mirror of
https://github.com/Akkudoktor-EOS/EOS.git
synced 2025-04-19 00:45:22 +00:00
48 Stunden Predcition & Optimierung
Ein paar Zeitfunktionen korrigiert (24h / 48h) Strompreis Cache stündlich leeren Strompreis bei nur 24h Daten, wird verdoppelt (Prognose fehlt noch)
This commit is contained in:
Bla Bla
parent
817e5e3cc9
commit
95b0ec5664
@ -13,6 +13,10 @@ Dieses Projekt bietet eine umfassende Lösung zur Simulation und Optimierung ein
|
||||
- `Simulation:` Lastverteilung 1h Werte -> Minuten (Tabelle)
|
||||
- `Dynamische Lasten:` z.B. eine Spülmaschine, welche gesteuert werdeb jabb,
|
||||
- `Simulation:` AC Chargen möglich
|
||||
- `Optimierung:` E-Auto Akku voll = in der 0/1 Liste keine Möglichkeit mehr auf 1 (aktuell ist der Optimierung das egalm ändert ja nichts) Optimierungsparameter reduzieren
|
||||
- `Backend:` Visual Cleaner (z.B. E-Auto Akku = 100%, dann sollte die Lademöglichkeit auf 0 stehen. Zumindest bei der Ausgabe sollte das "sauber" sein)
|
||||
- `Backend:` Cache regelmäßig leeren können (API)
|
||||
|
||||
|
||||
|
||||
## Installation
|
||||
|
||||
@ -27,7 +27,7 @@ import os
|
||||
app = Flask(__name__)
|
||||
|
||||
|
||||
opt_class = optimization_problem(prediction_hours=24, strafe=10)
|
||||
opt_class = optimization_problem(prediction_hours=48, strafe=10)
|
||||
soc_predictor = BatterySocPredictor.load_model('battery_model.pkl')
|
||||
|
||||
|
||||
@ -69,7 +69,7 @@ def flask_optimize():
|
||||
return jsonify({"error": f"Fehlender Parameter: {p}"}), 400
|
||||
|
||||
# Simulation durchführen
|
||||
ergebnis = opt_class.optimierung_ems(parameter=parameter, start_hour=datetime.now().hour)
|
||||
ergebnis = opt_class.optimierung_ems(parameter=parameter, start_hour=datetime.now().hour) # , startdate = datetime.now().date() - timedelta(days = 1)
|
||||
|
||||
return jsonify(ergebnis)
|
||||
|
||||
|
||||
@ -98,32 +98,7 @@ class EnergieManagementSystem:
|
||||
verluste_wh_pro_stunde[-1] += verluste
|
||||
eigenverbrauch_wh_pro_stunde.append(eigenverbrauch)
|
||||
|
||||
|
||||
|
||||
# Mehr erzeugt als verbraucht
|
||||
# if erzeugung > verbrauch:
|
||||
# überschuss = erzeugung - verbrauch
|
||||
# #geladene_energie = min(überschuss, self.akku.kapazitaet_wh - self.akku.soc_wh)
|
||||
# geladene_energie, verluste_laden_akku = self.akku.energie_laden(überschuss, stunde)
|
||||
# verluste_wh_pro_stunde[-1] += verluste_laden_akku
|
||||
# #print("verluste_laden_akku:",verluste_laden_akku)
|
||||
# netzeinspeisung_wh_pro_stunde.append(überschuss - geladene_energie-verluste_laden_akku)
|
||||
# eigenverbrauch_wh_pro_stunde.append(verbrauch)
|
||||
# stündliche_einnahmen_euro = (überschuss - geladene_energie-verluste_laden_akku) * self.einspeiseverguetung_euro_pro_wh[stunde]
|
||||
# #print(überschuss," ", geladene_energie," ",verluste_laden_akku)
|
||||
# netzbezug_wh_pro_stunde.append(0.0)
|
||||
# # Noch Netzbezug nötig
|
||||
# else:
|
||||
# netzeinspeisung_wh_pro_stunde.append(0.0)
|
||||
# benötigte_energie = verbrauch - erzeugung
|
||||
# aus_akku, akku_entladeverluste = self.akku.energie_abgeben(benötigte_energie, stunde)
|
||||
# verluste_wh_pro_stunde[-1] += akku_entladeverluste
|
||||
# #print("akku_entladeverluste:",akku_entladeverluste)
|
||||
|
||||
# stündlicher_netzbezug_wh = benötigte_energie - aus_akku
|
||||
# netzbezug_wh_pro_stunde.append(stündlicher_netzbezug_wh)
|
||||
# eigenverbrauch_wh_pro_stunde.append(erzeugung+aus_akku)
|
||||
# stündliche_kosten_euro = stündlicher_netzbezug_wh * strompreis
|
