Towards a Sustainable Microgrid on Alderney Island Using a Python-based Energy Planning Tool
Date Issued
October 2020
Abstract
In remote or islanded communities, the use of microgrids (MGs) is necessary to ensure electrification and resilience of supply. However,
even in small-scale systems, it is computationally and mathematically challenging to design low-cost, optimal, sustainable solutions
taking into consideration all the uncertainties of load demands and power generations from renewable energy sources (RESs). This
paper uses the open-source Python-based Energy Planning (PyEPLAN) tool, developed for the design of sustainable MGs in remote
areas, on the Alderney island, the 3rd largest of the Channel Islands with a population of about 2000 people. A two-stage stochastic
model is used to optimally invest in battery storage, solar power, and wind power units. Moreover, the AC power flow equations are
modelled by a linearised version of the DistFlow model in PyEPLAN, where the investment variables are here-and-now decisions and
not a function of uncertain parameters while the operation variables are wait-and-see decisions and a function of uncertain parameters.
The k-means clustering technique is used to generate a set of best (risk-seeker), nominal (risk-neutral), and worst (risk-averse) scenarios
capturing the uncertainty spectrum using the yearly historical patterns of load demands and solar/wind power generations. The proposed
investment planning tool is a mixed-integer linear programming (MILP) model and is coded with Pyomo in PyEPLAN.
even in small-scale systems, it is computationally and mathematically challenging to design low-cost, optimal, sustainable solutions
taking into consideration all the uncertainties of load demands and power generations from renewable energy sources (RESs). This
paper uses the open-source Python-based Energy Planning (PyEPLAN) tool, developed for the design of sustainable MGs in remote
areas, on the Alderney island, the 3rd largest of the Channel Islands with a population of about 2000 people. A two-stage stochastic
model is used to optimally invest in battery storage, solar power, and wind power units. Moreover, the AC power flow equations are
modelled by a linearised version of the DistFlow model in PyEPLAN, where the investment variables are here-and-now decisions and
not a function of uncertain parameters while the operation variables are wait-and-see decisions and a function of uncertain parameters.
The k-means clustering technique is used to generate a set of best (risk-seeker), nominal (risk-neutral), and worst (risk-averse) scenarios
capturing the uncertainty spectrum using the yearly historical patterns of load demands and solar/wind power generations. The proposed
investment planning tool is a mixed-integer linear programming (MILP) model and is coded with Pyomo in PyEPLAN.
File(s)![Thumbnail Image]()
Name
manuscript.pdf
Size
1.09 MB
Format
Adobe PDF
Checksum (MD5)
0d2b4339903b2dbf4ae12656d768aceb