R/calcCapacityFactor.R

In pik-piam/mrremind: MadRat REMIND Input Data Package

#' @title calc Capacity Factor
#' @description provides capacity factor values
#' @return magpie object of the capacity factor data
#' @author Renato Rodrigues, Stephen Bi
#' @importFrom plyr round_any
#' @examples
#' \dontrun{
#' calcOutput("CapacityFactor")
#' }
#'
calcCapacityFactor <- function() {
  ### calculation of coal power capacity factor
  GWh_2_EJ <- 3.6e-6

  # Read capacity factor inputs
  global <- readSource("REMIND_11Regi", subtype = "capacityFactorGlobal", convert = FALSE)

  # remove no longer used items
  notUsed <- c("apcarelt", "aptrnelt", "apcarh2t", "apcarpet", "apcardit", "apcardiefft", "apcardieffH2t")
  global <- global[, , notUsed, invert = TRUE]

  # Set coal plant capacity factor long-term assumption to 50% (down from 60%)
  global[, , "pc"] <- 0.5
  # Read capacity factor rules
  rules <- readSource("REMIND_11Regi", subtype = "capacityFactorRules")

  #   Creating new MAgPIE object to store the final capacity values
  output <- new.magpie(getRegions(rules), seq(2005, 2150, 5), getNames(global))

  # Merging global and rules values
  # Filling MagPIE object with global values
  output[, , getNames(global)] <- global[, , getNames(global)]

  ### Global Coal Plant Tracker Calcs ###
  # Read coal capacity and generation data to derive historical capacity factor rules
  # Read generation data from Energy Balances
  coalgen_c <- calcOutput("IO", subtype = "output", aggregate = FALSE)[, , "pecoal.seel"]
  coalgen_c <- dimSums(coalgen_c, dim = 3)
  map <- toolGetMapping(getConfig("regionmapping"), type = "regional", where = "mappingfolder")
  coalgen_R <- toolAggregate(coalgen_c, map, weight = NULL)
  getNames(coalgen_c) <- "pc"
  getNames(coalgen_R) <- "pc"

  # Read coal capacity data from GCPT
  hist_cap_coal_c <- readSource("GCPT", subtype = "historical", convert = FALSE)
  hist_cap_coal_R <- toolAggregate(hist_cap_coal_c, rel = map, weight = NULL)

  # Calculate historical 5-year average capacity factors by country
  coal_factor_c <- new.magpie(getRegions(hist_cap_coal_c), years = seq(2005, round_any(max(getYears(coalgen_c, as.integer = TRUE)), 5, f = floor)), names = "pc", fill = 0)
  coal_factor_R <- new.magpie(getRegions(hist_cap_coal_R), years = getYears(coal_factor_c), names = "pc", fill = 0)
  for (i in getYears(coal_factor_c, as.integer = TRUE)) {
    if ((i + 2) %in% getYears(coalgen_c, as.integer = TRUE)) {
      coal_factor_c[, i, ] <- dimSums(coalgen_c[, (i - 2):(i + 2), ], dim = 2) / (dimSums(hist_cap_coal_c[, (i - 2):(i + 2), ], dim = 2) * 365 * 24 * GWh_2_EJ)
      coal_factor_R[, i, ] <- dimSums(coalgen_R[, (i - 2):(i + 2), ], dim = 2) / (dimSums(hist_cap_coal_R[, (i - 2):(i + 2), ], dim = 2) * 365 * 24 * GWh_2_EJ)
      # If latest available data falls in year 4 of a 5-year timestep
    } else if ((i + 1) %in% getYears(coalgen_c, as.integer = TRUE)) {
      # Special treatment of 2020 due to COVID-19: half-weight 2020
      if (i == 2020) {
        coal_factor_c[, i, ] <- (dimSums(coalgen_c[, (i - 2):(i + 1), ], dim = 2) - 0.5 * dimSums(coalgen_c[, (i), ], dim = 2)) /
          ((dimSums(hist_cap_coal_c[, (i - 2):(i + 1), ], dim = 2) + dimSums(hist_cap_coal_c[, (i + 1), ], dim = 2)) * 365 * 24 * GWh_2_EJ)
        coal_factor_R[, i, ] <- (dimSums(coalgen_R[, (i - 2):(i + 1), ], dim = 2) + dimSums(coalgen_R[, (i + 1), ], dim = 2)) /
          ((dimSums(hist_cap_coal_R[, (i - 2):(i + 1), ], dim = 2) - 0.5 * dimSums(hist_cap_coal_R[, (i), ], dim = 2)) * 365 * 24 * GWh_2_EJ)
        # For other timesteps, just use a 4-year average
      } else {
        coal_factor_c[, i, ] <- dimSums(coalgen_c[, (i - 2):(i + 1), ], dim = 2) / (dimSums(hist_cap_coal_c[, (i - 2):(i + 1), ], dim = 2) * 365 * 24 * GWh_2_EJ)
        coal_factor_R[, i, ] <- dimSums(coalgen_R[, (i - 2):(i + 1), ], dim = 2) / (dimSums(hist_cap_coal_R[, (i - 2):(i + 1), ], dim = 2) * 365 * 24 * GWh_2_EJ)
      }
      # If latest available data falls in the middle of a timestep
    } else {
      # if only data until 2020 is available, assign a double-weight to 2018 and 2019
      if (i == 2020) {
        coal_factor_c[, i, ] <- (dimSums(coalgen_c[, (i - 2):i, ], dim = 2) + dimSums(coalgen_c[, (i - 2):(i - 1), ], dim = 2)) /
          ((dimSums(hist_cap_coal_c[, (i - 2):i, ], dim = 2) + dimSums(hist_cap_coal_c[, (i - 2):(i - 1), ], dim = 2)) * 365 * 24 * GWh_2_EJ)
        coal_factor_R[, i, ] <- (dimSums(coalgen_R[, (i - 2):i, ], dim = 2) + dimSums(coalgen_R[, (i - 2):(i - 1), ], dim = 2)) /
          ((dimSums(hist_cap_coal_R[, (i - 2):i, ], dim = 2) + dimSums(hist_cap_coal_R[, (i - 2):(i - 1), ], dim = 2)) * 365 * 24 * GWh_2_EJ)
        # For other timesteps, assign triple-weight to final year
        # Valid assumption for 2015 capacity factors
      } else {
        coal_factor_c[, i, ] <- (dimSums(coalgen_c[, (i - 2):i, ], dim = 2) + 2 * coalgen_c[, i, ]) /
          ((dimSums(hist_cap_coal_c[, (i - 2):i, ], dim = 2) + 2 * hist_cap_coal_c[, i, ]) * 365 * 24 * GWh_2_EJ)
        coal_factor_R[, i, ] <- (dimSums(coalgen_R[, (i - 2):i, ], dim = 2) + 2 * coalgen_R[, i, ]) /
          ((dimSums(hist_cap_coal_R[, (i - 2):i, ], dim = 2) + 2 * hist_cap_coal_R[, i, ]) * 365 * 24 * GWh_2_EJ)
      }
    }
    # Replace countries without coal power with regional capacity factor.
    coal_factor_c[, i, ][which(!is.finite(coal_factor_c[, i, ]))] <-
      coal_factor_R[map$RegionCode[which(map$CountryCode %in% getRegions(coal_factor_c)[which(!is.finite(coal_factor_c[, i, ]))])], i, ]
    # The derived coal capacity factors for Bosnia, Ireland and Myanmar are > 1 in some years for as yet unclear reasons.
    coal_factor_c[, i, ][which(coal_factor_c[, i, ] > 1)] <- max(coal_factor_c[, i, ][which(coal_factor_c[, i, ] < 1)])
  }