||||
|
||||
|
||||
if self.eauto:
|
||||
eauto_soc_pro_stunde.append(eauto_soc)
|
||||
|
||||
@ -67,12 +67,19 @@ class LoadForecast:
|
||||
|
||||
# Beachten, dass bei Schaltjahren der Tag des Jahres angepasst werden muss
|
||||
stats_for_range = self.data_year_energy[start_day_of_year:end_day_of_year] # -1 da die Indizierung bei 0 beginnt
|
||||
# print(start_day_of_year,"-",end_day_of_year)
|
||||
# print(stats_for_range.shape)
|
||||
stats_for_range =stats_for_range.swapaxes(1, 0)
|
||||
|
||||
stats_for_range = stats_for_range.reshape(stats_for_range.shape[0],-1)
|
||||
# print(stats_for_range.shape)
|
||||
# print(stats_for_range)
|
||||
# print()
|
||||
# print(stats_for_range)
|
||||
# print(start_day_of_year, " ",end_day_of_year)
|
||||
# Hier kannst du entscheiden, wie du die Daten über den Zeitraum aggregieren möchtest
|
||||
# Zum Beispiel könntest du Mittelwerte, Summen oder andere Statistiken über diesen Zeitraum berechnen
|
||||
return stats_for_range[0]
|
||||
return stats_for_range
|
||||
|
||||
|
||||
|
||||
|
||||
@ -70,7 +70,7 @@ class optimization_problem:
|
||||
self.toolbox.register("individual", create_individual)#tools.initCycle, creator.Individual, (self.toolbox.attr_bool,self.toolbox.attr_bool), n=self.prediction_hours+1)
|
||||
self.toolbox.register("population", tools.initRepeat, list, self.toolbox.individual)
|
||||
self.toolbox.register("mate", tools.cxTwoPoint)
|
||||
self.toolbox.register("mutate", tools.mutFlipBit, indpb=0.05)
|
||||
self.toolbox.register("mutate", tools.mutFlipBit, indpb=0.1)
|
||||
self.toolbox.register("select", tools.selTournament, tournsize=3)
|
||||
|
||||
def evaluate_inner(self,individual, ems,start_hour):
|
||||
@ -131,7 +131,7 @@ class optimization_problem:
|
||||
|
||||
# Genetischer Algorithmus
|
||||
def optimize(self,start_solution=None):
|
||||
population = self.toolbox.population(n=1000)
|
||||
population = self.toolbox.population(n=400)
|
||||
hof = tools.HallOfFame(1)
|
||||
|
||||
stats = tools.Statistics(lambda ind: ind.fitness.values)
|
||||
@ -145,7 +145,7 @@ class optimization_problem:
|
||||
population.insert(0, creator.Individual(start_solution))
|
||||
|
||||
#algorithms.eaMuPlusLambda(population, self.toolbox, 100, 200, cxpb=0.2, mutpb=0.2, ngen=500, stats=stats, halloffame=hof, verbose=True)
|
||||
algorithms.eaSimple(population, self.toolbox, cxpb=0.2, mutpb=0.2, ngen=200, stats=stats, halloffame=hof, verbose=True)
|
||||
algorithms.eaSimple(population, self.toolbox, cxpb=0.1, mutpb=0.1, ngen=400, stats=stats, halloffame=hof, verbose=True)
|
||||
|
||||
member = {"bilanz":[],"verluste":[],"nebenbedingung":[]}
|
||||
for ind in population:
|
||||
@ -159,14 +159,19 @@ class optimization_problem:
|
||||
return hof[0], member
|
||||
|
||||
|
||||
def optimierung_ems(self,parameter=None, start_hour=None,worst_case=False):
|
||||
def optimierung_ems(self,parameter=None, start_hour=None,worst_case=False, startdate=None):
|
||||
|
||||
|
||||
############
|
||||
# Parameter
|
||||
############
|
||||
date = (datetime.now().date() + timedelta(hours = self.prediction_hours)).strftime("%Y-%m-%d")
|
||||
date_now = datetime.now().strftime("%Y-%m-%d")
|
||||
if startdate == None:
|
||||
date = (datetime.now().date() + timedelta(hours = self.prediction_hours)).strftime("%Y-%m-%d")
|
||||
date_now = datetime.now().strftime("%Y-%m-%d")
|
||||
else:
|
||||
date = (startdate + timedelta(hours = self.prediction_hours)).strftime("%Y-%m-%d")
|
||||
date_now = startdate.strftime("%Y-%m-%d")
