Background
Big electrical transformers usually have automatic on-load tap changers.
These are mechanisms that adjust the relation between the in- and out-coming voltages, helping keep the latter as constant as possible.
Whenever an adjustment is needed, a motor changes connections between different possible taps, making a characteristic sound.
When the mechanical parts of tap changers age or start malfunctioning, the sound they produce might vary or the time needed to change between connections may become longer.
The transformers on which we applied our solution did not have a real time monitoring system to supervise the health of the tap changers, which are actually the only mechanical parts of a transformer and one of the most sensible ones. In particular, a failure in operating properly can create severe damages to the full machine and cause disruptions in the energy distribution. These failures imply important costs: only the reparation of the tap changer itself can cost between 200k-300k NOK and losses associated with energy disruption during a certain period of time can be much higher.
Solution
We installed a microphone close to the transformer connected to an edge device with our Resonyx software. The system listened continuously to the sounds the machine produced and our analysis models automatically registered all tap changer sounds and measured their duration. With this data we could create statistics that allowed the detection of anomalous behaviour of these mechanisms.
On top of this, thanks to the continuous monitoring together with the immediate detection and analysis of the sound we could develop a fast alarm system that would trigger if the tap changer lasted longer than a certain threshold. With this alarm, the owners of the transformer could react fast in case of failure.
Furthermore, all tap changer operations were recorded and their sounds were stored in the cloud, allowing for manual inspection by an expert person in case there was the suspicion that something might be wrong. This inspection, as well as the monitoring of the general sound of the environment can be done remotely and, especially for machinery located in remote places, can help to reduce the frequency of inspections by the maintenance crew.
Value
With our audio analytics solution, we could keep track of all tap changer mechanisms in all monitored transformers. We could detect anomalous behaviors, propose inspections of the mechanism before they failed and also provide fast alarms when a failure actually happened.
Therefore, we could help increase the lifetime of transformers, since maintenance can be done before failure. In addition, the monitoring can prevent the failure of the whole transformer and helps to ensure the uptime of the electricity grid.
Once this system was set, all analysis was done automatically and stored in a cloud, allowing the transformer experts to keep track of the health of their machines without having to physically travel to all of them.
Comments