In EnergyEconomyDecoupling/PFUDatabase: Functions and Workflows for Primary-Final-Useful Societal Exergy Analysis Within the PSUT Framework

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

library(dplyr)
library(PFUPipeline)
library(PFUSetup)
library(pins)
library(Recca)

Introduction

This vignette demonstrates the features of the PFUPipeline package.

Load PFUPipeline

PFUPipeline is a package that runs a targets pipeline and stores data in Dropbox using the pins package.

To retrieve data, one needs

• access to the Dropbox folder containing pipeline releases and
• a valid pin name.

One valid pin name is "psut" and a valid pin release is "20220828T174526Z-60a07". "20220828T174526Z-60a07" contains data for all countries (155) and all years (1960--2019).

folder <- PFUSetup::get_abs_paths()[["pipeline_releases_folder"]]
folder
# version <- "20220828T174526Z-60a07" # All countries
version <- "20220909T125050Z-e5a95" # Only GHA
psut_mats <- folder %>% 
  pins::board_folder(versioned = TRUE) %>% 
  pins::pin_read("psut", version = version)

Data structure

psut_mats is an R data frame. Let's look at its contents.

dplyr::glimpse(psut_mats)

The metadata columns contain the following values.

psut_mats$Country %>% unique()
psut_mats$Method %>% unique()
psut_mats$Energy.type %>% unique()
psut_mats$Last.stage %>% unique()
c(psut_mats$Year %>% min(), psut_mats$Year %>% max())
psut_mats$IEAMW %>% unique()

Example ECC matrices

Let's look at a few of matrices using Ghana 1971 as an example. Simple matrices can be found in the muscle work data frames.

GHA1971_MW_final <- psut_mats %>% 
  dplyr::filter(Country == "GHA", Year == 1971, Energy.type == "E", 
                Last.stage == "Final", IEAMW == "MW")
dplyr::glimpse(GHA1971_MW_final)
GHA1971_MW_final$R[[1]]
GHA1971_MW_final$U[[1]]
GHA1971_MW_final$V[[1]]
GHA1971_MW_final$Y[[1]]

These matrices can be dumped to an Excel file for easier browsing.

gha_mw_file <- file.path("~", "GHA1971_MW_final.xlsx")
Recca::write_ecc_to_excel(GHA1971_MW_final, path = gha_mw_file, overwrite_file = TRUE)

Or a Sankey diagram can be created.

GHA1971_MW_final %>% 
  Recca::make_sankey() %>% 
  magrittr::extract2("Sankey") %>% 
  magrittr::extract2(1)

When useful energy is the last stage, we see the food and feed creating muscle work.

GHA1971_MW_useful <- psut_mats %>% 
  dplyr::filter(Country == "GHA", Year == 1971, Energy.type == "E", 
                Last.stage == "Useful", IEAMW == "MW")
GHA1971_MW_useful %>% 
  Recca::make_sankey() %>% 
  magrittr::extract2("Sankey") %>% 
  magrittr::extract2(1)

The IEA data give rise to more complicated matrices. This one is has the last stage being final energy.

GHA1971_IEA_final <- psut_mats %>% 
  dplyr::filter(Country == "GHA", Year == 1971, Energy.type == "E", 
                Last.stage == "Final", IEAMW == "IEA")
gha_iea_file <- file.path("~", "GHA1971_IEA_final.xlsx")
Recca::write_ecc_to_excel(GHA1971_IEA_final, path = gha_iea_file, overwrite_file = TRUE)

GHA1971_IEA_final %>% 
  Recca::make_sankey() %>% 
  magrittr::extract2("Sankey") %>% 
  magrittr::extract2(1)

When we push to useful energy, the matrices become more complex.

GHA1971_IEA_useful <- psut_mats %>% 
  dplyr::filter(Country == "GHA", Year == 1971, Energy.type == "E", 
                Last.stage == "Useful", IEAMW == "IEA")
gha_iea_file <- file.path("~", "GHA1971_IEA_useful.xlsx")
Recca::write_ecc_to_excel(GHA1971_IEA_useful, path = gha_iea_file, overwrite_file = TRUE)

And the Sankey diagram is unsurprisingly busier, but the matrices keep everything organized.