|
||||
#print("Start_date:",date_now)
|
||||
|
||||
akku_size = parameter['pv_akku_cap'] # Wh
|
||||
year_energy = parameter['year_energy'] #2000*1000 #Wh
|
||||
@ -187,7 +192,7 @@ class optimization_problem:
|
||||
eauto.set_charge_per_hour(laden_moeglich)
|
||||
min_soc_eauto = parameter['eauto_min_soc']
|
||||
start_params = parameter['start_solution']
|
||||
gesamtlast = Gesamtlast()
|
||||
gesamtlast = Gesamtlast(prediction_hours=self.prediction_hours)
|
||||
|
||||
###############
|
||||
# spuelmaschine
|
||||
@ -205,7 +210,12 @@ class optimization_problem:
|
||||
###############
|
||||
lf = LoadForecast(filepath=r'load_profiles.npz', year_energy=year_energy)
|
||||
#leistung_haushalt = lf.get_daily_stats(date)[0,...] # Datum anpassen
|
||||
leistung_haushalt = lf.get_stats_for_date_range(date_now,date)[0,...].flatten()
|
||||
|
||||
leistung_haushalt = lf.get_stats_for_date_range(date_now,date)[0] # Nur Erwartungswert!
|
||||
|
||||
|
||||
|
||||
|
||||
gesamtlast.hinzufuegen("Haushalt", leistung_haushalt)
|
||||
|
||||
###############
|
||||
@ -229,8 +239,11 @@ class optimization_problem:
|
||||
###############
|
||||
filepath = os.path.join (r'test_data', r'strompreise_akkudokAPI.json') # Pfad zur JSON-Datei anpassen
|
||||
#price_forecast = HourlyElectricityPriceForecast(source=filepath)
|
||||
price_forecast = HourlyElectricityPriceForecast(source="https://api.akkudoktor.net/prices?start="+date_now+"&end="+date+"")
|
||||
price_forecast = HourlyElectricityPriceForecast(source="https://api.akkudoktor.net/prices?start="+date_now+"&end="+date+"", prediction_hours=self.prediction_hours)
|
||||
specific_date_prices = price_forecast.get_price_for_daterange(date_now,date)
|
||||
#print(price_forecast)
|
||||
print(specific_date_prices)
|
||||
#print("https://api.akkudoktor.net/prices?start="+date_now+"&end="+date)
|
||||
|
||||
###############
|
||||
# WP
|
||||
|
||||
@ -14,20 +14,34 @@ utc_time = utc_time.replace(tzinfo=pytz.utc)
|
||||
local_time = utc_time.astimezone(pytz.timezone('Europe/Berlin'))
|
||||
print(local_time)
|
||||
|
||||
def repeat_to_shape(array, target_shape):
|
||||
# Prüfen , ob das Array in die Zielgröße passt
|
||||
if len(target_shape) != array.ndim:
|
||||
raise ValueError("Array and target shape must have the same number of dimensions")
|
||||
|
||||
# die Anzahl der Wiederholungen pro Dimension
|
||||
repeats = tuple(target_shape[i] // array.shape[i] for i in range(array.ndim))
|
||||
|
||||
# np.tile, um das Array zu erweitern
|
||||
expanded_array = np.tile(array, repeats)
|
||||
return expanded_array
|
||||
|
||||
|
||||
|
||||
class HourlyElectricityPriceForecast:
|
||||
def __init__(self, source, cache_dir='cache', abgaben=0.00019):
|
||||
def __init__(self, source, cache_dir='cache', abgaben=0.000, prediction_hours=24): #228
|
||||
self.cache_dir = cache_dir
|
||||
if not os.path.exists(self.cache_dir):
|
||||
os.makedirs(self.cache_dir)
|
||||
self.cache_time_file = os.path.join(self.cache_dir, 'cache_timestamp.txt')
|
||||
self.prices = self.load_data(source)
|
||||
self.abgaben = abgaben
|
||||
self.prediction_hours = prediction_hours
|
||||
|
||||
def load_data(self, source):
|
||||
cache_filename = self.get_cache_filename(source)
|
||||
if source.startswith('http'):
|
||||
cache_filename = self.get_cache_filename(source)
|
||||
if os.path.exists(cache_filename):
|
||||
if os.path.exists(cache_filename) and not self.is_cache_expired():
|
||||
print("Lade Daten aus dem Cache...")
