Skip to content

Commit

Permalink
Averaged all csv files of non-residential load shape dataset
Browse files Browse the repository at this point in the history
  • Loading branch information
eggimasv authored and eggimasv committed Mar 6, 2017
1 parent 1e56d63 commit a2f5c86
Show file tree
Hide file tree
Showing 2 changed files with 162 additions and 40 deletions.
2 changes: 1 addition & 1 deletion energy_demand/building_stock_generator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -94,7 +94,7 @@ def virtual_building_stock(data, assumptions, data_ext):
l = reg_building_stock_cur_yr[0].dwellings
for i in l:
print(i.__dict__)

# Add to data
data['reg_building_stock_by'] = reg_building_stock_by
data['reg_building_stock_cur_yr'] = reg_building_stock_cur_yr
Expand Down
200 changes: 161 additions & 39 deletions energy_demand/load_profiles_generator.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -6,49 +6,81 @@
#import datetime
from datetime import date
import unittest
import matplotlib.pyplot as plt




def read_in_non_residential_load_curves():
"""
Folder with all csv files stored
"""
# Dictionary: month, daytype, hour
#
# 0. Find what day it is (with date function). Then define daytype and month
# 1. Calculate total demand of every day
# 2. Assign percentag of total daily demand to each hour

folder_path = r'C:\Users\cenv0553\Dropbox\00-Office_oxford\07-Data\09_Carbon_Trust_advanced_metering_trial_(owen)\Community'
# out_dict_av: Every daily measurment is taken from all files and averaged
# out_dict_not_average: Every measurment of of every file is plotted

folder_path = r'C:\Users\cenv0553\Dropbox\00-Office_oxford\07-Data\09_Carbon_Trust_advanced_metering_trial_(owen)\_all_elec' #Community
#folder_path = r'C:\Users\cenv0553\Dropbox\00-Office_oxford\07-Data\09_Carbon_Trust_advanced_metering_trial_(owen)\__OWN_SEWAGE' #Community

all_csv_in_folder = os.listdir(folder_path) # Get all files in folder
print(all_csv_in_folder)

# Initialize dict with list for every hour dict = {daytype: {month: {}}
main_dict = {0: {}, 1: {}}
for i in main_dict:
month_dict = {}
for f in range(12):
#day_dict = dict.fromkeys(range(24))
day_dict = {k: [] for k in range(24)}
month_dict[f] = day_dict
main_dict[i] = month_dict

# Dict
dict_result = {0: {}, 1: {}}

# Initial dict
dict_results = np.zeros((2, 12, 24), dtype = float) # datyape, month, hours
#array_results = np.zeros((2, 12, 24), dtype = float) # datyape, month, hours

nr_of_line_entries = 0

# Itreateu folder
# Itreateu folder with csv files
for path_csv_file in all_csv_in_folder:
path_csv_file = os.path.join(folder_path, path_csv_file)
print("path: " + str(path_csv_file))

# Read csv file
with open(path_csv_file, 'r') as csv_file: # Read CSV file
read_lines = csv.reader(csv_file, delimiter=',') # Read line
with open(path_csv_file, 'r') as csv_file: # Read CSV file
read_lines = csv.reader(csv_file, delimiter=',') # Read line
_headings = next(read_lines) # Skip first row
#print("read_lines: " + str(read_lines))

# Iterate day
for row in read_lines:
#print("RO " + str(row))

#DO SOME TEST TO CHECK FIL FILE CORRECT
if len(row) != 49:
# Skip row
# Test if file has correct form
if len(row) != 49: # Skip row
continue
#assert len(row) == 49 #one date plus 48 half hourly data

#try:
# Convert all values except date into float values
row[1:] = map(float, row[1:])

#e xcept ValueError:
# # Something is wrong with original raw file...()
# continue

# Total daily sum
daily_sum = sum(row[1:])

if daily_sum == 0: # Skip row if no demand of the day
print("TOTAL SUM IS 0")
continue

# Nr of lines added
nr_of_line_entries += 1

day = int(row[0].split("/")[0])
month = int(row[0].split("/")[1])
Expand All @@ -60,51 +92,141 @@ def read_in_non_residential_load_curves():
daytype = mf.get_weekday_type(date_row)

# Month Python
month_python = month - 1

# Convert all values except date into float values
row[1:] = map(float, row[1:])
# Total daily sum
daily_sum = sum(row[1:])
#print("dailyUM: " + str(daily_sum))
month_python = month - 1

