#21 a try to make understandThePrice ....

rte-antares-rpackage · Nov 27, 2018 · 7586bbf · 7586bbf
1 parent 30cf7a3
commit 7586bbf
Show file tree

Hide file tree

Showing 4 changed files with 428 additions and 0 deletions.
diff --git a/NAMESPACE b/NAMESPACE
@@ -1,26 +1,32 @@
 # Generated by roxygen2: do not edit by hand
 
 export(.setAliasH5)
+export(addConvergencePriceArea)
+export(addConvergencePriceSystem)
 export(addDownwardMargin)
 export(addExportAndImport)
 export(addLoadFactorLink)
 export(addMonotones)
+export(addNeighbours)
 export(addNetLoad)
 export(addProcessingH5)
 export(addUpwardMargin)
 export(compare)
 export(correctBalance)
 export(externalDependency)
+export(getAllNeighbours)
 export(getValues)
 export(loadFactor)
 export(mergeAllAntaresData)
 export(modulation)
+export(neighbours)
 export(netLoadRamp)
 export(surplus)
 export(surplusClusters)
 export(surplusSectors)
 export(synthesize)
 export(thermalGroupCapacities)
+export(understandThePrice)
 import(antaresRead)
 import(data.table)
 importFrom(methods,is)

diff --git a/R/understandPrice.R b/R/understandPrice.R
@@ -0,0 +1,280 @@
+#Copyright © 2016 RTE Réseau de transport d’électricité
+
+#' understandThePrice
+#'
+#' This function gives some explanations about the price.
+#' @param antaresData Object of class \code{antaresData} created with function
+#'   \code{\link[antaresRead]{readAntares}}. antaresData must contains hourly detail data
+#'   for areas, clusters and links.
+#'   Areas data must contains "H. STOR".
+#'   Links data must contains transCapacityDirect and transCapacityIndirect.
+#'   Clusters data must contains minGenModulation.
+#'
+#' @return
+#' \code{understandPrice} modifies its input by adding to clusters
+#' data columns "market.bid.cost" and "must.run".
+#'
+#' @examples
+#' \dontrun{
+#' # First simulation
+#' studyPath <- "path/to/study/"
+#'
+#' setSimulationPath(studyPath, 1)
+#' myData <- readAntares(showProgress = FALSE,
+#'                       mcYears = "all",
+#'                       areas = "all",
+#'                       clusters = "all")
+#' understandThePrice(antaresData = myData)
+#'
+#' }
+#'
+#' @export
+#'
+understandThePrice <- function(antaresData = NULL, opts = simOptions()){
+
+  .check_x(antaresData)
+
+  if(attr(antaresData, "synthesis")){
+    stop("antaresData are synthesis, addMonotones() needs detail data.")
+  }
+  if(is(antaresData, "antaresDataList")){
+    for(tableAnta in c("areas", "links", "clusters")){
+      if(!(tableAnta %in% names(antaresData))){
+        stop("antaresData does not contain areas, clusters and links data")
+      }
+    }
+  }else{
+    stop("antaresData must be an antaresDataList with areas, clusters and links data.")
+  }
+
+  if(is.null(antaresData$links$congestion)){
+    if(is.null(antaresData$links$transCapacityDirect) & is.null(antaresData$links$transCapacityIndirect)){
+      stop("antaresData links must contains transCapacityDirect and transCapacityIndirect.")
+    }
+  }
+
+  if(is.null(antaresData$clusters$minGenModulation)){
+    stop("antaresData clusters must contains minGenModulation.")