  # Derive coal capacity factor rule assumptions for 2020 - 2035 (linear convergence of historical trends to global default)
  hist_yr <- "y2015"
  start_yr <- "y2020"
  end_yr <- "y2030"
  conv_yr <- "y2035"

  coal_factor_c_years <- intersect(getYears(coal_factor_c), getYears(output))
  coal_factor_c_years <- coal_factor_c_years[coal_factor_c_years < start_yr]
  output[, getYears(output)[getYears(output) < start_yr], "pc"] <-
    coal_factor_c[, coal_factor_c_years, ]

  output[, getYears(output)[which(getYears(output) >= conv_yr)], "pc"] <- global[, , "pc"]
  slope <- (output[, conv_yr, "pc"] - output[, hist_yr, "pc"]) / (as.numeric(gsub("y", "", conv_yr)) - as.numeric(gsub("y", "", hist_yr)))
  for (t in getYears(output)[which(getYears(output) >= start_yr & getYears(output) <= end_yr)]) {
    output[, t, "pc"] <- output[, hist_yr, "pc"] + slope * (as.numeric(gsub("y", "", t)) - as.numeric(gsub("y", "", hist_yr)))
  }

  # Overwriting MAgPie object with rules values
  output[getRegions(rules), getYears(rules), getNames(rules)] <- ifelse(rules[getRegions(rules), getYears(rules), getNames(rules)] != 0, rules[getRegions(rules), getYears(rules), getNames(rules)], output[getRegions(rules), getYears(rules), getNames(rules)])

  # change coal capacity factor in Germany to reflect observed decrease in coal electricity in recent years, note: to be checked whether necessary for other regions as well
  # https://static.agora-energiewende.de/fileadmin/Projekte/2021/2020_01_Jahresauswertung_2020/200_A-EW_Jahresauswertung_2020_WEB.pdf
  output["DEU", c("y2020", "y2025"), "pc"] <- 0.43
  output["DEU", c("y2030"), "pc"] <- 0.4

  weight <- calcOutput("FE", source = "IEA", aggregate = FALSE)[, 2015, "FE (EJ/yr)"]

  # Return regions aggregation weighted by final energy
  return(list(
    x = output, weight = weight,
    unit = "% of capacity",
    description = "Installed capacity availability - capacity factor (fraction of the year that a plant is running)"
  ))
}