A predictive model for maintenance in power lines
Almost six years ago I started working in my own startup. With my co-founders, we had the gut feeling that the Internet of Things (IoT) will be a disruptive technology. We weren't sure about which industry we should focus, so by coincidence, we ended up developing an Agtech solution for monitoring grain bags.
What do grain bags have in common with power lines? Short answer, nothing. But the long answer is the reason why I am writing this article.

While I was living in Chile, I attended to a Singularity University event where I presented our solution for monitoring grain bags.


This is a device with two functions: monitoring the biological activity inside the silo, for detecting when the grain is getting rotten, and also monitoring the movement, for detecting thieving in the bag! Yes, we measure the way the grain moves and we send an alarm by email or SMS.
While I was at the stand, Mauricio Retamal, who has experience in maintenance in high voltage power lines asked me if we can use our sensor for monitoring movement in a line, and I answered: “We haven’t done it before, but I think so”. So we did a first pilot to validate our technology could work in a power line.
After this first pilot, almost three years of work, we finish a product for easily gathering information in real-time from a power line.
The process was super interesting, we went from our original device, even the same case, we added just some clamps to hold the device to the power line. Those clamps are a little bit expensive, so later we designed a new device with the ability to work with only one clamp.

Also, we had some troubles with the electromagnetic field in a 140.000 V it is quite intensive, and also the weather conditions didn't help either. So we build an special IP67 case, for protecting our telecommunication module.
This device is able to measure the movement in the power line. Filtering the data we can understand if the movement is because there is a strong wind, because a tree falls in the line, because someone is trying to steal the power line and if the line is fallen.
This technology allows our customer to locate in a couple of seconds the specific place where the cable is fallen. In the past, it could take up to 4 hours.
This device, which has the ability to get data from a power line, opened us a new possibility: what if we measure other variables? For example, the temperature of the cable, or the height of the cable or the inclination of the power tower as the weather conditions in the area are crucial for the maintenance.

All this information flowing to our system is not easy to process by a human, so now, we are working on understanding how we can we decide the best moment to do maintenance in the line.

If you are interested in knowing more about this predictive model, feel free to contact me.