Real time oil tank monitoring — Modern Way IoT
I had been to India recently and while talking to one of my friends who is a site manager of a chemical plant; he was explaining his challenges of not getting timely information when oil tanks go below a certain % and management team having no visibility of data picked up during the shifts.
I was surprised to hear because company invested money in procuring machines, setting them up and they were relying on humans to check readings, informing site managers and then site managers informing management team sitting elsewhere.
Upon discussing it further, it was evident to me that they were hesitant to invest further on software not only because the products shown were expensive but also other customisation costs involved in that.
Seeing his helplessness, I decided to pay a visit to the site and see if I can help. Very next day I went to the site and was overjoyed seeing #opcua servers. My instant reaction was I can help him.
Next plan was to write an #opcua client to get data out of the server. I knew a NodeJS opcua sdk (https://node-opcua.github.io/) that helped me extracting data out (few gotchas in between :-)).
Next step was to see those data in real time without writing any code and here comes #grafana from Grafana Labs. Used Crate.io database to send data from #nodejs to CrateDB. Next step was super simple and that was visualising data in Grafana (video in this post). We went a step further to provide alerts sent to Slack and Email without any code. They are all within Grafana. I went to TOML (https://en.wikipedia.org/wiki/TOML#:~:text=TOML%20is%20a%20file%20format,source%2C%20and%20receives%20community%20contributions) for opc tags and configurations.
End result was quite impressive. Helped them preparing some visuals in #powerbi that management people can use.
Lesson learned — You don’t need building things from scratch all the time. There are plenty of tools out there which can be integrated in few hours time. You just need to reach out to the right people.
It’s been a while coming back to the UK and upon reflecting to this experience, I have now planned to productise it a bit and add more features.
Very first thing I have done is security everywhere hearing IoT vulnerability everyday and a developed a message hub called EagleHub which is a layer on the top of #mqtt for communication from and to machines/devices (devices could be servers as well. Yes you heard it right. Anyone interested in real time data can use it).
Currently I am working on integration of #azure anomaly detection service integrated with SynapseML (https://microsoft.github.io/SynapseML/). Reason being , I wanted to pick anomalies using one parameter (for example temperature) and multiple parameters (temperature, speed, wind etc together can create issues). Once you are happy with a model, it can be deployed on premise or in the cloud for real time anomaly detection.
I am really excited on this journey and will be posting updates here.
If you have any suggestions, feedback and want me to show you something in motion, please feel free to reach out to me at rakesh@sketchmyview.com .