+  }
+  if(is.null(antaresData$links$congestion)){
+    addLoadFactorLink(antaresData)
+  }
+
+  #global variable for this function
+  joinArea <- c("area", "mcYear", "timeId")
+
+  # we need column import, export
+  if(is.null(antaresData$areas$import)){
+    addExportAndImport(antaresData, addCapacities = TRUE, opts = opts)
+  }
+  #addPriceConvergenceArea to get the cheaper cluster in the system
+  if(is.null(antaresData$areas$priceConvergenceArea)){
+    addConvergencePriceArea(antaresData)
+  }
+
+  clDes <- readClusterDesc(opts = opts)
+  .cleanDataCluster(clDes)
+
+  areaInPriceConvergenceArea <- unique(unlist(strsplit(paste(unique(antaresData$areas$priceConvergenceArea),
+                                                             collapse = " "), split= " ")))
+
+  for(convergeArea in areaInPriceConvergenceArea){
+    realNeighbours <- getAllNeighbours(areasString = convergeArea, virtualAreas = virtualAreas)
+    for(nei in realNeighbours){
+      if(!(nei %in% unique(antaresData$clusters$area)) & nei %in% unique(clDes$area)){
+        stop(paste0("Import clusters data for ' ", nei, " 'to be able to get the cheapest cluster for ' ",
+                    convergeArea, " ' "))
+      }
+    }
+  }
+
+  #A compute the last cluster running when there is no hydro
+  #A.1 get the market bids of clusters and merge the data
+
+  byJoin <- c("area", "cluster")
+  colToAdd <- c("market.bid.cost", "must.run", "unitcount", "nominalcapacity", "group", "spinning")
+  antaresData$clusters[clDes,
+              (colToAdd) := mget(paste0("i.", colToAdd)),
+              on = byJoin]
+
+  #A.2.1 get for each area the most expensive cluster
+  antaresData$clusters[is.na(minGenModulation), minGenModulation:=0L]
+  antaresData$clusters[is.na(spinning), spinning:=0L]
+  antaresData$clusters[, maxProdStarted := as.integer(nominalcapacity*NODU*(1-spinning/100))]
+  antaresData$clusters[, remaiCapaStarted := as.integer(maxProdStarted - production)]
+  antaresData$clusters[remaiCapaStarted < 0, remaiCapaStarted:=0]
+  antaresData$clusters[, minProduction := as.integer(minGenModulation*nominalcapacity*NODU)]
+  antaresData$clusters[, marketBid := as.double(market.bid.cost*marketBidModulation)]
+  clMostExpensive <- antaresData$clusters[production > 0 ,
+                       list(cluster,
+                            marketBid,
+                            group,
+                            minProduction,
+                            production,
+                            remaiCapaStarted),
+                       by = joinArea]
+
+  nameOrigin <- c("cluster", "area", "group", "marketBid")
+  nameFinal <- c("marginalCluster", "marginalArea", "marginalGroup", "marBidMarginalCluster")
+  #get the cheapest cluster running in the priceConvergenceArea
+  clMostExpensive <- clMostExpensive[,.SD[remaiCapaStarted > 0,],
+                                     by = .(area, mcYear, timeId)][, .SD[marketBid == min(marketBid)]
+                                                                                   , by = .(area, mcYear, timeId)]
+
+  #get the cheaper cluster running in the priceconvergeArea
+  clMostExpensive[antaresData$areas, priceConvergenceArea := mget("priceConvergenceArea"),
+           on = joinArea]
+
+  convergenceMcYear <- unique(clMostExpensive[, .(priceConvergenceArea), by = .(mcYear, timeId)])
+
+
+  clMostExpensive[, priceConvergenceArea := convergenceMcYear[stringi::stri_detect_fixed(pattern = area,
+                                                                                str = priceConvergenceArea),
+                                                     priceConvergenceArea],
+        by = .(area, mcYear, timeId)]
+
+  #get a cluster that have a remaining positive capacity  and get the cheaper
+  clMostExpensive <- clMostExpensive[,.SD[remaiCapaStarted > 0,],
+                                     by = .(priceConvergenceArea, mcYear, timeId)][, .SD[marketBid == min(marketBid)]
