DistributedSolar
L.Doyle
23/04/2021
Exploratory Analysis of City Run Distributed Solar Power Production Sites in Calgary
download solar energy production from Open Calgary data portal https://data.calgary.ca/Environment/Solar-Energy-Production/ytdn-2qsp
citysolar <- read.csv("./Solar_Energy_Production.csv")Explore the data
str(citysolar)## 'data.frame': 169785 obs. of 5 variables:
## $ name : chr "Southland Leisure Centre" "Southland Leisure Centre" "Southland Leisure Centre" "Southland Leisure Centre" ...
## $ id : int 164440 164440 164440 164440 164440 164440 164440 164440 164440 164440 ...
## $ address: chr "2000 SOUTHLAND DR SW" "2000 SOUTHLAND DR SW" "2000 SOUTHLAND DR SW" "2000 SOUTHLAND DR SW" ...
## $ date : chr "2017/01/02 11:00:00 AM" "2017/01/02 12:00:00 PM" "2017/01/02 01:00:00 PM" "2017/01/03 11:00:00 AM" ...
## $ kWh : num 0 0 0 0 0 0 0 0 0 0 ...names <- table(citysolar$name) ##investigate how many solar sites
namesDF <- as.data.frame(names)
namesDF## Var1 Freq
## 1 Bearspaw Water Treatment Plant 15593
## 2 Calgary Fire Hall Headquarters 16684
## 3 CFD Firehall #7 13629
## 4 City of Calgary North Corporate Warehouse 18007
## 5 Glenmore Water Treatment Plant 17077
## 6 Hillhurst Sunnyside Community Association 15849
## 7 Manchester Building M 13517
## 8 Richmond - Knob Hill Community Hall 17610
## 9 Southland Leisure Centre 23144
## 10 Telus Spark 317
## 11 Whitehorn Multi-Service Centre 18358Check for missing data
citysolar$date <- as.POSIXct(citysolar$date, format = "%Y/%m/%d %I:%M:%S %p") ##type str to date
colSums(is.na(citysolar)) ## check for missing data## name id address date kWh
## 0 0 0 0 0Check for data sampling interval
library(dplyr)
citysolar <- group_by(citysolar, name)
citysolar <- arrange(citysolar, date) ##sort grouped data by date
head(citysolar) ##indicates hourly measurements## # A tibble: 6 x 5
## # Groups: name [1]
## name id address date kWh
## <chr> <int> <chr> <dttm> <dbl>
## 1 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 14:30:00 27.8
## 2 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 15:30:00 30.6
## 3 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 16:30:00 23.6
## 4 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 17:30:00 15.3
## 5 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 18:30:00 7.90
## 6 Southland Leisure Centre 164440 2000 SOUTHLAND DR SW 2015-09-01 19:30:00 1.88Seasonal Variation in Solar Power Production at City Run Distributed Solar Sites
First apply a 30day moving average to the data to get a general sense of power production trends throughout the year. Seasonal variation in solar power production (high in the summer, low in the winter) is obvious for all the sites shown.
Use a 30 day window for the moving average. Because the data is sampled in hours we use k = 720 to get 30 day moving average (24hours * 30 days = 720) check for any rows with missing data introduced by rollmean() and remove them.
library(zoo)
citysolar <- mutate(citysolar, MA30day = rollmean(kWh, k = 720, fill = NA))
colSums(is.na(citysolar))## name id address date kWh MA30day
## 0 0 0 0 0 7507citysolar <- citysolar[complete.cases(citysolar),]
colSums(is.na(citysolar))## name id address date kWh MA30day
## 0 0 0 0 0 0Create plot
library(ggplot2)
g <- ggplot(citysolar, aes(date, MA30day, group = name, colour = name)) + geom_line() +
scale_color_brewer(palette = "Paired") + theme_dark() + theme(legend.position = "bottom") +
guides(colour = guide_legend(ncol = 2)) + ggtitle("Yearly Solar Power Production Trends") + xlab("Date") +
ylab("Power(kWh)")
g
Glenmore Water Treatment Plant in Detail
Take a closer look at the data for the Glenmore Water Treatment Plant to investigate smaller scale trends in solar power production at this location.
glenmore <- citysolar %>% filter(name == "Glenmore Water Treatment Plant")
g_glenmore <- ggplot(glenmore, aes(date, MA30day)) + geom_line(colour = "#FB9A99") +
theme_dark() + theme(legend.position = "bottom") +
guides(colour = guide_legend(ncol = 2)) + ggtitle("Glenmore Water Treatment Plant Solar Power Production") + xlab("Date") +
ylab("Power(kWh)")
g_glenmore
Year on Year Comparison
Plot the 4 years of available data against eachother to see if there are any significant differences in solar power production from year to year.
glenmore_yearly <- glenmore %>% mutate(year = format(date,"%Y"), month_day = format(date, "%m-%d %I:%M:%S %p"))
glenmore_yearly$month_day <- as.Date(glenmore_yearly$month_day, format = "%m-%d %I:%M:%S %p")
g_yearly <- ggplot(glenmore_yearly, aes(month_day, MA30day, group = year, color = year)) + geom_line() +
theme_dark() + theme(legend.position = "bottom") + scale_color_brewer(palette = "Paired") +
ggtitle("Glenmore Water Treatment Plant Year on Year Comparison") + xlab("Date") + ylab("Power(kWh)") + scale_x_date(date_labels = "%m", date_breaks = "1 month")
g_yearly
Daily Cyclicity in Solar Power Production at Glenmore Water Treatment in July
Select rows for one week in July, Look at hourly readings rather than 30day rolling average so we can see the finer grained cyclicity. Diurnal cyclicity (high production at noon, no production at night) is clearly visible
glenmore2018 <- glenmore %>% filter(date >= "2018-07-01" & date <= "2018-07-08")
##glenmore2018$date <- as.Date(glenmore2018$date)
g2018 <- ggplot(glenmore2018, aes(date, kWh)) + geom_line(colour = "#FB9A99") +
theme_dark() + theme(legend.position = "bottom") +
guides(colour = guide_legend(ncol = 2)) + ggtitle("Glenmore Water Treatment Plant 2018 Solar Power Production") + xlab("Date") + ylab("Power(kWh)")
g2018
Explore power consumption at city run distributed solar sites
THIS SECTION STILL UNDER CONSTRUCTION
EnCons <- read.csv("./Corporate_Energy_Consumption.csv")Types of energy consumption at city sites
EnergyTypes <- table(EnCons$Energy.Description)
EnergyTypesDF <- as.data.frame(EnergyTypes)
EnergyTypesDF## Var1 Freq
## 1 District Energy 309
## 2 Electricity 109896
## 3 Natural Gas 35865
## 4 Solar Power 969
## 5 Solar Thermal 68Select solar energy consumption only
SolarCons <- subset(EnCons, Energy.Description == "Solar Power")