GHA1971_IEA_useful %>% 
  Recca::make_sankey() %>% 
  magrittr::extract2("Sankey") %>% 
  magrittr::extract2(1)

Example calculations

The following subsections illustrate example calculations.

Matrix math

Here we calculate the value added matrix (W) within the data frame using dplyr::mutate().

with_W <- GHA1971_MW_final %>% 
  dplyr::mutate(
    W = matsbyname::transpose_byname(GHA1971_MW_final$V) %>%
      matsbyname::difference_byname(GHA1971_MW_final$U)
  )
dplyr::glimpse(with_W)
with_W$W[[1]]

Upstream swim

The PSUT framework allows upstream and downstream "swims."

Here is the final demand matrix for the IEA data.

# View(GHA1971_IEA_final$Y[[1]])

Let's say we wanted to know the energy (at all stages) required to provide residential energy in the country. We can use an input-output ``upstream swim'' for that analysis.

First, we calculate all the input-output matrices.

with_io_mats <- GHA1971_IEA_final %>% 
  Recca::calc_io_mats()
dplyr::glimpse(with_io_mats)

Next, we define a Y_prime matrix that contains only final demand from residences.

with_Y_prime <- with_io_mats %>% 
  dplyr::mutate(
    Y_prime = Y %>% 
      matsbyname::select_cols_byname("Residential")
  )
with_Y_prime$Y_prime[[1]]

Now let's swim upstream from the Y_prime matrix to see what upstream energy is needed to supply Residential final demand.

residential_ecc <- with_Y_prime %>% 
  Recca::new_Y()
dplyr::glimpse(residential_ecc)

The _prime matrices are what we want. So we keep those and make a new Sankey diagram.

residential_ecc_trimmed <- residential_ecc %>% 
  dplyr::select(Country, Method, Energy.type, Last.stage, Year, IEAMW,
                R_prime, U_prime, U_feed_prime, U_EIOU_prime, 
                r_EIOU_prime, V_prime, Y_prime, S_units) %>% 
  dplyr::rename(
    R = R_prime, 
    U = U_prime, 
    U_feed = U_feed_prime,
    U_EIOU = U_EIOU_prime,
    r_EIOU = r_EIOU_prime,
    V = V_prime, 
    Y = Y_prime
  ) %>% 
  dplyr::mutate(Sector = "Residential")
dplyr::glimpse(residential_ecc_trimmed)

Make a Sankey diagram of energy required to supply Residential demand only.

residential_ecc_trimmed %>% 
  Recca::make_sankey() %>% 
  magrittr::extract2("Sankey") %>% 
  magrittr::extract2(1)

From the Sankey diagram, we can see all the ways in which Ghanaian residences demanded upstream primary energy in 1971.

Efficiencies

With the Residential sector isolated, we can calculate the efficiency of energy use in residences.

primary_industries <- IEATools::tpes_flows
primary_industries
finaldemand_sectors <- IEATools::fd_sectors
finaldemand_sectors
with_efficiencies <- residential_ecc_trimmed %>% 
  Recca::primary_aggregates(p_industries = primary_industries, pattern_type = "leading") %>% 
  Recca::finaldemand_aggregates(fd_sectors = finaldemand_sectors) %>%
  dplyr::mutate(
    eta_pf_net = as.numeric(EX.fd_net) / as.numeric(EX.p),
    eta_pf_gross = as.numeric(EX.fd_gross) / as.numeric(EX.p)
  )
dplyr::glimpse(with_efficiencies)

Pipelines

A different R package (PFUAggDatabase) does the above calculations (and more!) for all countries; all years; energy and exergy; last stage final and useful; IEA, MW, and both; etc. The calculations are parallelized across countries. Despite parallelization, the calculations take several hours on a 10-core laptop, We will (eventually) generating the PFUAggDatabase on a supercomputer.

Conclusion

In combination, the PSUT framework and the R packages matsbyname, matsindf, and Recca provide powerful analysis tools for energy conversion chains.

EnergyEconomyDecoupling/PFUDatabase documentation built on Feb. 8, 2024, 8:45 a.m.