|
||||
with open(cache_filename, 'r') as file:
|
||||
data = json.load(file)
|
||||
@ -38,6 +52,7 @@ class HourlyElectricityPriceForecast:
|
||||
data = response.json()
|
||||
with open(cache_filename, 'w') as file:
|
||||
json.dump(data, file)
|
||||
self.update_cache_timestamp()
|
||||
else:
|
||||
raise Exception(f"Fehler beim Abrufen der Daten: {response.status_code}")
|
||||
else:
|
||||
@ -49,6 +64,18 @@ class HourlyElectricityPriceForecast:
|
||||
hash_object = hashlib.sha256(url.encode())
|
||||
hex_dig = hash_object.hexdigest()
|
||||
return os.path.join(self.cache_dir, f"cache_{hex_dig}.json")
|
||||
|
||||
def is_cache_expired(self):
|
||||
if not os.path.exists(self.cache_time_file):
|
||||
return True
|
||||
with open(self.cache_time_file, 'r') as file:
|
||||
timestamp_str = file.read()
|
||||
last_cache_time = datetime.strptime(timestamp_str, '%Y-%m-%d %H:%M:%S')
|
||||
return datetime.now() - last_cache_time > timedelta(hours=1)
|
||||
|
||||
def update_cache_timestamp(self):
|
||||
with open(self.cache_time_file, 'w') as file:
|
||||
file.write(datetime.now().strftime('%Y-%m-%d %H:%M:%S'))
|
||||
|
||||
|
||||
|
||||
@ -68,13 +95,17 @@ class HourlyElectricityPriceForecast:
|
||||
|
||||
# Extrahieren aller Preise für das spezifizierte Datum
|
||||
date_prices = [entry["marketpriceEurocentPerKWh"]+self.abgaben for entry in self.prices if date_str in entry['end']]
|
||||
print("getPRice:",len(date_prices))
|
||||
|
||||
# Hinzufügen des letzten Preises des vorherigen Tages am Anfang der Liste
|
||||
date_prices.insert(0, last_price_of_previous_day)
|
||||
if len(date_prices) == 23:
|
||||
date_prices.insert(0, last_price_of_previous_day)
|
||||
|
||||
return np.array(date_prices)/(1000.0*100.0) + self.abgaben
|
||||
|
||||
def get_price_for_daterange(self, start_date_str, end_date_str):
|
||||
print(start_date_str)
|
||||
print(end_date_str)
|
||||
"""Gibt alle Preise zwischen dem Start- und Enddatum zurück."""
|
||||
start_date_utc = datetime.strptime(start_date_str, "%Y-%m-%d").replace(tzinfo=pytz.utc)
|
||||
end_date_utc = datetime.strptime(end_date_str, "%Y-%m-%d").replace(tzinfo=pytz.utc)
|
||||
@ -84,13 +115,16 @@ class HourlyElectricityPriceForecast:
|
||||
|
||||
price_list = []
|
||||
|
||||
while start_date <= end_date:
|
||||
while start_date < end_date:
|
||||
date_str = start_date.strftime("%Y-%m-%d")
|
||||
daily_prices = self.get_price_for_date(date_str)
|
||||
#print(len(self.get_price_for_date(date_str)))
|
||||
print(date_str," ",daily_prices)
|
||||
print(len(self.get_price_for_date(date_str)))
|
||||
|
||||
if daily_prices.size ==24:
|
||||
price_list.extend(daily_prices)
|
||||
start_date += timedelta(days=1)
|
||||
|
||||
return np.array(price_list)
|
||||
|
||||
price_list = repeat_to_shape(np.array(price_list),(self.prediction_hours,))
|
||||
|
||||
return price_list
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user