Example

TulipaProfileFitting.jl is primarily used to fit power production curves of renewable sources such as wind and solar. In this example, we have generated a time series for a wind power plant's power availability production using a method developed in [1], which is available as a online tool here Renewables Ninja.

The file wind_power_profile.csv contains all the details of the generated profile.

So, Let's start loading the values from the file!

[1] Staffell, I., & Pfenninger, S. (2016). Using bias-corrected reanalysis to simulate current and future wind power output. Energy, 114, 1224-1239. https://doi.org/10.1016/j.energy.2016.08.068

using TulipaProfileFitting
using CSV
using Plots
using DataFrames
using HTTP

file_url = "https://raw.githubusercontent.com/TulipaEnergy/TulipaProfileFitting.jl/main/docs/src/files/wind_power_profile.csv"

df = DataFrame(CSV.File(HTTP.get(file_url).body, header=4))

From the dataframe we can get the profile values as follows:

profile_values = df.electricity

The current capacity factor of full load hours (e.g., mean value) can be determined using the following command:

current_cp = round(sum(profile_values)/8760;digits=2)

Let's define a new capacity factor of full load hours as 0.6

target_cp = 0.6

We can use the function find_solution in this package to obtain the coefficient that fit the values to the target.

coefficient = find_solution(profile_values, target_cp)

The coefficient is the output of solving a non-linear equation $f(x)=0$. To learn more, please refer to the mathematical formulation section.

Using the coefficient, we can determine the new profile and plot the power production curves.

fitted_profile = profile_values.^coefficient

We can plot the results using:

plot(profile_values, label="profile")
plot!(fitted_profile, label="fitted")
plot!(title="Profiles values - not sorted")

plot(sort(profile_values,rev=true), label="profile")
plot!(sort(fitted_profile,rev=true), label="fitted")
plot!(title="Profiles values - sorted")

It can be observed that the fitted curve primarily affects the intermediate value within the entire range. The values closer to one or zero remain relatively unchanged. This outcome is expected since the primary objective of the package is to adjust the intermediate value, thereby increasing or decreasing the capacity factor.

This example can also be found as a notebook in the repository here: TulipaProfileFitting.jl/notebooks