Skip to contents

Log Files

logfiles.Rmd

During a picoinjection or droplet making experiment, it can be important to know the fluid flow parameters at any given time, e.g. in order to link them to an image that was taken at a specific time. Both, the OB1 pressure controller and the Cetoni syringe pump allow to record the history of parameters, i.e. pressure or flowrate, over time and store them in a log file.

In this vignette, I will format and analyse those log files in order to show them in a single plot.

library(picoinjector)
library(dplyr)
library(ggplot2)
# show fractional seconds with three digits
op <- options(digits.secs = 3)

What do I want?

  • plot with q_oil, q_emulsion and p_injection on the y-axis, time on the x-axis
  • function to give me pressure and flowrates for a given time

Challenge:

  • align pressure and syringe pump profiles such that the timesteps are the same

Overall workflow:

  1. Read in the log files
  2. Find values of elapsed time of both syringe and pressure log in order to align them
  3. Plot

syringe pump

The Cetoni syringe pump log file can be loaded using the convenience function logfile_read_syringepump. This reads the CSV file and formats the columns a little bit.

(This function works only for a particular configuration of the log file in the Cetoni Elements software.)

syringe <- logfile_read_syringepump("data/logfiles/syringe_pump_log_20220805_150634.csv")
syringe
#> # A tibble: 149,607 × 4
#>    elapsed_time timestamp               channel flowrate
#>           <dbl> <dttm>                  <chr>      <dbl>
#>  1          0   2022-08-05 15:06:34.569 q1            10
#>  2          0   2022-08-05 15:06:34.569 q2            10
#>  3          0   2022-08-05 15:06:34.569 q3             0
#>  4          0.1 2022-08-05 15:06:34.669 q1            10
#>  5          0.1 2022-08-05 15:06:34.669 q2            10
#>  6          0.1 2022-08-05 15:06:34.669 q3             0
#>  7          0.2 2022-08-05 15:06:34.769 q1            10
#>  8          0.2 2022-08-05 15:06:34.769 q2            10
#>  9          0.2 2022-08-05 15:06:34.769 q3             0
#> 10          0.3 2022-08-05 15:06:34.868 q1            10
#> # … with 149,597 more rows

It can be seen that every row contains an absolute time stamp which will allow us to associate the flowrate with images data acquired on the microscope.

As will be shown later, the pressure controller log file does not contain absolute timestamps, just a column for the elapsed time. However, by aligning flowrate and pressure profiles, the absolute time stamp can also be inferred for the pressure log. To that end, one has to include an alignment mark during the experiment, e.g. by setting both, pressure and syringe pump to a specific value at the exact same time. By inspecting the resulting profiles, one can identify the alignment mark and shift the pressure profile so that both alignment marks conincide with the same time.

An overview graph of the syringe pump flowrate can be plotted using logfile_plot_syringepump.

Here, the alignment mark is at the end of the experiment, around minute 81, where flowrate on channel 2 and pressure on channel 2 were set to zero. Zoom on the x-axis to find the precise time:

logfile_plot_syringepump(syringe) +
  xlim(c(81.5, 82))
#> Warning: Removed 148704 rows containing missing values (`geom_line()`).

Investigate the logfile rows in order to find the exact time when q2 was set to zero:

syringe %>% 
  filter(channel == "q2",
         elapsed_time > 4904 & elapsed_time < 4905)
#> # A tibble: 9 × 4
#>   elapsed_time timestamp               channel flowrate
#>          <dbl> <dttm>                  <chr>      <dbl>
#> 1        4904. 2022-08-05 16:28:17.611 q2           300
#> 2        4904. 2022-08-05 16:28:17.711 q2           300
#> 3        4904. 2022-08-05 16:28:17.812 q2           300
#> 4        4904. 2022-08-05 16:28:17.911 q2           300
#> 5        4904. 2022-08-05 16:28:18.012 q2           300
#> 6        4905. 2022-08-05 16:28:18.111 q2             0
#> 7        4905. 2022-08-05 16:28:18.211 q2             0
#> 8        4905. 2022-08-05 16:28:18.312 q2             0
#> 9        4905. 2022-08-05 16:28:18.411 q2             0

pressure controller

A drawback of the pressure controller log files is, that the system time is not stored. Therefore, we need to align the pressure profile with the syringe pump profile in order to get that information.

