The document describes a case study of using wireless technology and a tablet-based process control software platform connected to wireless sensors to create simple laboratory tools for an honours research project. A survey found that honours students typically have less than 3 months to set up experiments and collect data manually, while PhD students have more time but experiments are more complex. The case study involved using wireless sensors connected to a tablet to automatically collect data from a vacuum membrane distillation experiment over several hours with minimal setup time. The outcomes were a shorter learning curve, reduced errors from manual data collection, and simpler experimental setup compared to traditional wired systems.
1. Using Wireless Technology to Create
Simple Laboratory Tools: A Case-Study
of Honours Research
Dr. Shane Cox
Instrument Works Pty Ltd
Lloyd Lian, Prof. Greg Leslie, Joel Tan
UNESCO Centre for Membrane Science & Technology
UNSW Australia
2. Introduction
• Mobile technology has progressively become more
visible within higher education
• But seems to be lagging when it comes to research
• It presents as a cost-effective and portable option that
can increase the efficiency of research
• Today are presenting a case study has investigated the
use of a tablet-based process control software platform
connected to wireless sensors
3. • Sydney based startup company
• Ex-researchers – building better tools
for engineers and researcher
• Developing a range of wireless
sensors for data collection and process
control
4. Why is doing research slow?
• It can take up to 6 months to design and build experimental
apparatus before they are ready to use
• Experiments are becoming more complex - meaning more sensors
and more data
• Researchers aren’t able to manage all of these resources
• This substantially reduces the amount of research they
generate and also affects its quality.
5.
6.
7. What do researchers say?
• Conducted a survey of 45 engineering researchers
• Predominantly PhD and Honours Students
Build my own equipment
Adapt from existing equipment
Outsource building of my equipment
< 1 month
< 3 months
< 6 months
> 6 months
How long to get data
HonoursPhD
8. What do researchers say?
PhD
Matlab
Labview
Proprietary for Hardware
None
Honours
Matlab
Labview
Proprietary for Hardware
None
What control system / DAQ system do you use?
9. PhD Vs Honours
PhD
• 3+ years to complete their
research
• Have the time to invest in
setting up their experiments
• Longer experiments makes
manual collect infeasible
• More experienced researchers
Honours
• 10-14 weeks to complete their
research
• Experiments are simpler, and
have less sensors.
• Shorter experiments make
manual collection possible
• Usually their first introduction to
research
10. A Different Approach
• Common device most people already
have.
• Familiar user interface that require
little or no training.
• Ability to collect many types of data in
one place.
• Wireless technology makes
connection simple
11. Honours Case Study
• Project investigated the use of vacuum membrane distillation
as an alternative method to recover solvents from shipping
industry waste water which is currently treated through
incineration.
• This projects was suited to a trial as it required
• Experiments were only short term
• Required to be setup in a fume cupboard due to the volatile
components making a traditional wired system problematic
• Needed to collect data from a number of sensors at once.
13. Honours Case Study
• Data acquisition up and running in
less than an hour.
• Data recorded at regular intervals
during experiment.
• Alarms were used to ensure setpoints
were maintained.
• Exports as a CSV file, ready for data
analysis
14. Outcomes
• Shorter learning curve than existing tool such as Labview
• Reduced chance of transcription errors introduced with manual
data collection
• Reduced setup complexity through a lack of wiring and
connection decreasing the opportunity for problems and errors
• Provided a unified collection tool for the current project.
Editor's Notes
We’ve seen in recent years that mobile technology has become more visible in the higher education sector.
This is particularly the case in teaching through tools like moodle – blackboard, lecture videos and other online collaborative tool.
However, the uptake of new technology seems to be somewhat lagging when it comes to research.
The tools we use today are often the same as those we used a decade ago.
The use of mobile technologies often present a cost-effective and portable alternative to existing tools that also potentially enable a greater efficiency in our research practice.
And today we are going to present a case study that has investigated the use of a tablet based process control platform that connects to wireless sensors.
And just briefly – the process control system that we’ve used has been supplied by a company called Instrument Works.
They are a sydney based startup company
They have developing a range of wireless sensors that are able to connect
To their data collection and process control software platform
The company was created by a number of researchers who left academia to start the company
With an aim to build better tools for engineers and researchers to collect and manage data
That embraces new technology particularly the use of mobile devices.
So in term of context and motivation for this work – an looking at how mobile technology might help us improve our efficiency in our research and focusing on the field of engineering.
Its important to look at the problem we are trying to address.
Research – whether it be academic or commercial is a slow process
Its takes a long time to go from having an idea – to having the results to validate (or invalidate) that idea
We start with planning out ourexperiments / doing the experimental design.
We then go an obtain all of the parts, tools, sensors and equipment.
One interesting fact here is that in academia 80% of researchers report that they’ve had to beg/borrow and steal for equipment to complete there research.
We then assemble our apparatus
Setup our data collection and where relevant our control processes
And its only then that we can conduct our experiments and start to obtain results
And this isn’t a straight forward linear processs. We often have to iterate to get everything right, for the right results.
And it can take upwards of 6 months to get this process right.
What we’ve also seen in the past decade or so it that the availability and cost of sensors has reduced significantly meaning we are now able to measure more parameters than previously would have.
This invariably makes experimental setups more complex and is generating signicantly larger amounts of data for us to analyse.
