Climate change is affecting everyone in far-reaching ways. Today 80% of electricity in the world is sourcing from fossil fuels which are a major contributor to green houses gases and are widely believed to have caused the climate change. The delivery of electricity to the customers has been relied upon by utility companies and through long power lines from central power stations vulnerable to weather conditions i.e. Porto Rico's 6 month long power outage, California Wild Fires, or cyber attacks. There is a revolution happening in terms of rise of distributed power systems known as microgird that can deliver electricity via renewable resources i.e. solar panels directly to the customers. This further has been accelerated due to dropping cost of the solar cells , and batteries. However the renewable resources are unpredictable and in order for the system to be operational it should be able to control loads such as lamps, diswasher, Air Conditioner and batteries in a dynamic way to be able to keep the system up and running. Therefore we introduce a prototype of Microgrid Energy Management System or Mems which is able to operate microgrids based on our control algorithms to keep the system running by controlling the load (lights) and batteries.

What it does

Mems is a prototype of a microgird controls which establishes a balance with Solar generation, and loads. It uses data analytics to predict the output of solar and batteries and decides whether to charge/discharge the batteries or shed the loads (lights). It uses integer programming to optimize for the longest run time given its preconditions i.e. batteries state of charge, solar output, state of the loads.

How I built it

A device called eGauge has been used for monitoring the loads. A standard control PWM plug has been used for controlling the plugs via a raspberry pi. the The battery is a regular APC modified with better batteries. APC data was intercepted via a script which is used to provide the charge status and voltage and power monitoring of the battery. A charge controller manages the charge/discharge time of the battery which we override via a data link. The data is sent to the cloud via a raspberry pi and is viewable through an smart phone app.

Challenges I ran into

Building hardware is always hard and is full of challenges. The biggest ones had to with communication of the charge controller and our system.

Accomplishments that I'm proud of

The fact the system works and all the bits and pieces are coming together last minute.

What I learned

Learned a lot about the compatibility issues battery and load control.

What's next for Microgrid Testbed

Planning to implement what has been learned from this experiment to apply it to a real world home.

Share this project: