experimental-set-up.Rmd

---
title: "Experimental Set Up & Data"
author: "Tina Lasisi"
date: "`r format(Sys.time(), '%B %d, %Y')`"
output:
  workflowr::wflow_html:
    toc: yes
    number_sections: yes
editor_options:
  chunk_output_type: console
---


```{r setup, message=FALSE, warning=FALSE, include=FALSE, paged.print=FALSE}

library(tidyverse)
library(lubridate)
library(janitor)

F = rprojroot::is_rstudio_project$make_fix_file()

knitr::opts_chunk$set(echo = TRUE, include = TRUE, eval = TRUE, warning = FALSE, message = FALSE, fig.retina = 2, fig.width = 8, fig.height = 6, out.width = "100%")
```

## Introduction

In brief, the experiments were carried out in a climate controlled chamber at the University of Loughborough (UK) using a full-body thermal manikin with 20 independently controllable zones. Human hair wigs ranging from straight to tightly curled were used to test the effect of hair on heat gain/loss in a number of conditions.

The experimental variables were:
- head covering (no wig, straight wig, moderately curled wig, tightly curled wig)
- radiation (on vs. off)
- wind speed (0.3m/s, 1m/s, 2.5m/s)
- evaporation (manikin with dry scalp vs. manikin with wet scalp)

## Set-up
Dry heat resistance experiments were carried out from October to November 2018 in a climate controlled chamber with a custom-built wind tunnel (see below).


```{r}
knitr::include_graphics("assets/manikin-chamber-setup.jpeg", error = FALSE)

```

```{r}
knitr::include_graphics("assets/manikin-chamber-setup-diagram.png", error = FALSE)

```


## Data

The raw data from the manikin is structured as follows

```{r}

raw_manikin <- read_csv(F("data/raw_manikin_output/round_1/HighCurv_RadOff_wind0_3 (run with new conditions).csv"), skip = 16) %>% 
  clean_names()

head(raw_manikin)

```

The data can be cleaned using the following pipeline.

```{r}

library(lubridate)
library(fs)
data_dir1 <- F("data/raw_manikin_output/round_1")
data_dir2 <- F("data/raw_manikin_output/round_2")
data_dir3 <- F("data/raw_manikin_output/round_3")
round1 <- fs::dir_ls(data_dir1, regexp = "\\.csv$")
round2 <- fs::dir_ls(data_dir2, regexp = "\\.csv$")
round3 <- fs::dir_ls(data_dir3, regexp = "\\.csv$")
round<-c(round1,round2,round3)
cleandata<- round %>%   
  map_dfr(read_csv,.id="source",skip=16)%>%
    clean_names() %>%
    select(c(source, x1, x2, contains("head"), ambient:ambient_3)) %>% 
    rename(filepath = source,
           time_elapsed = x1,
           time_hms = x2,
           surface_temp = head,
           heatflux = head_1,
           avg_amb= ambient,
           amb_1 = ambient_1,
           amb_2 =  ambient_2,
           rh = ambient_3)%>%
    slice(-(1:2)) %>% 
    as_tibble() %>% 
    mutate(time_hms = hms(time_hms)) %>% 
    mutate(time_elapsed = dseconds(time_elapsed)) %>% 
    select(time_elapsed, time_hms, everything())
cleandata$round=substr(cleandata$filepath,regexpr("round",cleandata$filepath),regexpr("round",cleandata$filepath)+6) 
cleandata$filepath <- sub( ".*output/","", cleandata$filepath)
head(cleandata)
```
### Variables extracted
Since we are only interested in the head, all other zones have been dropped. The variables are as follows:

- `time_elapsed` (in seconds) describes how long the experiment has been running
- `surface_temp` (deg C) describes surface temperature for head zone
- `heatflux` (W/m2) describes heat loss in watts per meter square.
- `avg_amb` (deg C) describes average ambient temperature based on two measurements
- `amb_1` (deg C) first ambient temperature sensor
- `amb_2` (deg C) second ambient temperature sensor
- `rh` (percentage) describes relative humidity in the chamber.


```{r}

raw_manikin_long <- raw_manikin %>% 
  pivot_longer(cols = surface_temp:rh, names_to = "var")


```

```{r}

raw_manikin_long %>% 
  ggplot(aes(x = time_elapsed, 
             y = value,
             col = var)) +
  geom_line() +
  facet_grid(var ~ ., scales = "free")

```

### Variables calculated
A stable period is then chosen to average.

Some information required for the calculations includes the Area of skin surface on the manikin (Ask), which is `0.0978` in meters squared.

```{r}

start_time_mins = 15
end_time_mins = 40
area = 0.0978

start_time = start_time_mins*60
end_time = end_time_mins*60

cleandata %>% 
  filter(time_elapsed > start_time & time_elapsed < end_time) %>% 
    mutate(heat_resistance = (surface_temp-avg_amb)/max(0.01, heatflux),
           clo = (heat_resistance/0.155))
     
  

```

Two additional variables are calculated: `heat_resistance` and `clo`. Both are measures of insulation. 

Heat resistance is calculated from the difference between the surface temperature and the ambient temperature


- 

Other calculations

net solar influx (W/m2) is calculated as: 

$$heatflux_{(radiation)}- heatflux_{(no \ radiation)}$$

temperature correction for 4C (no radiation) is calculated as:

$$28/heat\ resistance_{(no \ radiation)} $$
temperature correction for 4C (radiation) is calculated as:

$$temp\ corrected \ heat loss_{(4C)} - solar\ influx $$

temperature correction for 30C (no radiation) is calculated as:

$$5/heat\ resistance_{(no \ radiation)} $$
temperature correction for 30C (radiation) is calculated as:

$$temp\ corrected \ heat loss_{(30C)} - solar\ influx $$