a python script to analyze cleaning effectiveness of filter cleaning using ilo sensor data.

ilo_cleaning_analyze.py

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from datetime import datetime
import pytz

# Read the CSV file
df = pd.read_csv('~/Downloads/history.csv')

# Convert last_changed to datetime (keeping the timezone information)
df['last_changed'] = pd.to_datetime(df['last_changed'])

# Create separate dataframes for each sensor
fan_df = df[df['entity_id'] == 'sensor.hp_ilo_fan_1_speed'].copy()
cpu_df = df[df['entity_id'] == 'sensor.hp_ilo_cpu_temperature'].copy()
inlet_df = df[df['entity_id'] == 'sensor.hp_ilo_inlet_ambient'].copy()
exhaust_df = df[df['entity_id'] == 'sensor.hp_ilo_system_exhaust_temperature'].copy()

# Convert state to numeric
fan_df['state'] = pd.to_numeric(fan_df['state'])
cpu_df['state'] = pd.to_numeric(cpu_df['state'])
inlet_df['state'] = pd.to_numeric(inlet_df['state'])
exhaust_df['state'] = pd.to_numeric(exhaust_df['state'])

# Define cleaning time (with UTC timezone)
cleaning_time = pd.to_datetime('2025-03-28 20:00:00').tz_localize('Asia/Shanghai')

# Rename state columns and drop unnecessary columns
fan_df = fan_df[['last_changed', 'state']].rename(columns={'state': 'fan_speed'})
inlet_df = inlet_df[['last_changed', 'state']].rename(columns={'state': 'inlet_temp'})
exhaust_df = exhaust_df[['last_changed', 'state']].rename(columns={'state': 'exhaust_temp'})
cpu_df = cpu_df[['last_changed', 'state']].rename(columns={'state': 'cpu_temp'})

# Merge all data
# First merge fan and inlet
merged_df = pd.merge_asof(
    fan_df.sort_values('last_changed'),
    inlet_df.sort_values('last_changed'),
    on='last_changed',
    tolerance=pd.Timedelta('5min')
)

# Then merge exhaust
merged_df = pd.merge_asof(
    merged_df,
    exhaust_df.sort_values('last_changed'),
    on='last_changed',
    tolerance=pd.Timedelta('5min')
)

# Then merge CPU
merged_df = pd.merge_asof(
    merged_df,
    cpu_df.sort_values('last_changed'),
    on='last_changed',
    tolerance=pd.Timedelta('5min')
)

# Add period column after merging
merged_df['period'] = merged_df['last_changed'].apply(
    lambda x: 'After Cleaning' if x > cleaning_time else 'Before Cleaning'
)

# Print column names for debugging
print("Available columns:", merged_df.columns.tolist())

# Calculate temperature difference and efficiency metrics
merged_df['delta_T'] = merged_df['exhaust_temp'] - merged_df['inlet_temp']
merged_df['cooling_efficiency'] = merged_df['delta_T'] / merged_df['fan_speed']

# Create plots
fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(20, 15))

# Plot 1: Time series of all metrics
ax1.plot(merged_df['last_changed'], merged_df['fan_speed'], label='Fan Speed')
ax1.plot(merged_df['last_changed'], merged_df['inlet_temp'], label='Inlet Temp')
ax1.plot(merged_df['last_changed'], merged_df['exhaust_temp'], label='Exhaust Temp')
ax1.plot(merged_df['last_changed'], merged_df['cpu_temp'], label='CPU Temp')
ax1.axvline(x=cleaning_time, color='r', linestyle='--', label='Cleaning Time')
ax1.legend()
ax1.set_title('Time Series of All Metrics')
ax1.tick_params(axis='x', rotation=45)
ax1.grid(True)

# Plot 2: Fan Speed vs Inlet Temperature
sns.scatterplot(data=merged_df,
                x='inlet_temp',
                y='fan_speed',
                hue='period',
                alpha=0.6,
                ax=ax2)
ax2.set_title('Fan Speed vs Inlet Temperature')
ax2.grid(True)

# Plot 3: Temperature Difference vs Fan Speed
sns.scatterplot(data=merged_df,
                x='fan_speed',
                y='delta_T',
                hue='period',
                alpha=0.6,
                ax=ax3)
ax3.set_title('Temperature Difference vs Fan Speed')
ax3.grid(True)

# Plot 4: Cooling Efficiency Over Time
scatter = ax4.scatter(merged_df['last_changed'], merged_df['cooling_efficiency'],
                     c=merged_df['cpu_temp'], cmap='viridis', alpha=0.6)
ax4.axvline(x=cleaning_time, color='r', linestyle='--', label='Cleaning Time')
ax4.set_title('Cooling Efficiency Over Time (color = CPU Temperature)')
ax4.tick_params(axis='x', rotation=45)
ax4.grid(True)
plt.colorbar(scatter, ax=ax4, label='CPU Temperature')

plt.tight_layout()
plt.show()

# Print comprehensive statistics
print("\nComprehensive Statistics for both periods:")
for period in ['Before Cleaning', 'After Cleaning']:
    period_data = merged_df[merged_df['period'] == period]
    if len(period_data) > 0:
        print(f"\n{period}:")
        print(f"Number of data points: {len(period_data)}")
        print(f"Average fan speed: {period_data['fan_speed'].mean():.2f}")
        print(f"Average inlet temperature: {period_data['inlet_temp'].mean():.2f}")
        print(f"Average exhaust temperature: {period_data['exhaust_temp'].mean():.2f}")
        print(f"Average CPU temperature: {period_data['cpu_temp'].mean():.2f}")
        print(f"Average temperature difference (Exhaust - Inlet): {period_data['delta_T'].mean():.2f}")
        print(f"Average cooling efficiency (ΔT/Fan Speed): {period_data['cooling_efficiency'].mean():.3f}")
        print("\nTemperature Ranges:")
        print(f"Inlet: {period_data['inlet_temp'].min():.1f} - {period_data['inlet_temp'].max():.1f}")
        print(f"Exhaust: {period_data['exhaust_temp'].min():.1f} - {period_data['exhaust_temp'].max():.1f}")
        print(f"CPU: {period_data['cpu_temp'].min():.1f} - {period_data['cpu_temp'].max():.1f}")