Edge AI predictive maintenance solution for industrial equipment
About the client
Senzoro specializes in producing artificial intelligence systems for industrial machines. Specifically, the company focuses on maintenance solutions, aiming to use sensor data to the fullest potential. Among customers that integrate their condition-monitoring systems and ultrasonic sensors are Siemens, Voith, and Windkraft Simonsfeld AG.
The challenge
Rolling bearings are the most common cause of breakdowns in industrial equipment. And Senzoro aimed to develop a solution that would validate bearing conditions inside industrial motors in real time, using ultrasonic signal analysis. While the company had its custom AI algorithms for this task, it lacked expertise in embedded systems, particularly in edge AI development.
The client needed a system that could:
Delivered value
Solution
From the beginning, the focus was on identifying hardware that could capture high-frequency ultrasound signals with the precision required for accurate diagnostics. We started by searching for the right Analog-to-Digital Converter (ADC) board. It was necessary so that the client’s algorithm could process ultrasonic signals and convert them into digital data.
In parallel, we selected suitable computing platforms on which an AI algorithm can run efficiently at the edge. Among the multiple options were Raspberry Pi Zero and Jetson Nano. However, both had either performance or compatibility issues, so our engineers excluded them. The choice fell on the Raspberry Pi 4 — because of the right balance of performance and cost-efficiency.
To ensure seamless AI processing, we created a custom orchestration script. It records short ultrasonic audio samples, feeds them into the client's model for analysis, and reads back the results.
After the main development phase, our team needed to validate that the algorithm could detect the normal rolling bearing sound from the faulty one. Tested with both simulated and real ultrasonic signals, the system accurately flagged each condition.
As a result, the developed device provides immediate feedback about the remaining useful life (RUL) of equipment, so maintenance teams know how to act before a breakdown occurs.
The challenge we presented to Lemberg Solutions was huge and it involved a complex set of different disciplines not found in many companies. We were all impressed by the competency of each and everyone in their team. They have a very transparent co-working style and really share the responsibility that the desired product works.
How it works
