Why Exception based Operational Systems are the Only Practical Way?
As you unify your operations the more and more you roll up data, the more you drown in data and events etc. It becomes very quickly impractical to digest this information at the rate it is coming, and then it gets ignored or situations missed.
This calls for different thinking and taking the users away from MONITORING to now been naturally made aware of a situation through presentation, of the condition.
Concepts like the Advanced Process Graphics are changing the thinking in the way data is presented so time to awareness to situations is dramatically reduced. This I will discuss in the future.
But the big point of this discussion is move to Exception Based thinking meaning that the system notifies the correct people of a condition. This reduces the data needing to be consumed by a operator, and reduces the data that needs to be sent over networks. The issue with post analysis like through historian it is not real-time, and means you has gathered all the data not the critical data.
Now the concept of Exception based is not new most instruments today have exception based handling in them. But that is at a too lower level for most operations, especially as the operational scope of a operation worker increases.
The system now calls for “agents” (objects) that represent a Asset, including the operations within that asset that it must perform. This agent therefore is not a data structure, it is model representation of the asset, which is “self aware” of it’s condition and it’s expectations. It is then able to raise alarms, and trigger escalations.
This is the concept of “self Aware’ you may be asking what is he talking about, but it is logical. Take yourself do you have doctors monitoring you and advising you when you have to go for check ups or health activities? No it is the other way around, you are aware of your health condition and when you feel unwell, you visit a doctor to investigate and take action. This is the concept of being “self Aware”.
So why cannot we effectively make the assets “self Aware” but naturally modeling this in your supervisory system/ operational system. Working in conjunction with your control system, which should be controlling the asset execution. You may say “well I do that in my control system” you may have a lot of the data, but do you have the full understanding of expectations relative to production for that asset, are you going to put your KPIs down at that level. No the natural division is the supervisory platform should provide a representation of the asset that worked closely with the control, instruments to understand the current condition, but then there is additional intelligence of the expectations of that asset from a performance, safety, environmental to perform.
The need to make the asset “Intelligent and self-aware” as close to source as possible, where for example the “oil/ gas well” is aware of it’s expectation and current state, the incoming data is VALIDATED, so the information in the well is “trusted”. So now the “well” can act on exception. This allows the operations center to be able to manage much larger numbers of processes and assets, and where the “the situational awareness” is pre-emptive, vs monitoring.
Intelligent objects can also have inbuilt workflows to automatically guide the operator to corrective actions and procedures e.g.
· Alarm/ safety condition escalation when an operator does not acknowledge in time the escalation continues through email, alerts and/or notifications
Again this is one of the big mis-understandings in the ArchestrA concepts is the fact an Object is not a data structure it is a representation of the asset of what ever it is representing, and is “living” in memory so can be intelligent to understand conditions from a number of inputs to create a say “an unsafe condition”. Now if you have these objects built as standards you are able to evolve them with new capability and roll that new innovation out in a consistent manner.