Isolated Island Energy System

[brunalfaustino]

Hi Andrew, your model is really amazing. Thank you for sharing it.

I am actually trying to use it for a class where we are modelling an Isolated Island Energy System. Here is the issue, the island we are modeling is located in northeastern Brazil. Very hot and sunny weather, with no heating demand and cooling accounted for within the electrical demand.

The island is mostly powered by diesel generators (4 of them with a total capacity 4978kW) and 2 solar plants that account for 10% of generation (a combined 950kwp). I have the data on the islands generation and consumption from 2013 until 2018, along with the types and number of consumers (meters). The place has a fixed resident population and a fixed fluctuating population of visitors regulated by government, both without any prospect of growing. However, there is an electrical demand growth projected for 2026, when the entire vehicle fleet of the island is to be replaced by electric vehicles.

We are set on a proposal that seems to work for this scenario and which includes at least doubling the solar generation and installing two wind turbines. This scenario runs fine in the model. However, as we try to incorporate batteries to our system, we can't seem to get it right.

Do you have any tips or instructions on how we can adjust the model to the described above. We are pretty ok with the weather data inputs we have used for both solar and wind generation. We have some technical specifications for the diesel generators and the batteries that were recently installed in the island, but we may be using these specifications incorrectly, or we may have incorrectly generated the demand, or may not have been able to fully disconnect our system from the grid... all guesses to what may be going wrong...

In any case, every and all input from you will very, very much appreciated. Thank you again!

Bruna

[andrewlyden]

Hi Bruna,

Great to hear of your interest in PyLESA. I'm very happy to support you on this.

The batteries should work for both the MPC and fixed order controllers. With the MPC you don't need to have a setpoint price and this makes it easier to integrate the batteries, have you tried using this controller? With the fixed order controller you need to input a setpoint which needs to take into consideration the grid prices. In your case with an islanded system it shouldn't matter and the controller should charge the battery with RES when there is excess available, and then discharge when there is a deficit between demand and RES.

It would be useful to see the hourly operation over a day or week so that I can figure out what is happening. It would also be very useful to see the inputs workbook which you are using incase I can help too :)

All the best, Andrew

[brunalfaustino]

Hi Andrew. Happy new year!

It looks like our model is finally working with the battery. It worked with both controllers, although it took several hours to run using the MPC. We now have a different issue. We only added the existing battery to the model, but we may need to propose the addition of a new one to make the system more efficient while isolated from the grid.

Right now we are exporting energy to the grid the days when demand is met and the battery is already full. Other days, when demand is not met and the battery doesn't have enough to complement, we are importing from the grid.

Since we are not connected to the grid, I am reading the imports as if they were coming from the generator, but we need eliminate the exports and maximize consumption from renewables. Would that extra battery work? Any tips on specifying the best fit.

If you are ok with it, please take a look at our the inputs file attached.

Thank you so much for getting back to me. By the way, I got a hold of you thesis. It really helped us better understand the inner workings of PyLesa.

Cheers,

Bruna

input_base.xlsx

[andrewlyden]

Happy new year to you too!

Yes, I should put a warning in the documentation that the MPC takes several hours to solve.

At the moment it is only possible to add a single battery. You could carefully choose the input parameters in order to represent two batteries. Simple example is doubling the energy capacity. It should work as you expect, the battery will store RES excess generation and discharge to avoid imports. Alternatively, you could add to the code with functionality for a second battery.

Thanks for sharing your input file. It looks like the fixed order controller is set with a grid import setpoint of 60. This means there are two different order of control for when the grid import price is above or below this setpoint. For an isolated system I think it would be best to ensure that the import prices input are always above the import setpoint. In this case imports will always be used as a backup (like your backup generator). If the import price is below the setpoint then imports will charge the storage (which I'm guessing you don't want to do). The input sheet is showing it is using a time-of-use tariff, this should be changed to a flat rate tariff, and this set to above the import setpoint (i.e., above 60).

Hope this helps :)

Andrew