Broadsens, a leader in low-power industrial sensors, and Elipsa a leader in plug and play AI solutions for IIOT are combining machine data with machine learning to provide easy to configure predictive maintenance at scale.
Industrial equipment such as pumps, compressors, generators, HVAC, etc. are critical to successful business operation. Unplanned downtime of such equipment can be catastrophic in terms of economic loss to an organization. Historically, maintaining such equipment has been reactive, incorporating onsite diagnostic or notifications of failures that have already occurred.
With the growth of the Industrial Internet of Things (“IIOT”) this critical equipment is becoming smart and connected. With this new stream of data, machine learning has proven capable of detecting patterns indicative of future problems allowing for predictions of downtime and a proactive approach to equipment maintenance. This growing area is known as Predictive Maintenance.
Machine learning algorithms can learn from numerous types of sensors and data points. Often times, the most effective data points for successfully building predictive maintenance algorithms have proven to be vibration data, learning that certain vibration profiles can predict the future failure of a machine.
The Broadsens SVT200-V is the only compact true real-time wireless vibration sensor that can monitor machine condition continuously with more than 3-year battery life in typical usage.
SVT200-V has two sampling modes: continuous low-speed sampling mode and high-speed sampling mode. It acquires x, y and z axis data in 16-bit resolution continuously. When the sensor detects a vibration event, it switches to high-speed 6.4kHz sampling rate, takes a fixed amount of data, calculates velocity RMS and acceleration RMS value, and sends the result plus temperature to the wireless gateway.
The smart algorithm inside SVT200-V adjusts the sampling mode dynamically and optimizes the energy usage. It can capture both high frequency and low frequency defects from the machine and structures.
The Broadsens wireless gateway connects and controls many sensors including the SVT200-V, SV200-A, SVT300-V, SVT300-A, SVT400-V and SVT400-A. The gateway includes 4-core ARM processor for real-time data visualization and processing. The gateway is configurable to transfer real-time sensor data to third-party systems via MQTT such as the Elipsa AI Platform.
Broadsens real time wireless vibration & temperature sensor
Elipsa’s AI-based Predictive Maintenance seamlessly deploys across any workflow on the edge or in the cloud, increasing the availability and output of critical equipment. Elipsa’s self-training AI models and bolt-on approach enable AI deployments that are simple, fast, and flexible.
Through the use of Elipsa’s Rapid Deployment Templates, users can start intelligently monitoring critical equipment in as little as five minutes. Elipsa’s Broadsens Vibration Analysis template takes in real-time velocity RMS data on the x, y, and z axis.
Once an SVT200-V is connected, data is streamed to Elipsa via MQTT. Elipsa’s Rapid Deployment Template automatically starts to learn the normal vibration profile of the machine that the sensor is attached to. After a set number of data points (as defined in the template), Elipsa automatically builds a machine learning model and returns predictions back to the Broadsens gateway.
Elipsa’s innovative approach learns normal vibration of a given machine in order to detect abnormalities of any kind reported back to the user via the Broadsens software.
In this case study, Broadsens's wireless vibration sensor SVT200-V and wireless gateway GU200S, vibration analysis software BroadVibra are used to monitor an air condenser. SVT200-V inside single screw-hole package is used with magnet mount. The sensor with magnet mount can be installed in several seconds. Data are fed to Elipsa’s AI platform continuously in real time via MQTT protocol. 24-hour data from the sensor is shown as below.
After several days, the sensor was removed to simulate accident to the air condenser. Elipsa AI algorithm detected the change with both velocity RMS and acceleration RMS value. Velocity RMS value showed more significant change, which is consistent with the industrial standard that prefers vibration velocity RMS over acceleration RMS for machine condition monitoring.
Wireless vibration sensor mounted on panel
Vibration velocity RMS AI prediction
24-hour vibration data from the panel
Acceleration RMS AI prediction
Predictive Maintenance at Scale
The combined Broadsens/Elipsa solution can easily be deployed across a range of equipment at scale. Make the promise of predictive maintenance a reality to start proactively managing critical equipment. There’s no need for a data scientist or a large capital ex spend. This plug-and-play solution will reduce downtime, increase operating efficiency, improve scheduling of key employees, making your machines more sustainable.