Start main page content

Engineering innovations at Open Day

- Wits University

Students at the 2019 School of Electrical and Information Engineering Open Day presented their engineering innovations for solutions to everyday challenges.

This year 140 students presented 70 projects at the annual School of Electrical and Information Engineering Open Day. The projects represent the class of 2019’s top engineering innovations and visions for solutions to everyday challenges.

As part of the curriculum, students are required to work in pairs on a project that demonstrates the knowledge they’ve gained over their undergraduate years. The aim is for them to design and build their solutions and to present these during the Open Day. In addition to the quality of their projects, the students were also assessed on their ability to effectively communicate and present their projects to a general audience.

The projects were on show in and around the Chamber of Mines building on 29 August. Projects ranged from a mobile off-grid electricity box that is charged by solar electricity to shear wave technology to detect anomalies in tissue that could indicate tumours and growths without invasive surgery. There was also a project demonstrating the use of machine learning to make matches between digitally generated identikits and photographs.

At the core students used their projects to think about engineering remedies to the problems they see in today’s world. Their projects showcase the intersection of modern engineering, social awareness, imagination and communication in coming together to solve problems.

For Chizeba Maulu and Mabatho Hashatsi their project, titled “Electricity theft detection in low voltage networks” their data-driven invention uses community-wide electricity consumption data to pinpoint irregularities in consumption patterns that could indicate theft of electricity.

“This is a huge problem in our country with illegal connections and we know we are running out of time to fix the problem of electricity theft that affects everybody and especially our hospitals and schools,” says Hashatsi.

Maulu adds: “We know that billions are being lost in electricity theft every year in South Africa. In one area in Joburg we found that in a two-week period they were able to recover R15 million in lost revenue just by being able to find out where the theft of electricity was taking place and taking action.”

Their classmates Seth Bulkin and Alexandra De Nooy also responded to one of society’s deepening but ignored crisis: the mental health burden. Their project “Detection of depression from speech” combines the development of an app with the findings from a large data set to form the basis of an assessment tool for depression. The app works through prompting the user to complete an audio questionnaire. The recording is uploaded and assessed within seconds for speech variations, tone, pitch and volume and that could be indicators of depression. The assessment is returned as a score and where it’s indicated, the app offers recommendations for interventions.

“There is still so much stigma attached to mental health but the app is private and confidential while allowing people to assess their own mental health on a regular basis, which people don’t do,” says Bulkin.

He says there’s opportunity for the app to be used by mental health care workers who can assess and monitor patients and pick up early warning signs and take appropriate action to protect patients from self-harm.

For other final-year students Jason Parry and Muhammed Cassim their project seeks to improve communication for deaf and hearing-impaired people. Their portable and cost effective sign language to speech translator features a glove that utilises force-dependant resistive sensors that are placed along the fingers of the glove. These sensors determine flexion of the fingers. The movement in turn gives an analogue signal which is converted into a digital signal by an analogue-to-digital chip. The text is then synthesised into speech through an amplifier-speaker circuit.

Parry says they are currently working at an 85% accuracy of the conversion of American Sign Language fingerspelling to speech and their prototype’s advantage is its low-cost and its portability (that can run off a power bank).

Each year the School give out prizes for the top projects that were best presented. This year’s  “Best Poster” prize went to Zakhele Skosana and Thandeko Khumalo for the design of their project title “Partial Discharge detection under non-power frequency voltages”, supervised by Professor Cuthbert Nyamupangedengu.

Senior lecturer Hugh Hunt said: “The overall quality of this year's work was very impressive and showed that the School of Electrical and Engineering is deeply engaged in modern data science and the 4th industrial revolution, while simultaneously focusing on the forefront of power and energy engineering.”

The annual Open Day ended with the tradition of planting the South African tree of the year (which for 2019 is the Sclerocarya birrea or Marula) and installing a plaque in honour of this year’s fourth-year class.