vivek.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Akash. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Akash]
We want to place a turbine in such a way that it produces a maximum power. So why don’t we place it anywhere because there is a problem. Turbine produces a wake effect which reduces the power of a turbine in a downward direction. That’s why we need an optimal position of turbine, for that we have a solution. Greedy, greedy is the best. You ask, “why?”. Because it takes less computation time. Also we have a better optimization method that is repetitive arrangement with greedy.
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Abhishek. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Abishek]
So the idea is, first divide the windfarm into grids and consider only the grid points for evaluation. The more dense grid is used the more close we will be to the optimum result after the completion of stage one but the computational time will also increase. In stage one we can start by selecting a random stationary point and the next point will be added one by one according to a greedy algorithm to yield maximum annual energy production and satisfy perimeter and proximity constraints. After getting all the fifty points we will be having a partially optimised wind farm layout.
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Vivek. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Vivek]
So now in stage 2, we start by removing one turbine and search again in the surrounding area around that turbine which will be divided into an even more dense grid to get a new location of the turbine using greedy method. Similarly other turbines will adjust their location. After multiple iterations we can see how our result got optimised from stage one to stage two.
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Awanit. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Awanit]
Currently our approach works for a flat terrain but in real life it is scalable for a complex terrain too and adaptable to the variable turbine heights even if integrating the cost factor is involved.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Blue animated background with Awanit centered on screen and the text “Shell.ia Hackathon” in the upper right corner.
[Awanit]
We can use NN for predicting the advance wind speed of our wind power output which can help several organisations for scheduling the power generation and can also serve as a prerequisite to execute sustainable integration of a wind power grid on a particular location on earth. Thank you.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
East Zone | Team DeepWhiz - A team of three AI lovers, united by their vision to give back to the community.
Title: Team deepwizai.com - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:10 minutes
Description: Three computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Team-7 - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM DEEPWIZAI.COM
Yellow wipe transition reveals the text ‘Shell ai HACKATHON’ and a centered frame with one of the team members' self-shot video.
[1st Team member]
Hi, this is Atif. I’m a PhD student from IIT Kharagpur and I ‘m the team lead of deepwizai.com.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a centered frame with one of the team members' self-shot video.
[2nd Team member]
Hi, this is Sayantan and I’m an MTech student from IIT Guwahati and currently I am working as a software developer in Cisco.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a centered frame with one of the team members' self-shot video.
[3rd Team member]
Hi, I am Gourab Chowdburry. I am a student at IIT Kharagpur.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a centered frame with Gourabs' self-shot video.
[Gourab]
Hi everyone, our objective is wind farm layout optimisation.
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Gourab. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Gourab]
So AI in industry, people are very skeptical about AI’s since the decisions are made without a human in the loop and not always explainable. So we proposed NDOpt, which is a novel algorithm, which takes humans’ decision into account and improves on top of that. So this is an algorithm which is a non-differentiable optimiser. That takes initial position, changes the position of each turnpoint, checks for the constraints and improves them if they are viable. They take the learning rate and the tolerance rate, then choose the layout that gives the highest score. So the layout is designed in such a way that the point, each point is a wind turbine that can be moved in any direction and the algorithm will move the turbine in the direction that gives the highest AEP score. This is a small visualisation tool we designed to help us give the initial seed. So we can play around with the points, which are the wind turbines and this gives us AEP score. So then we move to the ensemble meter, which takes two layout and then combines them to get the highest AEP score.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Gourab. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Gourab]
The energy transition drive is hindered by huge operational costs and reduced output. So we can use AI for early prediction of planned failures, for example, geothermal plants, and we can use AI to maximize the output in solar layout optimisation.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
West Zone | Team ChaiAtChea - Are two Chemical Engineering PhD students that want to use their skills to positively impact the energy transition… and they love chai!
Title: Team ChaiAtChea - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:20 minutes
Description: Two computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Team-7 - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM CHAIATCHEA
Yellow wipe transition reveals the blue animated background and a centered frame with a self-shot video of one of the team members.
[1st Team member]
Hello everyone. We are two PhD students from the department of chemical engineering at the Indian Institute of Technology Bombay.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a centered frame with a self-shot video of one of the team members.
[2nd Team member]
I am Om Prakash, we are team ChaiAtChea because we miss our chai-vala chea canteen at IIT campus and we love chai.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of the 1st team member. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[1st Team member]
We started out with some standard conventional and evolutionary methods. However they did not perform very well because of slow convergence, very high computational time. We tried a few other metaheuristics, one of which was the Yin-Yang_Pair Optimisation, which was a methoddI had developed during my Masters. This proved to work and generated good solutions in the first half of the competition. We used the 7 year wind data given to us to construct three new wind sets: high-wind, low-wind and mean wind. While all three were used by us at different stages of the competition, the high-wind set was observed to give us good improvements in regard to the public leaderboard and hence was used the most by us.
