An Australian university study has successfully used artificial intelligence (AI) algorithms in conjunction with neural networks to detect defects in bridges in real time.
A neural network, as defined in the Oxford Dictionary, is a computer system modelled on the human brain and nervous system.
The Australian Catholic University research developed a method for the real-time structural health monitoring of bridges, using the Chumchup, Gocong, Ongdau and Ongnhieu Bridges in Vietnam to test it.
This new AI program uses bridge vibration data to accurately identify minor structural flaws before they become critical and then can alert maintenance crews.
Australian Catholic University associate professor for computational intelligence and Social Good Lab women in AI director Niusha Shafiabady led the multinational research team.
To develop the machine learning program, the Australian and Vietnamese researchers used the concept of a loss factor, which represents the process of energy dissipation across different vibration states, as a key indicator of structural health.
“Loss factor is, if you want to think of it in our daily life? If you’ve ever jumped on a trampoline, some of the energy goes into making the trampoline stretch and move and some of that energy doesn't come back to you,” Shafiabady said.
“That is the lost energy in the bridges, the loss factor is actually that lost energy, so it measures the energy that doesn't come back.”
This energy that is lost is actually turned into heat or it causes internal friction. The research team analysed the vibration patterns using loss factor, in the bridges to assess their structural health status.
The results demonstrated that the energy dissipation of the bridge during operation could be categorised into signals from three distinct sources: structural responses, defects-related indicators and noise interference.
By monitoring variations in the loss factor over time, the model was able to identify early signs of structural deterioration.
To paint a complete picture, the study used three different scenarios on the various bridges.
“The first scenario was when we had a heavy vehicle load on those bridges, for example, trucks or containers and the vehicles that exceeded the standard load limit when they were crossing the bridge,” Shafiabady said.
“The second case study was related to the light vehicle load that usually happens with small cars and motorcycles when it is not rush hour. At that time, the traffic was really not very bad
“The third case study was when we had heavy traffic, because one of the aims of these studies was looking at managing the traffic. We considered the high traffic scenario when we had different types of cars on the bridge, and public transport.”
Using the four different bridges and three scenarios, the study assessed the loss factor and compared the results using different AI algorithms to collect data on the bridges structural health.
“[This was done] to detect early structural changes when we saw that the loss coefficient is changing,” Shafiabady said.
“Then we took it as a sign that those changes mean that there exists some fatigue or damage in some areas of the bridge.”
While using the neural networks can help to identify serious defects in bridges, Shafiabady said that the purpose of the study is for the AI to flag when pre-emptive maintenance is required.
“Applying these AI methods was primarily for pre-emptive maintenance,” she said. “It’s not necessarily something needed to have immediate attention to that bridge, but just to avoid issues like catastrophic problems that could happen if the maintenance teams didn't look after the bridge.”
The neural networks operate within the bridge and in conjunction with the AI by making specific decisions separately.
“The AI methods that we have applied for the analysis are a combination of different neural networks where one neural network will make a basic decision, or one part of decision, and then that decision goes to another neural network to look into it further and finalise the decision,” Shafiabday said.
“This process, we hope, will allow the AI to come up with the outcome that we're looking for.”
The team behind the study believes utilising artificial neural networks trained to detect defects in bridges could revolutionise safety practices and save lives from potential structural failures.
“This diagnostic framework could save lives,” Shafiabady said.
“People worldwide use bridges daily to travel between home, work and school. Yet there are many examples that show, without proper defect detection and maintenance, these structures can fail, risking injury and death.”
Website: International Research Awards on Computer Vision
#computervision #deeplearning #machinelearning #artificialintelligence #neuralnetworks, #imageprocessing #objectdetection #imagerecognition #faceRecognition #augmentedreality #robotics #techtrends #3Dvision #professor #doctor #institute #sciencefather #researchawards #machinevision #visiontechnology #smartvision #patternrecognition #imageanalysis #semanticsegmentation #visualcomputing #datascience #techinnovation #university #lecture #biomedical
Awards-Winners : computer-vision-conferences.scifat.com/awards-winners
Contact us : computer@scifat.com
No comments:
Post a Comment