SunCode 2019 is coming up! Check out the details here.

Want to build a clean energy startup in less than 24 hours with a chance at $10,000 in cash prizes? Powerhouse, a co-working space and venture fund in Oakland, is hosting their annual hackathon on April 13-14. Over 150 developers, designers, and clean energy professionals come together to build solutions for the clean energy industry. Startups that have resulted from SunCode include: Powerhive, SunSwarm, UtilityAPI, Nanogrid, and OhmConnect. You don't have to know anything about clean energy to participate. 

Schedule  

Friday Evening April 13 

6:00-7:00p - Happy Hour and Dinner 
7:00-8:30p - Form Teams
9:00p - Doors Close 

Saturday April 14 

8:00a - Doors Open and Breakfast 
8:00a-12:00p - Hacking
12:00-1:00p - Lunch
1:00-6:00p - Hacking - To pitch you must upload to DevPost by 6pm - (Takes 15+ mins) 
6:00-6:30p - Dinner 
6:30-8:45p - Pitches to Judges - 3 mins per team and 30 sec Q+A
8:45-9:15p - Happy Hour (Judges Deliberate)
9:15-9:30p - Winners Announced

Challenges

Schneider Electric

Challenge: Build an online marketplace enabling end to end offering for residential and commercial segments including:

• Customer acquisition
• Lead generation for: 
  • sizing and development of hardware supported systems such as energy storage
  • software features or services such as - Energy cost savings, availability of reliable & resilient power, project financing, vendor(s) selection, installation and commissioning, monitoring, maintenance and servicing and covering all stakeholders such as consumers (individual or as communities), installers.

Provide a business model that will help in scaling the marketplace for mature economies like US, Germany and Australia,  yet catering to the local needs of these geographies.

Tesla Challenge

1) Problem Statement: What type of home is most suitable for solar, and what attributes of the home has high value in predicting whether the homeowner will go solar?

Context: If we determine that a home with certain attributes are more likely to go solar, we can create different segments of homeowners. Then, we can utilize these segments to test different approaches such as having a different method for engineers to design a system, conducting site survey, presenting different options in the online configurator, etc. The solution is open-ended, and one can look at many different features of the home and its predictability. Here are some examples:

• Based on address or lat/long, what direction does the front of the house face and how does that correlate with solar? For example, does a north/east facing home have a higher likelihood of going solar?
• What roof size is most suitable for solar?
• What roof size with with X solar potential are more likely to go solar?
• Is there a certain geographical region where homes are more likely to go solar?
• Can we build a machine learning model with certain home attributes as a feature set to predict whether a homeowner will go solar?
• Can we rank homes based on solar potential and certain home attributes?

Datasets:
Google Sunroof Data (Lat/Long, Roof Area, Solar Potential by Deciles)
Sample of Tesla PV Installed Customers (Lat/Long, Mounting Planes, Azimuth)
Sample of Tesla PV Customers who cancelled or were deactivated (Lat/Long, Mounting Planes, Azimuth)
Zillow Data (Home Square Footage, Year Home Built, # of Bedrooms, etc.)
NREL regional hourly consumption data based on weather stations

2) Context: When we intend to install solar panels on a customer’s home, a surveyor goes onsite and measures the length of each roof edge so that a system designer can model the roof accurately and design a layout of solar panels we can be sure will fit on the roof. This is a time-consuming process for the surveyor, and any reduction in time spent on the roof translates to a significant reduction in overall installation costs. Therefore, given:

• Module dimensions (36” x 54”)
• Target module count (between 6 and 40)
• General roof shape with measurements accurate to 4” (see attached examples)
• Rectangular
• Hip
• Rectangular with dormers
• Single hip
• Locations of rooftop obstructions within 24” in any direction

And the following restrictions:

• No modules can be installed within 12” of any roof edge
• No modules can be installed within 6” of any rooftop obstruction
• Modules must be installed in landscape orientation

A surveyor on-site needs to know exactly what measurements to take, so the challenge is to:

1. Design a system that will automatically determine, given the above information, what combination of roof edge length measurements, obstruction locations relative to roof lines, and obstruction locations relative to each other are necessary to be gathered by the on-site surveyor in order to ensure the number of modules specified will fit on the roof.
2. Display this information in an easy-to-read UI such that the surveyor knows what measurements to take.