# Iterate hours
cnt = 0
h_day = 0
control_sum = 0
cnt, h_day, control_sum = 0, 0, 0

for half_hour_val in row[1:]: # Skip date
for half_hour_val in row[1:]: # Skip date
cnt += 1
if cnt == 2:
control_sum += (first_half_hour + half_hour_val)
dict_results[daytype][month_python][h_day] = (first_half_hour + half_hour_val) * (100 / daily_sum)
main_dict[daytype][month_python][h_day].append((first_half_hour + half_hour_val) * (100 / daily_sum)) # Calc percent of total daily demand
cnt = 0
h_day += 1

first_half_hour = half_hour_val #must be at this position
first_half_hour = half_hour_val # must be at this position

# Check if 100 %
assertions = unittest.TestCase('__init__')
assertions.assertAlmostEqual(control_sum, daily_sum, places=7, msg=None, delta=None)

return dict_results

out_array = read_in_non_residential_load_curves()

# -----------------------------------------------
# Calculate average and add to overall dictionary
# -----------------------------------------------

import matplotlib.pyplot as plt
# -- Calculate each value
out_dict_not_av = {0: {}, 1: {}}

# Initialise out_dict_not_av
for daytype in out_dict_not_av:
month_dict = {}
for f in range(12): # hours

# Add all daily values into a list (not average)
day_dict = {k: [] for k in range(24)}
month_dict[f] = day_dict
out_dict_not_av[daytype] = month_dict

# copy values from main_dict
for daytype in main_dict:
for month_python in main_dict[daytype]:
for h_day in main_dict[daytype][month_python]:
out_dict_not_av[daytype][month_python][h_day] = main_dict[daytype][month_python][h_day]

# Average (initialise dict)
out_dict_av = {0: {}, 1: {}}
for dtype in out_dict_av:
month_dict = {}
for month in range(12):
month_dict[month] = {k: 0 for k in range(24)}
out_dict_av[dtype] = month_dict

# Iterate daytype
for daytype in main_dict:

# Iterate month
for _month in main_dict[daytype]:

# Iterate hour
for _hr in main_dict[daytype][_month]:
nr_of_entries = len(main_dict[daytype][_month][_hr]) # nr of added entries

if nr_of_entries != 0: # Because may not contain data because not available in the csv files
out_dict_av[daytype][_month][_hr] = sum(main_dict[daytype][_month][_hr])/nr_of_entries

return out_dict_av, out_dict_not_av

out_dict_average, out_dict_not_average = read_in_non_residential_load_curves()



# Calculate average daily load shape for all mongth (averaged)

# Initiate
yearly_averaged_load_curve = {0: {}, 1: {}}
for daytype in yearly_averaged_load_curve:
yearly_averaged_load_curve[daytype] = {k: 0 for k in range(24)}

for daytype in out_dict_average:

# iterate month
h_average_y = 0
for hour in range(24):

for month in out_dict_average[daytype]:
h_average_y += out_dict_average[daytype][month][hour]

h_average_y = h_average_y/12

yearly_averaged_load_curve[daytype][hour] = h_average_y

print("Result yearly averaged:")
print(yearly_averaged_load_curve)

# -------------------
# plot
# -------------------
plot_average = True

# Select month to plot
# ---------------------
x_values = range(24)
daytype = 1 # Daytaoe


#print(y_values)


# --- if average is calculated
if plot_average == True:

# plot all month
for month in range(12):

y = []
# y-values
y_values = list(out_dict_average[daytype][month].values())
plt.plot(x_values, y_values, label = 'Month %s'%month)

for row in out_array[0][0]: # daytype = 0, January = 0
y.append(list(row))
plt.xlabel("hours")
plt.ylabel("percentage of daily demand")
plt.title("Plotting all month for the daytype {}".format(daytype))
plt.legend()
plt.show()

x = range(len(y))
# --- if individual days are plotted
if plot_average == False:
month = 3

print("Number of lines: " + str(len(y)))
# y-values
y_values = list(out_dict_not_average[daytype][month].values()) #) # daytype = 0, January = 0

for i in range(len(y)):
plt.plot(x, [pt[i] for pt in y])
plt.plot(x_values, y_values)
plt.xlabel("hours")
plt.ylabel("percentage of daily demand")
plt.title("Plotting the Month of {} for the daytype {}".format(month, daytype))
plt.show()

plt.show()
print("Finished loead profiles generator")

0 comments on commit a2f5c86

Please sign in to comment.