mirror of
https://github.com/Akkudoktor-EOS/EOS.git
synced 2025-04-19 08:55:15 +00:00
332 lines
11 KiB
Python
332 lines
11 KiB
Python
from datetime import datetime, timedelta
|
|
|
|
import mariadb
|
|
import matplotlib.pyplot as plt
|
|
import numpy as np
|
|
import pandas as pd
|
|
|
|
|
|
class BatteryDataProcessor:
|
|
def __init__(
|
|
self,
|
|
config,
|
|
voltage_high_threshold,
|
|
voltage_low_threshold,
|
|
current_low_threshold,
|
|
gap,
|
|
battery_capacity_ah,
|
|
):
|
|
self.config = config
|
|
self.voltage_high_threshold = voltage_high_threshold
|
|
self.voltage_low_threshold = voltage_low_threshold
|
|
self.current_low_threshold = current_low_threshold
|
|
self.gap = gap
|
|
self.battery_capacity_ah = battery_capacity_ah
|
|
self.conn = None
|
|
self.data = None
|
|
|
|
def connect_db(self):
|
|
self.conn = mariadb.connect(**self.config)
|
|
self.cursor = self.conn.cursor()
|
|
|
|
def disconnect_db(self):
|
|
if self.conn:
|
|
self.cursor.close()
|
|
self.conn.close()
|
|
|
|
def fetch_data(self, start_time):
|
|
query = """
|
|
SELECT timestamp, data, topic
|
|
FROM pip
|
|
WHERE timestamp >= %s AND (topic = 'battery_current' OR topic = 'battery_voltage')
|
|
ORDER BY timestamp
|
|
"""
|
|
self.cursor.execute(query, (start_time,))
|
|
rows = self.cursor.fetchall()
|
|
self.data = pd.DataFrame(rows, columns=["timestamp", "data", "topic"])
|
|
self.data["timestamp"] = pd.to_datetime(self.data["timestamp"])
|
|
self.data["data"] = self.data["data"].astype(float)
|
|
|
|
def process_data(self):
|
|
self.data.drop_duplicates(subset=["timestamp", "topic"], inplace=True)
|
|
|
|
data_pivot = self.data.pivot(index="timestamp", columns="topic", values="data")
|
|
data_pivot = data_pivot.resample("1T").mean().interpolate()
|
|
data_pivot.columns.name = None
|
|
data_pivot.reset_index(inplace=True)
|
|
self.data = data_pivot
|
|
|
|
def group_points(self, df):
|
|
df = df.sort_values("timestamp")
|
|
groups = []
|
|
group = []
|
|
last_time = None
|
|
|
|
for _, row in df.iterrows():
|
|
if last_time is None or (row["timestamp"] - last_time) <= pd.Timedelta(
|
|
minutes=self.gap
|
|
):
|
|
group.append(row)
|
|
else:
|
|
groups.append(group)
|
|
group = [row]
|
|
last_time = row["timestamp"]
|
|
|
|
if group:
|
|
groups.append(group)
|
|
|
|
last_points = [group[-1] for group in groups]
|
|
return last_points
|
|
|
|
def find_soc_points(self):
|
|
condition_soc_100 = (
|
|
self.data["battery_voltage"] >= self.voltage_high_threshold
|
|
) & (self.data["battery_current"].abs() <= self.current_low_threshold)
|
|
condition_soc_0 = (
|
|
self.data["battery_voltage"] <= self.voltage_low_threshold
|
|
) & (self.data["battery_current"].abs() <= self.current_low_threshold)
|
|
|
|
times_soc_100_all = self.data[condition_soc_100][
|
|
["timestamp", "battery_voltage", "battery_current"]
|
|
]
|
|
times_soc_0_all = self.data[condition_soc_0][
|
|
["timestamp", "battery_voltage", "battery_current"]
|
|
]
|
|
|
|
last_points_100 = self.group_points(times_soc_100_all)
|
|
last_points_0 = self.group_points(times_soc_0_all)
|
|
|
|
last_points_100_df = pd.DataFrame(last_points_100)
|
|
last_points_0_df = pd.DataFrame(last_points_0)
|
|
|
|
return last_points_100_df, last_points_0_df
|
|
|
|
def calculate_resetting_soc(self, last_points_100_df, last_points_0_df):
|
|
soc_values = []
|
|
integration_results = []
|
|
reset_points = pd.concat([last_points_100_df, last_points_0_df]).sort_values(
|
|
"timestamp"
|
|
)
|
|
|
|
# Initialisieren der SoC-Liste
|
|
self.data["calculated_soc"] = np.nan
|
|
|
|
for i in range(len(reset_points)):
|
|
start_point = reset_points.iloc[i]
|
|
if i < len(reset_points) - 1:
|
|
end_point = reset_points.iloc[i + 1]
|
|
else:
|
|
end_point = self.data.iloc[
|
|
-1
|
|
] # Verwenden des letzten Datensatzes als Endpunkt
|
|
|
|
if start_point["timestamp"] in last_points_100_df["timestamp"].values:
|
|
initial_soc = 100
|
|
elif start_point["timestamp"] in last_points_0_df["timestamp"].values:
|
|
initial_soc = 0
|
|
|
|
cut_data = self.data[
|
|
(self.data["timestamp"] >= start_point["timestamp"])
|
|
& (self.data["timestamp"] <= end_point["timestamp"])
|
|
].copy()
|
|
cut_data["time_diff_hours"] = (
|
|
cut_data["timestamp"].diff().dt.total_seconds() / 3600
|
|
)
|
|
cut_data.dropna(subset=["time_diff_hours"], inplace=True)
|
|
|
|
calculated_soc = initial_soc
|
|
calculated_soc_list = [calculated_soc]
|
|
integrated_current = 0
|
|
|
|
for j in range(1, len(cut_data)):
|
|
current = cut_data.iloc[j]["battery_current"]
|
|
delta_t = cut_data.iloc[j]["time_diff_hours"]
|
|
delta_soc = (
|
|
(current * delta_t) / self.battery_capacity_ah * 100
|
|
) # Convert to percentage
|
|
|
|
calculated_soc += delta_soc
|
|
calculated_soc = min(max(calculated_soc, 0), 100) # Clip to 0-100%
|
|
calculated_soc_list.append(calculated_soc)
|
|
|
|
# Integration des Stroms aufaddieren
|
|
integrated_current += current * delta_t
|
|
|
|
cut_data["calculated_soc"] = calculated_soc_list
|
|
soc_values.append(cut_data[["timestamp", "calculated_soc"]])
|
|
|
|
integration_results.append(
|
|
{
|
|
"start_time": start_point["timestamp"],
|
|
"end_time": end_point["timestamp"],
|
|
"integrated_current": integrated_current,
|
|
"start_soc": initial_soc,
|
|
"end_soc": calculated_soc_list[-1],
|
|
}
|
|
)
|
|
|
|
soc_df = (
|
|
pd.concat(soc_values)
|
|
.drop_duplicates(subset=["timestamp"])
|
|
.reset_index(drop=True)
|
|
)
|
|
return soc_df, integration_results
|
|
|
|
def calculate_soh(self, integration_results):
|
|
soh_values = []
|
|
|
|
for result in integration_results:
|
|
delta_soc = abs(
|
|
result["start_soc"] - result["end_soc"]
|
|
) # Use the actual change in SoC
|
|
if delta_soc > 0: # Avoid division by zero
|
|
effective_capacity_ah = result["integrated_current"]
|
|
soh = (effective_capacity_ah / self.battery_capacity_ah) * 100
|
|
soh_values.append({"timestamp": result["end_time"], "soh": soh})
|
|
|
|
soh_df = pd.DataFrame(soh_values)
|
|
return soh_df
|
|
|
|
def delete_existing_soc_entries(self, soc_df):
|
|
delete_query = """
|
|
DELETE FROM pip
|
|
WHERE timestamp = %s AND topic = 'calculated_soc'
|
|
"""
|
|
timestamps = [
|
|
(row["timestamp"].strftime("%Y-%m-%d %H:%M:%S"),)
|
|
for _, row in soc_df.iterrows()
|
|
if pd.notna(row["timestamp"])
|
|
]
|
|
|
|
self.cursor.executemany(delete_query, timestamps)
|
|
self.conn.commit()
|
|
|
|
def update_database_with_soc(self, soc_df):
|
|
# Löschen der vorhandenen Einträge mit demselben Topic und Datum
|
|
self.delete_existing_soc_entries(soc_df)
|
|
|
|
# Resample `soc_df` auf 5-Minuten-Intervalle und berechnen des Mittelwerts
|
|
soc_df.set_index("timestamp", inplace=True)
|
|
soc_df_resampled = soc_df.resample("5T").mean().dropna().reset_index()
|
|
# soc_df_resampled['timestamp'] = soc_df_resampled['timestamp'].apply(lambda x: x.strftime('%Y-%m-%d %H:%M:%S'))
|
|
print(soc_df_resampled)
|
|
|
|
# Einfügen der berechneten SoC-Werte in die Datenbank
|
|
