The reference design combines edge machine learning and Wi-Fi connectivity to enable predictive maintenance for industrial motors and rotating equipment.
Microchip Technology has introduced an edge AI-enabled reference design with Wi-Fi enabled for developing predictive maintenance systems for electric motors and other rotating equipment. Suitable for both brand-new equipment and retrofit applications, the compact, battery-powered device monitors machine health locally, helping reduce maintenance costs, reducing unplanned downtime, and increasing the lifespan of machines. Edge computing makes the design more private by eliminating dependence on cloud-based AI.
With edge AI, Wi-Fi connectivity, and security features, the ready-to-produce reference design provides continuous monitoring of equipment condition. Machine learning models analyse vibration and audio data to detect issues and predict maintenance requirements, while temperature readings are used as well for better condition monitoring. Sleep mode during idle periods with activation of Wi-Fi only when needed enables battery life of more than three years.
At the core of the hardware is the ultra-low power microcontroller that has built-in security capabilities along with the low-power SPI Wi-Fi module. The Wi-Fi operates from 3.0 V to 4.2 V and provides low-power wireless connectivity for industrial applications. Its compact footprint and low-power operation make it suitable in an industrial application for extended periods of time.
The reference design is designed to accelerate product development by providing the design files and firmware for production-quality hardware. Machine learning models are developed with the use of MPLAB Machine Learning Development Suite, while cloud connectivity and dashboards can be implemented using Avnet’s IOTCONNECT solution. The device can be installed externally as a retrofit module or integrated directly into new equipment.
For evaluation and further development of the product, Microchip recommends compatible hardware includes the Motor Control Plug-In Module, the Motor Control Development Board, and a three-phase brushless DC motor with an encoder having 24 V. Together with the supplied design resources provided by Microchip and its design files, engineers can evaluate the machine learning algorithms and deploy secure, edge-based predictive maintenance solutions.
For more information, click here.