 Did you know that one fire over four is triggered by an electrical issue? Most of them generated by an electrical arc. For decades our customer and regulatory bodies have been innovating to increase electrical installation safety and make sure all protections are properly installed. As the market evolves we see new and different types of applications emerging requesting more and more innovation to progress against electrical arcs. Solar panels, batteries, power tools, e-bikes and so on. One important point is the existing system generates a large number of false positives. For some of them this is a key issue as they drop the machine yields. Today we are going to demonstrate how you can simply build your own electrical arc detection mechanism with STM32 and AI. Let's dive into our demonstration. At the heart of this system is the STM32 G4, one of our most successful mainstream MCU. As you can imagine we will focus on current data. To do that we will gather it from a current transformer connected to a filters and an amplifier. Then pass it into a 12-bit ADC. Something you can do on all STM32 family by the way. In essence the demonstration mimics the electrical spark by positioning two electrodes in close proximity. We slowly bring them together until we can visualize an electrical arc. We use Nano Edge AI to generate and implement a classification library with two classes to differentiate whether the arc is happening or not. In this case Nano Edge is the tool of choice. It generates a library without any ML expertise than optimizes it. Here we are running with a model that takes less than 20 kilobyte of flash. All of that was completed within one day where in other cases it would take months. AI library classification is running in real time and then can allow the user as soon as it detects it. In this case you can easily imagine that fast edge processing is essential. We need to react as soon as possible as the arc is happening so we can cut the power. Combining the large ST microcontroller portfolio and optimized AI library allows our customers to address electrical arc detection in multiple systems. To find out more we invite you to visit STM32AI.ST.com and explore the important subject of electrical arc detection.