psrmodels: a package for adequacy of supply analysis in power systems

This packages implements some useful functions to do adequacy of supply analysis in single area and 2-area power systems. The focus is on time-collapsed models, but efficient simulation functionality is also implemented for time-sequential analysis. In the case of time-collapse models, some semi-parametric extreme value models are available for analysing both net demand (demand minus renewables) and power margins.

Installation

The package is on PyPi, so

pip install psrmodels 

is enough to install the package. It runs on Python >=3.6

Usage

You can see the full documentation here. Below some toy models are created for time-collapsed and time-sequential analysis

toy bivariate time-collapsed and time-sequential models

import numpy as np
import pandas as pd

from psrmodels.time_collapsed import BivariateHindcastMargin as tc_model
from psrmodels.time_collapsed import ConvGenDistribution as tc_convgen

from psrmodels.time_dependent import BivariateHindcastMargin as td_model
from psrmodels.time_dependent import ConvGenDistribution as td_convgen

# create toy demand and wind data with 100 observations
np.random.seed(1)

demand = np.random.normal(loc=1000,scale=50,size=(100,2))

wind = np.random.normal(loc=250,scale=50,size=(100,2))

# create toy generator data

gens =  pd.DataFrame({"Capacity": 50*np.ones(15), "Availability": 0.95*np.ones(15)})

# create 2 (identical) conventional generation distributions from generator data

convgen_dists = [tc_convgen(gens),tc_convgen(gens)]

# create time-collapsed bivariate hindcast model

model1 = tc_model(demand=demand,renewables=wind,gen_dists=convgen_dists)

## compute LOLE for area 1 under a 'veto' policy an 1000MW interconnection capacity
model1.lole(c=1000,policy="veto",axis=0)

# now, create a time-sequential model

# first, create time-sequential generators. We need to add a 'TTR' (time to repair) column to our generator data
gens["TTR"] = 50 #50 hours to repair on average
td_convgen_dists = [td_convgen(gens),td_convgen(gens)]

# create time-sequential model
model2 = td_model(demand=demand,renewables=wind,gen_dists=td_convgen_dists)

# simulate post-interconnection sequential power margin values under a veto policy and 1000MW interconnection
sim_data = model2.simulate_post_itc(n_sim=1000,c=1000,policy="veto")