450 lines
17 KiB
Python
450 lines
17 KiB
Python
# Author: David Tschan
|
|
# Institute: Paul-Scherrer Institute, Laboratory for Atmospheric Chemistry, Multiphase Chemistry Group,
|
|
# Forschungsstrasse 111, 5232 Villigen PSI, Switzerland
|
|
# Role: Master Student
|
|
# License: n/a
|
|
# Date: 2026
|
|
|
|
### ----------------------------------------------------------------------------------------------------------------------------------- ###
|
|
### SUBMODULE DOCSTRING
|
|
|
|
"""The docstring for this submodule
|
|
"""
|
|
|
|
### ----------------------------------------------------------------------------------------------------------------------------------- ###
|
|
### IMPORTS AND CONFIGURATIONS
|
|
|
|
import numpy as np
|
|
import pandas as pd
|
|
import math
|
|
import warnings
|
|
|
|
### ----------------------------------------------------------------------------------------------------------------------------------- ###
|
|
### OFR.py
|
|
|
|
class Distribution():
|
|
|
|
r"""Object representing a distribution.
|
|
|
|
Description
|
|
-----------
|
|
Bla Bla Bla
|
|
|
|
Attributes
|
|
----------
|
|
var: np.ndarray
|
|
A 1D array containing the (independent) variable, for instance the diameter $D$.
|
|
fct: np.ndarray
|
|
rA 1D array containing the dependent functional value, for instance count density $\frac{d\log_{10}N}{d\log_{10}D}$
|
|
"""
|
|
|
|
def __init__(self, variable_array, function_array, form = 'log10'):
|
|
|
|
self.var = variable_array
|
|
self.fct = function_array
|
|
|
|
self.__form = form
|
|
|
|
def transform(self, to = 'linear'):
|
|
|
|
"""A method to transform distributions between different representations
|
|
|
|
Description
|
|
-----------
|
|
Bla bla bla
|
|
|
|
Parameters
|
|
----------
|
|
to: str
|
|
Indicating into which representation the distribution should be transformed. Possible values are: `linear`, `ln` or `log10`
|
|
"""
|
|
|
|
if self.__form == to:
|
|
warnings.warn("Already in desired format. Transformation has no effect.")
|
|
return
|
|
|
|
# TODO: check the transformations!
|
|
# TODO: check the order of the transformation (first variable, then function, or vice versa) to avoid having a mess!
|
|
|
|
match to:
|
|
case 'linear':
|
|
if self.__form == 'ln':
|
|
self.var = np.exp(self.var)
|
|
self.fct = np.divide(self.fct, self.var)
|
|
self.__form = 'linear'
|
|
elif self.__form == 'log10':
|
|
self.var = np.power(10, self.var)
|
|
self.fct = np.divide(self.fct, self.var * np.log(10))
|
|
self.__form = 'linear'
|
|
case 'ln':
|
|
if self.__form == 'linear':
|
|
self.var = np.log(self.var)
|
|
self.fct = np.multiply(self.var, self.fct)
|
|
self.__form = 'ln'
|
|
elif self.__form == 'log10':
|
|
self.var = np.log(np.power(10, self.var))
|
|
self.fct = np.divide(self.fct, np.log(10))
|
|
self.__form = 'ln'
|
|
case 'log10':
|
|
if self.__form == 'linear':
|
|
self.fct = np.multiply(self.var, self.fct * np.log(10))
|
|
self.var = np.log10(self.var)
|
|
self.__form = 'log10'
|
|
elif self.__form == 'ln':
|
|
self.var = np.log10(np.exp(self.var))
|
|
self.fct = np.multiply(self.fct, np.log(10))
|
|
self.__form = 'log10'
|
|
|
|
def update_units(self, new_var_units, new_fct_units):
|
|
|
|
"""A method to update the units of the distribution
|
|
|
|
Description
|
|
-----------
|
|
Bla Bla Bla
|
|
|
|
Parameters
|
|
----------
|
|
new_var_units: float
|
|
The SI factor used to transform the independent variable `self.var`
|
|
new_fct_units: float
|
|
The SI factor used to transform the dependent variable `self.fct`
|
|
"""
|
|
|
|
old_from = self.__form
|
|
self.transform(to='linear')
|
|
|
|
self.var = self.var * new_var_units
|
|
self.fct = self.fct * new_fct_units
|
|
|
|
self.transform(to=old_from)
|
|
|
|
def integrate(self) -> np.float64:
|
|
|
|
"""A method to integrate over the spectrum $dN/d f(D)$
|
|
|
|
Description
|
|
-----------
|
|
Bla Bla Bla
|
|
"""
|
|
|
|
return np.trapezoid(y = self.fct, x = self.var)
|
|
|
|
def median():
|
|
# TODO: implement that method
|
|