Researchers aren’t able to manage all of these resources with the tools that we have available
And this not only affect the quantity of research we can complete – but also the quality of that research.
And by way of examples that you commonly see.
If you walk in to any research lab in any engineering faculty any where in the world what you see is this:
Researchers using tools that were developed 10-15 years ago.
The ability to collect data from these types of devices in limited
Such that the most common use case is for the data to be collected by hand
This typically results in two problems
Transcription errors either when writing the data down, or when subsequently transferrring them to a digital format.
The lack of collection of other metadata which can be used to ensure the results are valid
For example in this picture we have a research recording data from a pH meter.
Whilst the values are recorded what hasn’t been recorded are the
Details of the calibration, when it was calibrated, details of the device or the probe or the calibration solutions
Or perhaps she hasn’t recorded the temperature of the solution
And what this means is, if we later discover that there was a problem with the probe or the calibration – we aren’t able to go back and identify the data that may be contaminated
Or if we see some anomalies in the data aren’t able to look at some of those external influences that may have created that problem
Or at the other end of the spectrum – you might see this.
An elaborate experimental apparatus –
That had a lot of money spent on it to be built
thats been designed for a specific project / experiment or purpose but
That, isn’t flexible as if and when your experiments change
Or the experimental parameter or conditions change.
And significantly has been constructed in such a way that the users/researcher aren’t necessarily familiar with and are able to change/update or
Or the conditions change that you want to run those experiments
That uses tools that researchers aren’t familiar with and can’t modify
What you also find with these types of setups –
Is that when they are finished the project that they were designed for or the person that used them leaves
They end up sitting in the corner of the lab gathering dust
Or have some of their components removed and repurposed for other project – essentially rendering them useless for the next project that comes along.
And so based on this – we’ve been and dicussed these observations and conducted a survey with a number of engineering researchers who are predominantly PhD or Honours students.
What we see is that for PhD Students about 90% percent of them either build their own equipment or adapt what they are using from existing equipment.
And this makes sense when you consider much of what they are doing is extending the work of those before them.
But for honours students we see that very few build there own equipment – and this might be considered due to the time contraints of an honours project.
Most adapt it from other existing equipment similar to a PhD student.
But a significant portion outsource their equipment building. And when you drill in to this you find that the outsourcing is predominantly to the PhD student or post-doc that they are working most closely with.
We also looked at the software that they use to collect data and control their equipment.
For the PhD students – LabView is the tool that is most commonly used, follow by proprietary software rom the hardware suppliers, with a few using matlab – an in many cases they reported needing to use mulitple tools for collecting data.
For example – some sensors in a experiment being connected to labview and others using proprietary software.
But what also stands out is that in almost all cases they are using some tool to help them collect their data.
On the other hand for the Honours students
half of them, don’t use any tools and the other use either proprietary software or labview.
So what we see is that for an honours students are more likely to manually collect their data – hand writing out their data.
So what are the reasons for this difference
The first difference is that honours students only have a short period of time in which to do their research, typically only 10-14 weeks, compared to a PhD student which has 3 years.
And in that time they need to design, setup, run and analyse their data.
So their projects are simpler
The experiments are simpler
They collect less data and they are shorter
And this make it easier for them to collect data by hand rather than have to learn new tools
For honours students this is often their first introduction to research, so their skills are limited to begin with.
As a result the experiments tend to be simpler and shorter, which make it more convenient to collect stand
So an alternative approach – with the instrument works system
is to use a device that our students already have, their smart phone – or provie a deidcated device such as an ipod or ipad.
What that provides is a user interface that they are already familiar with.
What also interesting with with mobile devices you notice is that most apps don’t come with manuals or instructions but a common design paradigm that makes them easy to use and require little or no training
Which is important for students who don’t have the time to learn a new set of tools.
And with wireless connectivity – in this case Bluetooth low energy – we are able to connect to a range of different sensors and collect all of their data in one place.
As a first case study – we’ve had a student trialing the system during their Honours project in which they were investigating the recovery of solvents from wastewater in petroleum processing.
This was a particularly useful case for the trial as the experiment setup needed to be place in a fume cupboard due to the volatle components being treated, which made a wired DAQ systems that we might have otherwise used problematic.
This is experimental setup used in this project.
Where we are measuring the suction pressure on the vacuum side of the membrane.
We monitored the temperature of both the feed and permeate sides to ensure they remained at their set points
And a balance was employed to measure the membrane flux. (Although this was laterremoved as we weren’t able to get accurate results due to the low fluxes and the air flow in the fume cupboard)
Once the apparatus was setup – putting in the required sensors and setting up the data acquisition system on the phone took less than an hour.
Data was able to be recorded at set intervals of the students choice and alarms were used to watch the set points for the temperautre.
At the end of each experiment data was able to be exported via email as a CSV file where it was imported in to Excel for data analysis.
What we’ve found so far from this case study and using this tool – and we should point out that this case study is on going
It’s a shorter learning curve than existing tools such as labview, which we typically don’t or aren’t able to use with an honours project because they aren’t familiar with it.
By digitising the data collection process we reduce the chance of transcription errors due to manual data collection
We’ve also reduced the setup complexity through the use of wireless sensors – which makes the setup process quicker and reduces possible points of failure and the need for trouble shooting.
And finally we’ve provided a unified collection tool for sensor data for this project. Where previsouly we might have recorded the pressure data online but the balance data manually.