What is the root cause of this plant upset? Is instant online analysis available?

  • "SOE of Control system or alarm logger lists the sequence of events, but does not help identifying the Root Cause" We have set up an investigation committee to find out the Root Cause and it may even take Weeks.
  • "Process alarms & parameters alone may not be enough. Many times we need to use our experience in determining Root Cause. We have set up an investigation committee it may even take weeks".
  • SAMA RCA is solution to the above issues raised frequently in Process Industries, Power plants as well as Building automation.
  • SAMA RCA is online software with instant Root Cause analysis solution. Root Cause Analysis (RCA) package gives instant Root Cause of not only a shutdown, but also of an upset condition.

Root Cause Identification

Decision Tree mechanism based on Result Tag (trip or upset), Condition tags and Cause tags as well as statistical calculations.

The Decision tree has to be supplemented in some cases:

Cause tag or condition tag sensor fails More than one Cause tag Hidden causes, hardware failure Then, we use the most probable root cause

Most Probable reason

If auto mode is chosen, auto relevance calculated (self-learnt automatically) is used to calculate adjusted probability.

In Manual mode, Relevance and weightage (based on your experience) for each cause tag can be assigned upfront. Based on these, it self learns and calculates adjusted probability.


Design View is for defining developing trees with result tag, condition tags and cause tags and interconnecting them.

Normal View shows current Plant status. It helps operator and plant staff in identifying possible trips or upset conditions.

In Root Cause view, the Root cause, which fulfilled the conditions, is highlighted, along with the path.

Historical Root Causes can be selected and viewed as they are stored.

Overview gives glance of all defined trees and their current status.

Results can be viewed based on

Probability over Feedback status : Root cause is identified based on Probability. When probability is equal among the condition satisfying causes, Root Cause determined based on feedback status among the multiple causes.

Feedback status over probability : Root cause is identified based on the actual feedback. If no feedback cause is found or when multiple causes based on feedback is available, then the cause with highest probability, but satisfying conditions is highlighted as root cause.

User can also select the root cause manually which can also be used in Probability calculation.


Relevance A is Cause and B is Result. If B arrives within pre-defined time window of A, then it is relevant. SAMA RCA calculated Relevance in Semi-Automatic and Automatic methods. In semi-automatic, user can set initial relevance and relative weightage of one cause tag over other.

Other important parameters indicated include confidence level, Learn Factor.

Data Sources

Data collection from AE source : OPC AE or direct database connectivity of Historians or printer ports

Data collection from DA source : OPC DA or direct database connectivity of Historians


Windows 2008 server, windows XP, Windows 7. Windows 2008 SQL server.