Search
  • Elipsa

Elipsa: AI for IoT delivered via MQTT



MQTT is the standard messaging protocol of the Internet of Things (“IoT”) and now it can act as the conduit for generating automated predictive insights through Elipsa.


Elipsa’s MQTT broker allows users to publish streaming machine data to a unique topic while subscribing to the predictive insights generated. Any system with MQTT publish capabilities can now seamlessly connect to the Elipsa platform for automated intelligent monitoring of critical equipment.


Elipsa combines MQTT with our Rapid Deployment Templates to deliver AI-based Predictive Maintenance at scale.


In this post we will detail how to utilize MQTT for Elipsa integration using Node-Red. While we are using Node-Red to demonstrate the capabilities, you can utilize any MQTT enabled system to publish streaming data to Elipsa.


Elipsa users with an active MQTT broker account will be granted login credentials unique from their Elipsa platform credentials.


For automated predictive maintenance, each user will have access to publish to /template/<your-elipsa-mqtt-id>/pub and subscribe to the insights that Elipsa publishes to /template/<your-elipsa-mqtt-id>/sub.


AI Integration via MQTT


Similar to integrating with Elipsa via our REST APIs, the Rapid Deployment Templates require data to be sent to MQTT in a specific format.


Function Node


For this demonstration, we are utilizing our pre-existing template for monitoring vibration data on the X, Y, and Z axis. As detailed in previous posts, our template will need to receive the following information:

  • API Key of Template (saved when you create or copy a template)

  • User Access Key (accessible via your Elipsa profile page)

  • Unique ID of your machine (such as serial number)

  • Streaming data from device

In our example, we are receiving X, Y, and Z max acceleration into Node-Red from an NCD Predictive Maintenance sensor. Note: Elipsa is hardware agnostic and thus any sensor can be utilized for your use case.


In Node-Red, we can utilize a function node to set the msg.payload to the format required.



Simply drag in a function node and paste the following, replacing the necessary data points.


var data = msg.payload;

//CONSTRUCT OBJECT TO SEND TO ELIPSA
msg.payload = {
    "api_key":"<your api key>",
    "access_key":"<your access key>",
    "unique_id":"<your unique id>",
    "batch":"N",
    "Data": {
        "x": <x axis datapoint name in node-red>,
        "y": <y axis datapoint name in node-red>,
        "z": <z axis datapoint name in node-red>
    }
};

return msg;

Note: the X, Y, and Z data point source will differ based on the raw data name of your sensor.


MQTT Out


With the msg.payload in the necessary format, we are now ready to transmit that data to Elipsa via MQTT.


This is done through a simple MQTT Out node. When adding your first MQTT node, you will need to configure the connection to the Elipsa broker. On the properties sheet of your MQTT out node, click the edit icon next to the Server field.


This will bring up the properties sheet for your MQTT server connection. Name your server. On the connection tab, set the server field as mqtt.elipsa.ai on port 1883. Leave all the other remaining defaults.


On the security tab enter your MQTT user and access key found on your profile page within the Elipsa platform click Add.


Back on the properties sheet of your MQTT Out node, name your node, and enter the topic as /template/<your-elipsa-mqtt-id>/pub click Done.


Automated Predictive Maintenance


At this point, you are now streaming your live sensor data to Elipsa via MQTT. Elipsa automatically begins learning normal behavior from your data in order to build a model and start making predictions.


With our REST API integration, the status of your model and predictions are returned directly when data is sent to the API. For MQTT integration, the insights are returned to a separate MQTT topic that you must subscribe to.


To access Elipsa’s predictive insights, simply add an MQTT In node to your flow.



On the properties sheet, Name your node and select the Elipsa server configured above. Set the Action to ‘Subscribe to single topic’. Enter /template//<your-elipsa-mqtt-id>/sub as your topic leaving the remaining defaults.


Finally, for testing purposes, connect the output of your MQTT in node to a debug node.


As Node-Red reads in new sensor data, it publishes the data to Elipsa’s MQTT. Elipsa’s Rapid Deployment Templates utilize the streaming data to learn normal behavior over time. Elipsa automatically builds a predictive model, and return predictions back to Node-Red through the MQTT subscription topic.



Subscribing to the Elipsa insights topic will return predictions as to the current health of your equipment. Insights are returned back into Node-Red via a JSON formatted object to be utilized for additional workflow steps.


Summary

Elipsa's MQTT broker provides the simplest, fastest, and most flexible way of integrating predictive capabilities into your IoT application and workflows. Any application or hardware capable of publishing data via MQTT can not be seamlessly monitored via Elipsa's AI-based Predictive Maintenance and Intelligent Monitoring.



 

Elipsa is a leader in delivering fully automated, scalable AI solutions for industrial IoT applications. 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. To learn more, please visit https://www.elipsa.ai