FinalMethods.ipynb

Kernel: Python 3 (Ubuntu Linux)

Infiltration Calculations

Things needed to make calculations more accurate:

• No inf or zeros present in data

• Find true asymtote for CO2_inside (seems to go below 400 "outside ppm")

• Get accurate ETC volume. (used 40,000 ft$^3$ here)

• Isolate Negative slopes

• Determine best days/times to isolate further

VolumetricFlowRate:

$$CFM_{CO_2} = CFM_{allair} * ( \% CO_2 outside - \% CO_2 inside)$$$$VolumetricFlowRate = CFM_{CO_2} / (\% CO_2 outside - \% CO_2 inside)$$

• $\Delta CO_2$ = $CO_2$ cubic feet per minute (cfm)

• Volumetric Flow Rate = Total cubic feet per minute (cfm)

• $CO_2$_outside = proportion (unitless)

• $CO_2$_inside = proportion (unitless)

Infiltration:

$$Infiltration = \frac{\Delta CO2}{(CO2\_outside - CO2\_inside)} * \frac{1}{Volume} * \frac{60 min}{hour}$$

• Infiltration = "air changes" Per Minute ($\large{\frac{1}{min}}$)

• $\Delta CO_2$ = $CO_2$ cubic feet per minute (cfm)

• $CO_2$_outside = proportion (unitless)

• $CO_2$_inside = proportion (unitless)

• Volume = ETC cubic feet ( $ft^{3}$ )

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline

More Data Manipulation

Starting with just ppm and timestamp

Finshing withing percent CO2 inside and outside, as well as cfm of CO2, and volume of ETC.

df1 = pd.read_csv('NetatmoCleanedForInfiltrationAnalysis.csv', parse_dates=True)
df1.drop(df1.columns[[0]], axis =1 , inplace = True)
df1.head()

df1['Time'] = pd.to_datetime(df1.Time)
df1.head()

df1.rename(index=str, columns={"CO2": "ppm_inside"}, inplace = True)
df1.head()

df1['TimeDelta'] = df1['Time'].diff(1)
df1['TimeDelta'] = df1.TimeDelta /  np.timedelta64(1, 'm')
df1['ppm_outside'] = 400
df1['ETC_Volume'] = 4e4
df1['changeinppminside'] = df1['ppm_inside'].diff(1)
df1.head(4)

Cfm = percent change inside * Volume of ETC

df1['percentchangeinside'] = df1['changeinppminside'] / 1e4
df1.head()

df1['CFM_CO2'] = (df1['percentchangeinside'] / df1['TimeDelta'])* df1['ETC_Volume']
df1.head()

df1['percent_Inside'] = df1['ppm_inside'] / 1e4
df1['percent_Outside'] = df1['ppm_outside'] / 1e4
df1.head()

df1.drop(df1.columns[[0, 2, 3, 5, 6,]], axis =1 , inplace = True)
df1.head()

df1.percent_Inside.describe()

count    100985.000000
mean          0.054765
std           0.030451
min           0.020100
25%           0.036200
50%           0.043800
75%           0.061500
max           0.277700
Name: percent_Inside, dtype: float64

df1.to_csv('InfiltrationFINALCleanedData.csv')

df2 = pd.read_csv('InfiltrationFINALCleanedData.csv', parse_dates=True, index_col=0)
df2.iloc[32]

Time               2016-02-19 16:07:00
ETC_Volume                       40000
CFM_CO2                           12.8
percent_Inside                  0.0409
percent_Outside                   0.04
Name: 32, dtype: object

df2 = df2[df2['percent_Inside']>(df2['percent_Outside'] + 0.02)]
df2 = df2[df2['CFM_CO2'] < 0]
df2 = df2.replace([np.inf, -np.inf], np.nan)
df2.dropna()
df2.describe()

df2.head()

df2['CFM_CO2'].hist(bins = 50, range = (-80, 0))

<matplotlib.axes._subplots.AxesSubplot at 0x7f754f36c0f0>

df2['CFM_ETC'] = df2['CFM_CO2'] / (df2['percent_Inside'] - df2['percent_Outside'])
df2.describe()

df2['CFM_ETC'].plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e3131d0>

df2['CFM_ETC'].hist(bins = 10, range = (-1500, -1000))

<matplotlib.axes._subplots.AxesSubplot at 0x7f754df58198>

df2['ACH'] = (df2['CFM_ETC'] / df2['ETC_Volume']) * 60
df2.describe()

#df2['CFM_ETC'].plot.density( range = 200 )

df2['ACH'].hist(bins = 10, range = (-10, 0))

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e50c550>

df2.set_index('Time',inplace=True)
df2.head()

df2['ACH'].plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e8c6c18>

df2.reset_index(inplace = True)
df2.head(1)

df2['Time'] = pd.to_datetime(df2.Time)

df2['Day'] = df2.Time.dt.weekday_name
df2.head()

#df3.Day.value_counts()

InfiltrationData = df2.drop(df2.columns[[1,2,3,4,5]], axis =1 )
InfiltrationData.to_csv('FinalData.csv')

df3 = pd.read_csv('FinalData.csv', index_col = 0,  parse_dates=True)
df3.head()

df3.dtypes

Time     object
ACH     float64
Day      object
dtype: object

df3['Time'] = pd.to_datetime(df2.Time)
#df3['Day'] = pd.to_datetime(df2.Time)

df3.dtypes

Time    datetime64[ns]
ACH            float64
Day             object
dtype: object

df3 = df3[df3['ACH'] < -0.6]
df3 = df3[df3['ACH'] > -5.0]
df3.describe()

df3.head()

Fridays = df3[df3['Day'] == 'Friday']
Fridays.set_index('Time',inplace=True)
Fridays.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e76a668>

Tuesdays = df3[df3['Day'] == 'Tuesday']
Tuesdays.set_index('Time',inplace=True)
Tuesdays.plot()

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e767208>

Fridays.hist(bins = 100, range = (-2 , -0.5))

array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7f754e7f0828>]],
      dtype=object)

Tuesdays.hist(bins = 100, range = (-2 , -0.5))

array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7f754e719c18>]],
      dtype=object)

Tuesdays.plot.density()

<matplotlib.axes._subplots.AxesSubplot at 0x7f754e249b00>

Tuesdays.describe()