We do not use artificial intelligence (AI) for the sake of it – we use it for specific tasks, which we demonstrated to work better than traditional statistics. Conversely, we are not involved in robotics in the classical sense (of manufacturing or humanoid use), we use the same control systems and hold ourselves to the same industrial and safety standards. Our devices are mechanical too, and they need precise control in movement. We also offer research and development consulting services in these fields as well.
One of our applications is ‘Fault Detection and Diagnosis’ for refrigeration and air conditioning.
Sadly, home consumers regard their refrigerators and air conditioners as electronics, partially because these devices are commonly available in electronics stores.
In reality, anything that uses the vapour-compression refrigeration cycle is a mechanical device, and they need periodic maintenance just like a car. Instead, more often than not, consumers are commonly running their equpment down to the point of catastrophic failure, by which time, the only solution is complete replacement.
This is the equivalent of not doing any periodic maintenance of a car and then having a breakdown, resulting in an engine replacement.
In hot climate countries where most of the energy produced is used for air conditioning, good funcional maintenance is more essential than ever.
We created a neural network-based early warning system that predicts faults before they happen. It is the functional equivalent of the ‘Check Engine’ light of a car. It is cheap to implement, and we believe that it works better than the statistics-based systems used in larger building management platform. We represent the sensor values as pictures, and the system simply does image classification, like so:
In fact, we believe it works so much better, that it is unethical to profiteer on it. So we decided to give away the model for free along with its example generic training data, but we kept our data and the required know-how as AVZ’s intellectual property. This gives an advantage over our competitors, but at the same time, is a huge step in saving the environment.
Considering that refrigerant leak is the most common problem, and these materials are throusands of times worse in ecological harm than carbon dioxide, we maintain that this is the ethical thing to do.