Alternatively (but related), come up with a set of rules that can be applied to an arbitrary roof plane to determine whether the surveyor needs to take any measurements at all. For example,“If a roof plane is rectangular, has dimensions larger than 16’ x 36’, and has no more than 3 obstructions, an 8-module system will always fit and no rooftop measurements need to be taken”

Datasets

Roof Plane Description and Examples

Vestas Challenge

Challenge: A generic approach to optimal unit dispatch for renewable Hybrid plants in a C&I context.

Objective: Create an continuously optimising algorithm, utilising machine learning for optimal unit dispatch, including various renewable generators (wind + PV) connected to a battery energy system supplying a local electricity load (behind the meter). The local demand must be met at all times at the lowest possible cost on the electricity bill. .

Data available: Wind production (hourly), PV production (hourly), relevant storage data, wholesale electricity prices (hourly), local demand profile, tariffs for grid consumption. This data is provided as a show case, the solution needs to be able to adapt when the input data change within minutes. You can assume that the prices fluctuate by 20% off the given profiles as well as over seasons and years and the wind/PV fluctuate by 10% off from the given profiles.

Challenge: Enable energy prosumers to settle electricity bills through large MW renewable generation assets. 

Objective: Develop a platform for MW scale Hybrid renewable plant operators, that utilises p2p energy trading to certain customer segments, including but not limited to Large industrial clients, metropolitan areas, while also being active on traditional energy markets. Propose an approach to effectively balance production and demand using intelligent control i.e. machine learning (improve wholesale – p2p ratio’s and price forecast) or artificial intelligence (own and operate the settlement of p2p customers.

Data available: Wind production (hourly), PV production (hourly), wholesale electricity prices (hourly), local demand profile

Renault Nissan Challenge

Challenge 1: Assist non-homeowners access the Electric Vehicle market

Help non-homeowners who need to rely on public charging network to assess cost saving benefits of EVs vs combustion engine vehicles.

Develop an application taking into account local charge point coverage, rates, …

Challenge 2: Assist future customers of Electric Vehicles in better understanding the overall environmental benefits of buying an EV vs a combustion engine vehicle.

For EVs: How to account for Well To Wheel GHG emissions from vehicle and battery manufacturing, from battery materials mining and transportation, from local electricity generation & distribution…?

For combustion engine vehicles: How to account for Well To Wheel GHG emissions from primary fuel extraction & transportation, vehicle fuel refinement and distribution…?

Develop a GHG emissions comparison tool.

Sunrun Challenge

Challenge: Short Term Production and Consumption Forecasting using Commodity Weather Data.

Context: With the increased integration of Residential Solar Assets into the grid, it’s imperative that we get a looking forward view into the performance of solar + storage assets in order to determine our capacity to do grid services without impacting the backup capability of the battery.

Bonus: Determine what the potential financial impact to the customer is and evaluate a scenario where this is minimized and/or calculate compensation.

We are seeking to answer the following sorts of questions:

Given a minimum 20% storage reserve, how much power can we offset one hour from now?

Given a minimum 20% storage reserve, how much power can we offset four hours from now?

Given a minimum 20% storage reserve, how much power can we offset 24 hours from now?

Can we safely dip into the storage reserve today to handle an event and make it up over the next few days?

What is the confidence interval for production forecasting over the next 1, 3, and 7 days?

Datasets:

DarkSky API - Historical hourly weather, hourly forecasts out to 7 days, minute forecasts out to 1 hour.

Sample 100 Sunrun Assets with production/consumption data and system characteristics. (anonymized, located in 1-2 CA utilities.)

$10,000 in prizes

1. Purchase a ticket 

2. Register on DevPost 

3. Come ready to hack on Friday April 13 at 6:00pm at The Citizen Engagement Lab in Oakland 

• Mission
  How well does the solution address the goals defined for this challenge? (20%)
• Quality
  How creative, innovative, interesting, and unique is the solution in meeting contest requirements? (20%)
• Implementation
  How well is the idea executed by the team and how well is the solution integrated with potential customers or systems? (20%)
• User Experience
  How successful is the design, user functionality, graphics, typography, ease of use and visual aesthetic? (20%)
• Potential Impact
  To what extent will the submission impact the clean energy industry? How scaleable is the submission? (20%)