Whilst you might have data coming from multiple sources in an Industrial setting such as sensors, equipment, maintenance systems, production platforms and enterprise applications - this doesn’t necessarily provide you with meaningful information.
You need to link the information and understand it within the correct context.
The Context Issue
Let’s say you have a simple example of monitoring equipment.
You have the following data from a sensor:
Asset: Machine-04
Temperature: 82°C
Time: 10:45
Whilst you have some information, you still have some questions:
Is the machine running under normal conditions?
What is the expected temperature range?
Have there been any changes in temperature?
Has there been any recent maintenance?
Are there any other measurements to back this up?
Without the correct context it can be difficult for an AI to understand what is going on.
What Information Should Be Linked?
By linking the following information, you can get a more useful view of your operations.
Asset: What machine/component is it?
Location: Where is the equipment?
Sensor Data: What is the condition?
History: How does this compare to previous data?
Operational History: What do the maintenance and production records show?
Constraints: What are the constraints?
A Simple Flow Of Information
Industrial Data Sources
|
Data Identification
|
Context & Validation
|
AI Analysis
|
Insights & Recommendations
|
Human Review / Approved Action
|
All of these steps will help to provide quality output.
Why Validation Is Important
Industrial data can suffer from missing data, non-uniform units, incorrect time stamps and conflicting equipment names.
This should be addressed and not just combined. If there is uncertain information then this should be highlighted.
The Big Picture
Industrial AI isn’t just about having a model that provides an output. It is about ensuring that the information can be linked to the correct assets, operating conditions and business processes.
This is a very important aspect when building decision making systems with AI.
Find out more about the AI Decision Engine:
https://apertureventurestudio.com/ai-decision-engine/
Top comments (1)
suggestion and discussion is welcome