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Om. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Om]
We were parallelly looking at the grid based hybrid greedy search method. But before that, we pre-computed each fictitious turbine’s wake deficit at all points. Then we used this information for all our subsequent function evaluations. This trick reduced our computational burden by 99%. The first stage of the iteration of our approach is forward greedy step where a turbine is greedily added. Second, is the adjust stage, where added turbines are locally adjusted. Third, is backward elimination, where a bad turbine contributing little to the objective is removed. After we obtain all the required turbines, we fine tune each turbine in its vicinity in order of their maximum contribution to energy production. Finally, we extend this idea to a moving horizon framework, where we used the hybrid greedy approach to select turbines in each rectangular horizon.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Yellow wipe transition reveals the blue animated background and a centered frame with a self-shot video of the 1st team member.
[1st Team member]
Both of us have previously worked on different topics in this field, such as on biofuels and microgrid optimisation.
[Video footage]
Yellow wipe transition reveals the blue animated background and a centered frame with a self-shot video of Om.
[Om]
Analysing sustainability, ensuring safety operation and performing techno-economic analysis;
These are few of the many ways by which we can positively impact the energy transition.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
Public Vote
Team Energizer.de.ICT - A team of two friends that believe building an energy secure society is not just programming language or algorithms but also an understanding of energy systems.
Title: Team Energizers.de.ICT - Shell ai. Hackathon for Sustainable and Affordable Energy
Duration: 2:24 minutes
Description: Two computer science students that hope to use their experience in tackling novel research problems to drive the energy transition.
Team-7 - Transcript
[Background music plays]
Rhythmic instrumental music
[Visual]
Blue background with moving connection constellations.
Text appears:
Shell. Ai
HACKATHON
for Sustainable and Affordable Energy
In smaller font at bottom:
The information and opinions in this presentation are those of the author
and not endorsed by Shell International Limited
Yellow wipe transition reveals new text:
TEAM ENERGIZERS.DE.ICT
Yellow wipe transition reveals the blue animated background and a centered frame with a self-shot video of one of the team members.
[1st Team member]
Hello everyone this is Abhinandan Mohanty.
[Video footage]
Blue animated background with the text ‘Shell ai HACKATHON’ and a centered frame with a self-shot video of one of the team members.
[2nd Team member]
This is Jyotisman Rath and we are from team Energizers.de.ICT.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
TECHNICAL DOWNLOAD
[Video footage]
Yellow wipe transition reveals the blue animated background and a frame with a self-shot video of Jyotisman. Alongside corresponding slides and the text ‘Shell ai HACKATHON’.
[Jyotisman]
Wind Energy is a clean and sustainable source of energy. And by setting up wind farms, we can get energy on a very large scale. But the thing is, how do we exactly install the windturbines?
Maybe in a straight line or in a zig-zag fashion. We need to find out which windfarm layout will give the best energy output? And to answer this, we started with a basic analysis of the problem statement. We observed that by placing the turbines on the borders or some specific positions we got better results. But then, we needed something faster and better. Thus we came up with genetic algorithm. Now, this works exactly the way evolution occurs in nature, the way humans evolved and this nature inspired concept is fed into a programming language for making the best use. Coming to our approach, we started with a random layout as input to the algorithm, and as you see, the algorithm evolved to give the best result. By the end of the Hackathon, we were able to get an energy output of 548.6 Gigawatt hours. But wait! How do we know our solution is the best? For this, we calculated the efficiency and we got 93.7%. That is really interesting in terms of energy. Of course we can get better results and maybe cross the 550 mark, but the layout may come out to be specific to given conditions. Through our approach, we were able to get layouts that can perform good for different types of turbine or wind farm designs and also for different kinds of wind flow throughout the year. After all, we are looking for an optimum layout which is easily implementable and sustainable as well.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
HOW CAN YOUR SKILLS HELP WITH THE ENERGY TRANSITION?
[Video footage]
Yellow wipe transition reveals the blue animated background and a centered frame with a self-shot video of Jyotisman.
[Jyotisman]
For building a energy secure society, it’s not just programming language or algorithms but an understanding of energy systems that’s very much required as well. Since we have an interest in energy research, we feel we can help in the energy transition, that’s the goal for our society.
[Visual]
Yellow swipe revealing the blue animated background.
Text appears:
Vote for your favourite team at
shell.in/hackathon-vote
Help them win the public vote!
[Background music]
Music ends
[Visual]
White background with text revealed:
Powered by
Shell
.ai
Followed by a white screen showing the Shell pecten.
On the bottom of the screen it says:
#ShellHackathon
© Shell International Limited 2020
