Threshold Based Valve monitoring using Monitoring and Alerting

Background: 

A production facility had numerous valves that were being opened and closed at varying rates and amounts. The subject matter expert at hand wanted to be alerted when the pressure values of these valves exceeded or dropped below a certain value as set by the subject matter expert.

Problem: 

Today, the majority of valve maintenance is being done according to a fixed schedule. Without AIR and CDF the pressure sensors on these valves would have to be manually read and then converted into an excel spreadsheet. Once the data has been extracted into excel the subject matter expert had to analyze this data and figure out if the thresholds were being breached. If they were then the maintenance is done in order to prevent accidents. As it can be inferred this process was very manual and not easily scalable. The subject matter expert  also wastes their time on tedious tasks as compared to actual important tasks.

Solution: 

Using AIR a data scientist is able to easily define and deploy a threshold model that generates alerts when the conditions are breached. The subject matter expert is then able to go into the AIR front end and setup monitoring tasks on the pressure time series data in CDF. AIR then alerts the subject matter expert when a condition is breached through email or SMS thereby automating the monitoring process and making maintenance a much more automated process. This model can also be scaled over many such valves or time series enabling a more holistic monitoring solution.