+                                                                         , by = .(priceConvergenceArea, mcYear, timeId)][
+                                                                           , (nameFinal) := mget(nameOrigin),]
+
+  antaresData$areas[clMostExpensive,
+                    (nameFinal) := mget(paste0("i.", nameFinal)),
+                    on = c("priceConvergenceArea", "mcYear", "timeId")]
+
+  # to explain the price, we cannot take into account timeId when there is storage/prod STEP
+  addProdHydroConvergenceArea(antaresData)
+
+  resAttr <- attributes(antaresData)
+  colNameVirtualAreas <- resAttr$virtualNodes$storageFlexibility
+  #A.3 get the timeIds when thermalCluster make the price (H.STRO = 0 and
+  # storage/flex = 0) in all area in convergencePriceArea
+  # myDataTimeId <- antaresData$areas[prodHydroConvergeArea==FALSE,
+  #                                   mget(c("area", "timeId", "mcYear", colNameVirtualAreas))]
+  # for(cName in colNameVirtualAreas){
+  #   myDataTimeId <- myDataTimeId[get(cName)==0, ]
+  # }
+  # timeIdsToSelect <- myDataTimeId[,  .(timeId), by=c("area", "mcYear")]
+  #
+  # timeIdsToSelect[antaresData$areas, priceConvergenceArea := mget("priceConvergenceArea"),
+  #                 on = joinArea]
+  #
+  # #some times there is stro/flex in one area for one ConvergenceArea
+  # timeIdsToSelect[, lengthDim := length(unlist(strsplit(priceConvergenceArea,
+  #                                                       split = " " ))),
+  #                 by = .(area, mcYear, timeId)]
+  #
+  # timeIdsToSelect[, realDim := dim(.SD)[1],
+  #                 by = .(mcYear, timeId)]
+  # timeIdsToSelect[, indexOK := lengthDim==realDim]
+  # timeIdsToSelect <- timeIdsToSelect[indexOK==TRUE,]
+
+  for(mcY in unique(antaresData$areas$mcYear)){
+      #A.4 the price is made by the last running cluster in the priceConvergenceArea or by a cluster in another
+      #area (but we must have remaining capacity of import > 0)
+      antaresData$areas[prodFlexConvergeArea==FALSE & prodHydroConvergeArea==FALSE & mcYear==mcY,
+                        reasonPrice := paste0("The price is made by the cluster ",
+                                              get(nameFinal[1]), " in the area ",
+                                              get(nameFinal[2]), ". Its group is : ",
+                                              get(nameFinal[3]),  ". Its cost is : ",
+                                              get(nameFinal[4]), " euro/Mwh."),
+                        by = c("priceConvergenceArea", "mcYear", "timeId")]
+    }
+
+  # congestion == 1
+  # if all the link are congestionned then the price is determined by a thermal cluster in the area
+  #B.1 get the timeIds when thermalCluster make the price (congestion == 1)
+
+  # sapply(as.character(unique(antaresData$areas$area)), FUN = function(myArea){
+  #   #all the links from/to an area are congestionned
+  #   realLinksOneArea <- getLinks(opts = opts,
+  #            areas = myArea,
+  #            exclude = c(resAttr$virtualNodes$storageFlexibility,
+  #                        resAttr$virtualNodes$production))
+  #   #all links must be here
+  #   allLinksHere <- TRUE
+  #   for(realLink in realLinksOneArea){
+  #     if(realLink %in% unique(antaresData$links$link)){
+  #       allLinksHere <- allLinksHere & TRUE
+  #     }else{
+  #       allLinksHere <- allLinksHere & FALSE
+  #     }
+  #   }
+  #   if(allLinksHere){
+  #     #get the timeIds where all links are congestionned
+  #     congesData <- antaresData$links[link %in% realLinksOneArea,
+  #                                     congestion,
+  #                                     by = .(link, timeId, mcYear)]
+  #     #if no row, do nothing
+  #     nRowCongestionned <- congesData[, .N, ]
+  #     if(nRowCongestionned > 0){
+  #       resLinksAllCongestionned <- congesData[, allLinksCon := min(congestion),