pass
|
|
|
|
def mode():
|
|
# TODO: implement that method
|
|
pass
|
|
|
|
def mean():
|
|
# TODO: implement that method
|
|
pass
|
|
|
|
def variance():
|
|
# TODO: implement that method
|
|
pass
|
|
|
|
class OFRSample:
|
|
|
|
"""Object to represent a single sample from the OFR
|
|
|
|
Description
|
|
-----------
|
|
Bla bla bla
|
|
|
|
Attributes
|
|
----------
|
|
|
|
number: int
|
|
The sample number
|
|
start_time: pd.datetime
|
|
The starting time of the sample recording
|
|
sourcefile: str
|
|
The path to the data file
|
|
headlines: int
|
|
The number of lines in the data file until raw data is reached
|
|
lower_bound: float
|
|
The lower bound diameter (nm) of the SMPS
|
|
upper_bound: float
|
|
The upper bound diameter (nm) of the SMPS
|
|
scan_up: float
|
|
The time (s) to ramp up voltage in SMPS
|
|
scan_down: float
|
|
The time (s) to ramp voltage back down in SMPS
|
|
delta_time: float
|
|
The time interval (s) between two measured data points of the raw count density
|
|
SOA_density: float
|
|
The density (kg/m3) of the SOA produced in the OFR
|
|
"""
|
|
|
|
def __get_data(self) -> np.float64:
|
|
|
|
footer_flag = False
|
|
footer_start_index = 0
|
|
data_flag = False
|
|
data_lines = []
|
|
|
|
midpoints = []
|
|
counts = []
|
|
|
|
date_line = None
|
|
time_line = None
|
|
|
|
with open(self.source_file, 'r', errors = 'ignore') as file:
|
|
for i, line in enumerate(file.readlines()):
|
|
if i == 18:
|
|
date_line = line
|
|
continue
|
|
if i == 19:
|
|
time_line = line
|
|
continue
|
|
if i == 20:
|
|
data_flag = True
|
|
continue
|
|
if data_flag == True:
|
|
data_lines.append(line)
|
|
if line.split(self.__data_separator)[0] == 'Scan Up Time(s)':
|
|
data_flag = False
|
|
footer_flag = True
|
|
footer_start_index = i
|
|
self.scan_up_time = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
if footer_flag == False:
|
|
continue
|
|
else:
|
|
if i == footer_start_index + 1:
|
|
self.scan_down_time = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
elif i == footer_start_index + 6:
|
|
self.aerosol_flow = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
elif i == footer_start_index + 11:
|
|
self.lower_diameter_bound = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
elif i == footer_start_index + 12:
|
|
self.upper_diameter_bound = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
elif i == footer_start_index + 13:
|
|
self.SOA_density = float(line.split(self.__data_separator)[1])
|
|
continue
|
|
|
|
self.sampling_time = self.scan_up_time + self.scan_down_time
|
|
data_lines.pop(-1)
|
|
|
|
# obtaining the starting time
|
|
raw_date_items = date_line.split(self.__data_separator)
|
|
raw_time_items = time_line.split(self.__data_separator)
|
|
del raw_date_items[0]
|
|
del raw_time_items[0]
|
|
|
|
raw_date = raw_date_items[(self.sample_number - 1) * 2].splitlines()[0]
|
|
raw_time = raw_time_items[(self.sample_number - 1) * 2].splitlines()[0]
|
|
self.start_time = pd.to_datetime(raw_date + ' ' + raw_time, format=r'%m/%d/%y %H:%M:%S')
|
|
|
|
for line in data_lines:
|
|
items = line.split(self.__data_separator)
|
|
local_diameter_midpoint = float(items[0])
|
|
if self.lower_diameter_bound <= local_diameter_midpoint and local_diameter_midpoint <= self.upper_diameter_bound:
|
|
midpoints.append(local_diameter_midpoint)
|
|
items.pop(0)
|
|
counts.append(float(items[(self.sample_number - 1) * 2]))
|
|
else:
|
|
continue
|
|
|
|
self.aerosol_flow = self.aerosol_flow * 6E-4
|
|
self.SOA_density = self.SOA_density * 1E3
|
|
self.lower_diameter_bound = self.lower_diameter_bound * 1E-9
|
|
self.upper_diameter_bound = self.upper_diameter_bound * 1E-9
|
|
|
|
# Input data from file is in (nm)
|
|
return [np.array(midpoints), np.array(counts)]
|
|
|
|
def __init__(self, source_file, sample_number, data_separator = ','):
|
|
|
|
## VALUES READ IN FROM DISK
|
|
#--------------------------
|
|
|
|
# Sample information
|
|
self.source_file = source_file # where the data is stored, i.e. "C:/Users/tschan_d/..."