insert_query = """
|
|
INSERT INTO pip (timestamp, data, topic)
|
|
VALUES (%s, %s, 'calculated_soc')
|
|
"""
|
|
for _, row in soc_df_resampled.iterrows():
|
|
print(row)
|
|
print(row["timestamp"])
|
|
record = (
|
|
row["timestamp"].strftime("%Y-%m-%d %H:%M:%S"),
|
|
row["calculated_soc"],
|
|
)
|
|
try:
|
|
self.cursor.execute(insert_query, record)
|
|
except mariadb.OperationalError as e:
|
|
print(f"Error inserting record {record}: {e}")
|
|
|
|
self.conn.commit()
|
|
|
|
def plot_data(self, last_points_100_df, last_points_0_df, soc_df):
|
|
plt.figure(figsize=(14, 10))
|
|
|
|
plt.subplot(4, 1, 1)
|
|
plt.plot(
|
|
self.data["timestamp"],
|
|
self.data["battery_voltage"],
|
|
label="Battery Voltage",
|
|
color="blue",
|
|
)
|
|
plt.scatter(
|
|
last_points_100_df["timestamp"],
|
|
last_points_100_df["battery_voltage"],
|
|
color="green",
|
|
marker="o",
|
|
label="100% SoC Points",
|
|
)
|
|
# plt.scatter(last_points_0_df['timestamp'], last_points_0_df['battery_voltage'], color='red', marker='x', label='0% SoC Points')
|
|
plt.xlabel("Timestamp")
|
|
plt.ylabel("Voltage (V)")
|
|
plt.legend()
|
|
plt.title("Battery Voltage over Time")
|
|
|
|
plt.subplot(4, 1, 2)
|
|
plt.plot(
|
|
self.data["timestamp"],
|
|
self.data["battery_current"],
|
|
label="Battery Current",
|
|
color="orange",
|
|
)
|
|
plt.scatter(
|
|
last_points_100_df["timestamp"],
|
|
last_points_100_df["battery_current"],
|
|
color="green",
|
|
marker="o",
|
|
label="100% SoC Points",
|
|
)
|
|
# plt.scatter(last_points_0_df['timestamp'], last_points_0_df['battery_current'], color='red', marker='x', label='0% SoC Points')
|
|
plt.xlabel("Timestamp")
|
|
plt.ylabel("Current (A)")
|
|
plt.legend()
|
|
plt.title("Battery Current over Time")
|
|
|
|
plt.subplot(4, 1, 3)
|
|
plt.plot(
|
|
soc_df["timestamp"], soc_df["calculated_soc"], label="SoC", color="purple"
|
|
)
|
|
plt.xlabel("Timestamp")
|
|
plt.ylabel("SoC (%)")
|
|
plt.legend()
|
|
plt.title("State of Charge (SoC) over Time")
|
|
|
|
# plt.subplot(4, 1, 4)
|
|
# plt.plot(soh_df['timestamp'], soh_df['soh'], label='SoH', color='brown')
|
|
# plt.xlabel('Timestamp')
|
|
# plt.ylabel('SoH (%)')
|
|
# plt.legend()
|
|
# plt.title('State of Health (SoH) over Time')
|
|
|
|
plt.tight_layout()
|
|
plt.show()
|
|
|
|
|
|
if __name__ == "__main__":
|
|
# MariaDB Verbindungsdetails
|
|
config = {}
|
|
|
|
# Parameter festlegen
|
|
voltage_high_threshold = 55.4 # 100% SoC
|
|
voltage_low_threshold = 46.5 # 0% SoC
|
|
current_low_threshold = 2 # Niedriger Strom für beide Zustände
|
|
gap = 30 # Zeitlücke in Minuten zum Gruppieren von Maxima/Minima
|
|
bat_capacity = 33 * 1000 / 48
|
|
|
|
# Zeitpunkt X definieren
|
|
zeitpunkt_x = (datetime.now() - timedelta(weeks=100)).strftime("%Y-%m-%d %H:%M:%S")
|
|
|
|
# BatteryDataProcessor instanziieren und verwenden
|
|
processor = BatteryDataProcessor(
|
|
config,
|
|
voltage_high_threshold,
|
|
voltage_low_threshold,
|
|
current_low_threshold,
|
|
gap,
|
|
bat_capacity,
|
|
)
|
|
processor.connect_db()
|
|
processor.fetch_data(zeitpunkt_x)
|
|
processor.process_data()
|
|
last_points_100_df, last_points_0_df = processor.find_soc_points()
|
|
soc_df, integration_results = processor.calculate_resetting_soc(
|
|
last_points_100_df, last_points_0_df
|
|
)
|
|
# soh_df = processor.calculate_soh(integration_results)
|
|
processor.update_database_with_soc(soc_df)
|
|
|
|
processor.plot_data(last_points_100_df, last_points_0_df, soc_df)
|
|
|
|
processor.disconnect_db()
|