+  #                                              by = .(timeId, mcYear)][,c("congestion", "link"):=NULL]
+  #       resLinksAllCongestionned[, area := myArea]
+  #       resLinksAllCongestionned <- unique(resLinksAllCongestionned)
+  #       antaresData$areas[resLinksAllCongestionned,
+  #                         ("allLinksCon") := mget(paste0("i.", "allLinksCon")),
+  #                         on = .(area, timeId, mcYear)]
+  #       antaresData$areas[area==myArea & is.na(allLinksCon), allLinksCon:=0]
+  #       #TODO
+  #       #for this area, when allLinksCon is equal to 1 then determie which cluster is the more expensive
+  #       # for this timeIds
+  #       # get the timeIds and get the most expensiveCluster
+  #       nRowRealCongestion <- antaresData$areas[area==myArea & allLinksCon==1, .N, ]
+  #       if(nRowRealCongestion > 0){
+  #         myDataTimeId <- antaresData$areas[allLinksCon==1 & `H. STOR`==0 &
+  #                                             area==myArea, mget(c("area",
+  #                                                                  "timeId",
+  #                                                                  "mcYear",
+  #                                                                  colNameVirtualAreas))]
+  #         for(cName in colNameVirtualAreas){
+  #           myDataTimeId <- myDataTimeId[get(cName)==0, ]
+  #         }
+  #         timeIdsToSelect <- myDataTimeId[,  .(timeId), by=c("area", "mcYear")]
+  #         if(timeIdsToSelect[, .N] > 0){
+  #           #B.4 the price is made by the last running cluster in the area
+  #           for(mcY in unique(timeIdsToSelect$mcYear)){
+  #             antaresData$areas[timeId %in% timeIdsToSelect[area==myArea & mcYear==mcY, (timeId)]
+  #                               & area == myArea & mcYear==mcY,
+  #                               reasonPrice := paste0("The price is made by the cluster ",
+  #                                                     nameFinal[1], " in the ",
+  #                                                     nameFinal[2], " group. Its cost is ",
+  #                                                     nameFinal[3], " euro/Mwh."), by=c("area", "mcYear", "timeId")]
+  #           }
+  #         }
+  #       }
+  #     }
+  #   }
+  # })
+
+  # print a report
+  numberLines <- antaresData$areas[, .N]
+  numberLinesExplain <- antaresData$areas[!is.na(reasonPrice), .N]
+  message("Percentage of lines explained : ",
+          as.character(round(numberLinesExplain/numberLines*100)), "%")
+  invisible(antaresData)
+}
+
+addProdHydroConvergenceArea <- function(antaresData = NULL){
+  antaresData$areas[, maxHydroConvergeArea := max(`H. STOR`),
+                    by = c("priceConvergenceArea", "mcYear", "timeId")]
+  antaresData$areas[, prodHydroConvergeArea := ifelse(maxHydroConvergeArea>0,
+                                                      TRUE,
+                                                      FALSE)]
+
+  resAttr <- attributes(antaresData)
+  colNameVirtualAreas <- resAttr$virtualNodes$storageFlexibility
+  antaresData$areas[, maxProdStorVirtual := do.call(pmax, .SD[,mget(colNameVirtualAreas)]),
+                 by = c("priceConvergenceArea", "mcYear", "timeId")]
+  antaresData$areas[, minProdStorVirtual := do.call(pmin, .SD[,mget(colNameVirtualAreas)]),
+                 by = c("priceConvergenceArea", "mcYear", "timeId")]
+  antaresData$areas[, absMinMaxStorVirtual := max(abs(minProdStorVirtual),
+                                               maxProdStorVirtual),
+                 by = c("priceConvergenceArea", "mcYear", "timeId")]
+
+  antaresData$areas[, prodFlexConvergeArea := ifelse(absMinMaxStorVirtual>0,
+                                                      TRUE,
+                                                      FALSE)]
+  invisible(antaresData)
+}
diff --git a/man/understandThePrice.Rd b/man/understandThePrice.Rd