6 minute read
Meet the Speaker: Dr Aatif Ali Khan
Dr Aatif Ali Khan discusses his research into ‘smart firefighting’ and the development of an AI-assisted tool to enable firefighters to conduct real-time risk analysis based on information gathered at a fire scene.
Dr Aatif Ali Khan is a lecturer at The University of Canterbury. At the Fire NZ Conference, he will be speaking on the topic of “Smart firefighting: A framework for decision making using AI-assisted tool in a building fire.”
Dr Khan earned his PhD from The Hong Kong Polytechnic University in structural fire engineering. He has provided consulting services for a number of industrial projects globally. He is a Chartered Engineer (CEng) from the Institution of Fire Engineers (UK) and International Professional Engineer [IntPE(NZ)] from Engineering New Zealand. His research interests include fire modelling, structural fire safety, fire investigation of structural fire accidents and the application of artificial intelligence in fire engineering problems and smart firefighting. Currently, he is focusing on developing the fire safety design for firefighters during emergency response.
FNZM: How did you find your way into fire engineering?
AK: I completed my undergraduate degree in India in 2008 in Mechanical Engineering and joined an automobile firm. I didn’t enjoy that, and I found a job in the fire sector in Saudi Arabia. After working there for a few years, I went to South Korea to study a master’s degree in Computational Fluid Dynamics (CFD).
After finishing that, I got the chance to do a PhD in Structural Fire Engineering at the Hong Kong Polytechnic University, and as I was finishing the PhD my supervisor was working on a project on smart firefighting, so I started focusing on that – and that’s now part of my research.
When a lecturer job came up at the University of Canterbury I applied. Coming to New Zealand was a no brainer because of Canterbury’s reputation as one of the leading universities in fire engineering. I arrived here just last November.
FNZM: What is ‘Smart firefighting’?
AK: The ‘smart’ word is just used because of the harnessing of new technologies to assist firefighters. It’s not because firefighters aren’t smart; they’re very smart; it’s just that it’s using technologies such as artificial intelligence, virtual reality, and augmented reality, and using sensors and real time data from inside buildings to assist firefighters in doing their job.
FNZM: What aspect of smart firefighting does your research focus on?
AK: It’s about utilising a library of critical events, implementing building IoT (Internet of Things) sensor networks, and training artificial intelligence (AI) models to forecast dynamic critical events in real-time. What we want to be able to forecast in the case of a fire, for example, is what might happen after a certain number of minutes, because in the case of fire one or two minutes can make a big difference. If we can tell a firefighter that in five minutes a window will break or that after 10 minutes a structure may destabilise before these events occur, it would be helpful to them.
We’re developing scenarios through computer simulation to create a large database so that the model is able to make predictions. Most of the time we are taking data from CFD simulations but also from experiments, knowledgebased data from historical events, and sometimes we create models from physics-based empirical correlations. Each has their limitations.
FNZM: What is the output of the research, and the audience?
AK: The first thing we did was to collect data on 50-60 years of accidents, and to do this we looked at firefighter fatalities predominantly from the US because most of the US data is open to the public. From this data, I created a library of critical and precursor events detailing what happened in each case, including such factors as whether a door was open, whether there was a fire in the corridor, whether fire reached the false ceiling, etc.
I created around 200 events from the data, and then discussed these with experienced firefighters in Hong Kong to verify whether these types of events had occurred in Hong Kong too. They checked their data and added to it from their experience.
Most of the time, we can see that there are a reasonable number of scenarios that are common across different countries, so we then looked to survey their frequency and consequences. I am still in the process of documenting this.
The idea is to provide firefighters with timely information in a language they can understand, and data outputs, such as temperatures, tend to be numeric, which may not mean much to a firefighter. Our software, therefore, will convert the number into what it actually means, so if a 500-kilowatt fire will likely cause a window to break after five minutes, the software will talk about the window breaking rather than simply stating the temperature.
One more thing, I’d like to highlight is that this project is highly interdisciplinary, calling for collaboration from a diverse range of experts, including those in fire engineering, computer science, structural engineering, and beyond.
FNZM: How will firefighters consume the outcomes of the research?
AK: We are developing software that will connect with the BIM model and looking to provide a device that provides a user interface. This user interface will provide the firefighter with the probability of critical events that are in the database library. The incident command officer can use this to conduct real-time risk analysis based on information gathered at the fire scene.
This ultimately gives the firefighter more time for their dynamic risk assessment so that they can make more informed decisions based on real-time information.
FNZM: How far are we from that today in NZ?
AK: I’m looking to conduct a similar survey study in New Zealand as well and am in the data collection phase at the moment.
I’m also working on the use of virtual reality in firefighting. Essentially, we can use the critical event data we’ve collected for our AI assisted tool to train firefighters, and we can make training scenarios based on what actually happened.
Accidents are sad things, but they are more tragic when we don’t learn from them. There is a gap between research and what firefighters are actually doing, and we want to bridge it. Fortunately, New Zealand is an amazing place for this because our firefighters and academics are so connected. The rest of the world should look to us and do the same.