pressure <- logfile_read_pressure("data/logfiles/pressure_log.txt")
pressure
#> # A tibble: 394,856 × 4
#>    elapsed_time channel value    mbar
#>           <dbl> <chr>   <chr>   <dbl>
#>  1          0   p1      target  0    
#>  2          0   p1      read    0.612
#>  3          0   p2      target 50    
#>  4          0   p2      read   50.0  
#>  5          0   p3      target  0    
#>  6          0   p3      read    0.970
#>  7          0   p4      target  0    
#>  8          0   p4      read    0.833
#>  9          0.1 p1      target  0    
#> 10          0.1 p1      read    0.605
#> # … with 394,846 more rows

Find time when p2 was set to zero:

pressure %>% 
  filter(channel == "p2",
         value == "target",
         elapsed_time > 4861.5 & elapsed_time < 4862.5)
#> # A tibble: 9 × 4
#>   elapsed_time channel value   mbar
#>          <dbl> <chr>   <chr>  <dbl>
#> 1        4862. p2      target   554
#> 2        4862. p2      target   554
#> 3        4862. p2      target   554
#> 4        4862. p2      target   554
#> 5        4862  p2      target   554
#> 6        4862. p2      target     0
#> 7        4862. p2      target     0
#> 8        4862. p2      target     0
#> 9        4862. p2      target     0

align profiles

What are the times where pressure and syringe pump logs should be aligned?

tPressure <- 4862.1
tSyringe <- 4904.6

Align pressure log to syringe pump log:

pressureAligned <- logfile_align_pressure(pressure, syringe, tPressure, tSyringe)
pressureAligned
#> # A tibble: 394,856 × 5
#>    elapsed_time timestamp               channel value    mbar
#>           <dbl> <dttm>                  <chr>   <chr>   <dbl>
#>  1         42.5 2022-08-05 15:07:17.069 p1      target  0    
#>  2         42.5 2022-08-05 15:07:17.069 p1      read    0.612
#>  3         42.5 2022-08-05 15:07:17.069 p2      target 50    
#>  4         42.5 2022-08-05 15:07:17.069 p2      read   50.0  
#>  5         42.5 2022-08-05 15:07:17.069 p3      target  0    
#>  6         42.5 2022-08-05 15:07:17.069 p3      read    0.970
#>  7         42.5 2022-08-05 15:07:17.069 p4      target  0    
#>  8         42.5 2022-08-05 15:07:17.069 p4      read    0.833
#>  9         42.6 2022-08-05 15:07:17.168 p1      target  0    
#> 10         42.6 2022-08-05 15:07:17.168 p1      read    0.605
#> # … with 394,846 more rows

plot

Remove unused channels:

syringe <- syringe %>% 
  dplyr::filter(channel != "q3")
pressureAligned <- pressureAligned %>% 
  dplyr::filter(channel == "p2")
# find out by how much the second y-axis has to be scaled:
r <- max(syringe$flowrate) / max(pressureAligned$mbar)

Merge pressure and syringe dataframes:

df1 <- syringe %>% 
  dplyr::rename(value = flowrate) %>% 
  dplyr::mutate(type = "syringe pump")

df2 <- pressureAligned %>% 
  dplyr::filter(value == "target") %>% 
  dplyr::select(!value) %>% 
  dplyr::rename(value = mbar) %>% 
  dplyr::mutate(type = "pressure controller")

# scale pressure
df2 <- df2 %>% 
  dplyr::mutate(value = value * r)


df <- dplyr::bind_rows(df1, df2)
ggplot(df, aes(x = elapsed_time / 60, y  = value, color = channel, linetype = type)) + 
  geom_line() + 
  scale_y_continuous(sec.axis = sec_axis(trans = ~./r, name = "Pressure [mbar]")) + 
  theme_pretty() + 
  labs(y = "Flowrate [ul/h]",
       x = "Elapsed time [min]",
       linetype = "Instrument",
       color = "Channel")