|
|
self.sample_number = sample_number # sample number, i.e. 33
|
|
self.start_time = None
|
|
self.__data_separator = data_separator
|
|
|
|
self.scan_up_time = None # Time to scan up the voltage range in DMA, in (s)
|
|
self.scan_down_time = None # Time to scan down the voltage range in DMA, in(s)
|
|
self.sampling_time = None # Total time to scan voltage up and back down in DMA, in (s)
|
|
|
|
self.lower_diameter_bound = None # Lower Diameter bound of CPC, in (nm)
|
|
self.upper_diameter_bound = None # Upper Diameter bound of CPC, in (nm)
|
|
self.aerosol_flow = None # Aerosol flow, in (L/min)
|
|
self.SOA_density = None # Expected density of SOA, in (g/cm3)
|
|
|
|
# Distributions of the relevant quantities (all relative to nanometer)
|
|
data = self.__get_data()
|
|
|
|
# Getting Distribution objects
|
|
self.Count = Distribution(variable_array = np.log10(data[0]), function_array = data[1], form='log10')
|
|
self.Count.update_units(new_fct_units=1E15, new_var_units=1E-9)
|
|
|
|
self.Length = Distribution(variable_array = np.log10(data[0]), function_array = data[0]*data[1], form = 'log10')
|
|
self.Length.update_units(new_fct_units=1E6, new_var_units=1E-9)
|
|
|
|
self.Surface = Distribution(variable_array = np.log(data[0]), function_array = ((math.pi)* data[0] ** 2) * data[1], form = 'log10')
|
|
self.Surface.update_units(new_fct_units=1E-3, new_var_units=1E-9)
|
|
|
|
self.Volume = Distribution(variable_array = np.log10(data[0]),function_array = ((math.pi / 6) * data[0] ** 3) * data[1], form = 'log10')
|
|
self.Volume.update_units(new_fct_units=1E-12, new_var_units=1E-9)
|
|
|
|
self.Mass = Distribution(variable_array = self.Volume.var, function_array = self.Volume.fct * self.SOA_density, form='log10')
|
|
|
|
class OFRAnalyser:
|
|
|
|
"""An object representing an OFR experiment
|
|
|
|
Description
|
|
-----------
|
|
Bla Bla Bla
|
|
|
|
Attributes
|
|
----------
|
|
data_source_file: str
|
|
A string pointing to the data source file
|
|
flow_source_file: str
|
|
A string pointing to the flow data source file
|
|
sample_number_list: [int]
|
|
A list of integers indicating which samples to consider
|
|
on_time_interval: [pd.Timestamp]
|
|
A list with two elements containing the pd.Timestamp objects when the VOC flow was switched on and off, respectively.
|
|
max_samples: int
|
|
The highest sample available in the data
|
|
data: pandas.DataFrame
|
|
A dataframe containing all the data
|
|
"""
|
|
|
|
def __init__(self, data_source_file, flow_source_file, smpl_number_list, on_time_interval, data_separator = ',') -> None:
|
|
|
|
self.data_source_file = data_source_file
|
|
self.flow_source_file = flow_source_file
|
|
self.smpl_number_list = smpl_number_list
|
|
self.on_time_interval = on_time_interval
|
|
self.__data_separator = data_separator
|
|
|
|
|
|
with open(self.data_source_file, 'r', errors = 'ignore') as file:
|
|
for i, line in enumerate(file.readlines()):
|
|
if i == 17:
|
|
items = line.split(',')
|
|
max_sample_number = int(len(items) / 2)
|
|
else:
|
|
continue
|
|
|
|
self.max_samples = max_sample_number
|
|
self.data = None
|
|
|
|
def __smpl_number_list_check(self, smpl_number_list) -> None:
|
|
|
|
# TODO: implement a similar check for the flow data file!
|
|
|
|
last_sample = smpl_number_list[-1]
|
|
|
|
if last_sample > self.max_samples:
|
|
raise IndexError("The last sample(s) are out of range of the number of recorded samples")
|
|
|
|
def GetData(self) -> None:
|
|
|
|
"""Method to get the data
|
|
|
|
Description
|
|
-----------
|
|
Bla Bla Bla
|
|
"""
|
|
|
|
# TODO make this work also for a specific sample!
|
|
# TODO convert everything is SI units, making live a lot easier...
|
|
|
|
self.__smpl_number_list_check(self.smpl_number_list)
|
|
|
|
# Quantities of interest:
|
|
numbers_of_sample = [] # (-)
|
|
starting_times = [] # (timestamp)
|
|
sampling_times = [] # (s)
|
|
number_concentrations = [] # (#/m^3)
|
|
length_concentrations = [] # (m/m^3)
|
|
surface_concentrations = [] # (m^2/m^3)
|
|
volume_concentrations = [] # (m^3/m^3)
|
|
aerosol_densities = [] # (kg/m^3)
|
|
mass_concentrations = [] # (kg/m^3)
|
|
filter_flows = [] # (m^3/s)
|
|
mass_production_rate = None # (kg/s)
|
|
mass_produced = None # (kg)
|
|
cumulative_mass_produced = None # (kg)
|
|
|
|
for sample_number in self.smpl_number_list:
|
|
sample = OFRSample(source_file=self.data_source_file, sample_number = sample_number, data_separator = self.__data_separator)
|
|
|
|
numbers_of_sample.append(sample_number)
|
|
starting_times.append(sample.start_time)
|
|
|
|
number_concentrations.append(sample.Count.integrate())
|
|
length_concentrations.append(sample.Length.integrate())
|
|
surface_concentrations.append(sample.Surface.integrate())
|
|
volume_concentrations.append(sample.Volume.integrate())
|
|
aerosol_densities.append(sample.SOA_density)
|
|
mass_concentrations.append(sample.Mass.integrate())
|
|
|
|
sampling_times.append(sample.sampling_time)
|
|
|
|
out_flag = False
|
|
|
|
with open(self.flow_source_file, 'r', errors='ignore') as file:
|
|
|
|
j = 1
|
|
|
|
for line_counter, line in enumerate(file.readlines()):
|
|
|
|
if out_flag == True:
|
|
continue
|
|
|
|
if line_counter >= 6:
|
|
|
|
if line.startswith('-'):
|
|
out_flag = True
|
|
continue
|
|
|
|
i = 0
|
|
items = line.split()
|
|
|
|
while i < int(items[-1]):
|
|
|
|
if j in self.smpl_number_list:
|
|
filter_flows.append(float(items[0]) * (1E-3) / 60)
|
|
|
|
i += 1
|
|
j += 1
|
|
|
|
else:
|
|
continue
|
|
|
|
mass_production_rate = [(mass * flow) for mass, flow in zip(mass_concentrations, filter_flows)]
|
|
mass_produced = [(mpr * duration) for mpr, duration in zip(mass_production_rate, sampling_times)]
|
|
cumulative_mass_produced = np.cumsum(np.array(mass_produced)).tolist()
|
|
|
|
self.data = pd.DataFrame({'Sample (-)': numbers_of_sample,
|
|
'Starting Time (-)': starting_times,
|
|
'Sampling Time (s)': sampling_times,
|
|
'Number (#/m^3)': number_concentrations,
|
|
'Length (m/m^3)': length_concentrations,
|
|
'Surface (m^2/m^3)': surface_concentrations,
|
|
'Volume (m^3/m^3)': volume_concentrations,
|
|
'Density (kg/m^3)': aerosol_densities,
|
|
'Mass (kg/m^3)': mass_concentrations,
|
|
'Filter Flow (m^3/s)': filter_flows,
|
|
'Mass Production Rate (kg/s)': mass_production_rate,
|
|
'Mass Produced (kg)': mass_produced,
|
|
'Cumulative Mass Produced (kg)': cumulative_mass_produced})
|
|
|
|
|
|
### ----------------------------------------------------------------------------------------------------------------------------------- ###
|
|
### |