Degradation Signal Denoising Method for AC Contactors Based on Adaptive Modal Screening

In the world of electrical infrastructure, the AC contactor is a humble but vital workhorse. Found in countless control systems, these devices act as heavy-duty switches, frequently opening and closing circuits to power motors and machinery. Their job is simple in concept but demanding in practice: they must handle high currents and withstand the violent physical forces of repeated switching. Over time, this constant stress takes a toll. The metal contacts inside erode from electrical arcs, and mechanical parts wear down, gradually degrading the device's performance until it eventually fails. Because these failures can disrupt entire power systems, engineers are eager to monitor the health of contactors before they break. The challenge lies in listening to the machine. When a contactor operates, it generates complex signals—vibrations and electrical surges—that carry clues about its internal condition. However, these signals are often drowned out by a chaotic mix of background noise from electromagnetic interference, mechanical vibrations, and the harsh industrial environment. Extracting the faint, whispering signs of wear from this roar of static is like trying to hear a single conversation in a crowded stadium.

Researchers at Shenyang University of Technology have developed a new way to solve this problem, focusing on a method to clean up these noisy signals so the true signs of degradation become visible. Instead of relying on old techniques that often smooth out the data too much, losing the very details engineers need, the team created a system that acts like a sophisticated filter. They start by breaking the messy signal into smaller, manageable pieces, a process that separates the signal into different layers based on their frequency and behavior. Once the signal is split, the researchers face the task of deciding which pieces contain useful information and which are just random noise. Traditional methods often require a human expert to look at the data and guess which parts to keep, a process that is slow and prone to error. The new approach removes this guesswork entirely.

The team built a smart scoring system that automatically evaluates each piece of the signal. They look at two main characteristics: how much energy the piece carries and how complex or random its pattern appears. Real degradation signals, such as those caused by a contact bouncing or an arc forming, tend to have high energy and a certain regularity. Pure noise, on the other hand, usually has low energy and looks highly chaotic. By combining these two measurements into a single score, the system can instantly tell the difference between a meaningful signal and background static. It then discards the noisy pieces and stitches the clean, meaningful pieces back together. This creates a clear, high-quality version of the original signal, revealing the subtle patterns of wear that were previously hidden.

To test if this method actually works, the researchers ran a full-life experiment on Chint CJX2-8011 AC contactors. They operated the devices under heavy, realistic conditions, switching them on and off at a rate of 300 times per hour until they reached the end of their useful life. Throughout this process, they recorded the signals and applied their new cleaning method. They compared the results against several older, standard techniques used in the industry. The new method proved superior in every way. It preserved the sharp, important spikes in the data that indicate mechanical impacts and electrical arcs, while stripping away the surrounding fuzz. When the researchers measured the clarity of the cleaned signals, their method consistently produced the highest scores, indicating that it kept the most useful information while removing the most noise.

The impact of this improvement goes beyond just having a cleaner picture. Because the signals are now so clear, the researchers could more accurately group the contactors into different stages of health, ranging from normal operation to severe wear. This ability to distinguish between different states of degradation is crucial for predicting when a device will fail. By providing a clearer view of the machine's condition, this new method offers a more reliable foundation for maintenance teams to decide when to replace a contactor, potentially preventing unexpected breakdowns. The study demonstrates that by letting the data speak for itself through an automated, intelligent screening process, engineers can finally hear the quiet warnings of a machine before it stops working.