SISREMO

Simple Storage Reqirement Model for a 100% Renewable Energy System

There are many researches dedicated to 100% renewable energy system modelling and their models are much more sophisticated than the model presented here. Some publications in this field are:

1. T. M. Clack etal., Evaluation of a proposal for reliable low-cost grid power with 100% wind, water, and solar, Proc Natl Acad Sci USA, 14, 16, (2017), www.pnas.org/cgi/doi/10.1073/pnas.1610381114
2. Ziegler etal., Storage Requirements and Costs of Shaping Renewable Energy Toward Grid Decarbonization, Joule 3, 2134–2153 (2019), https://doi.org/10.1016/j.joule.2019.06.012
3. Ch. Breyer, etal., On the History and Future of 100% Renewable Energy Systems Research, IEEE Access, 10, (2022), DOI: 10.1109/ACCESS.2022.3193402
4. D. Bogdanov, Radical transformation pathway towards sustainable electricity via evolutionary steps, Nature Communications volume 10, Article number: 1077 (2019), https://www.nature.com/articles/s41467-019-08855-1
5. O. Ruhnau, O. Qvist, Storage requirements in a 100% renewable electricity system: extreme events and inter-annual variability, Environ. Res. Lett. 17 044018, (2022), DOI 10.1088/1748-9326/ac4dc8

The model described here is intended to show the relationships between energy production, energy consumption and storage requirements to clarify a little. It is easier to understand than the models described in the literature and anyone who is interested can play with the code themselves with a little Julia programming experience.

Data

The data used are from the site Energy-Charts API

1. download power data as json from https://api.energy-charts.info/power using the REST API:
   • execute: load_ise_energy_chart_data(start_year, end_year) in storage/data_energy_charts.jl the minimum start_year is 2015
2. Parse downloaded json files and store date in a hdf5 file
   • execute run_ise_json_to_hdf5(false, 2015, 2022) in storage/data_energy_charts.jl

Energy-Chart data used in this model are:

1. Dates: UNIX timestamps are converted to Julia DateTime objects
2. Load
3. Sum of Wind offshore, Wind onshore and Solar

Load and renewable time series data

The real data are adapted to mimic a 100% renewable eneryg system by detrending and scaling.

Detrend and Scale Date

Detrending is done by fitting teh data to a second order polynomial

Detrend Load

$k$ - polynomial order $n$ - number of data points

1. Load trend, polynomial fit $L_{t} = \operatorname{polynomial_fit}(L, k)$
2. Detrend $L_{d} = \dfrac{L_{t}[n/2]}{L_{t}} ; L$

$L$ - Load [MW] $L_{t}$ - trend of Load, poynomial fit $L_{d}$ - detrended Load

Detrended Load data

Detrend and Scale Renewables

function scale_and_detrend(Load::Vector{Float64}, RP::Vector{Float64})

1. First scaling $R_{s} = R ; \dfrac{\operatorname{mean}(L)}{\operatorname{mean}(R)}$

2. Renewables trend, polynomial fit $R_t = \operatorname{polynomial_fit}(R_s, k)$

3. Detrend $R_d = \dfrac{R_t[n/2]}{R_t} R_s$

4. Scale again $R_{ds} = R_d \dfrac{\operatorname{mean}(L_d)}{\operatorname{mean}(R_d)}$

5. Diffeernce between Renewables and Laod $\Delta P = (R_{ds} - L_d)$

$R$ - renewable energy [MW] $R_{s}$ -scaled $R$ $R_d$ - detrended $R$ $R_t$ - detrended and scaled $R$

Detrended and scaled renewable power data

Differences beteeen reneable power and load

Compute Storage Fill Level as Function of Time

Given storage capacity and an overproduction capacity factor the storage fille level is computed;

compute_storage_level(dates, Load, RP, eunit, over_production, storage_capacity)

The algorithm in short is:

$\Delta P = R - L$

if $\Delta P > 0$ and $S < S_{capacity}$

$\quad S = S + \Delta P \Delta t$

elseif $\Delta P < 0$ and $S > 0$

$\quad S = S - \Delta P \Delta t$

end

Storage fill level over time for different combinations of storage capacity and renewable overproduction factor op