Predictive analysis is Key to Effective Decision Making, and Future of Industrial Operations

Last week I toured the east coast of Australia, engaging with Food and Beverage customers about opportunity. In Australia like US, the focus is on becoming competitive and effective to deliver high value products, in a more timely manner.

At the ARC conference in Orlando in Feb, we saw increased discussion around predictive analyst ices.

The whole objective is to move from the "as is" (which is traditional alarms) to the "to be" state, so the impact of time to detect, time to react to a conditions, has less impact on the situation, reducing cost, risk.

The diagram below I have shown before but is still one examples of why we must start a shift to predict through patterns, and relationship with between variables to see conditions developing before they develop into risk conditions.

As can be seen from the ARC diagram below most of the systems we build today are on now (dashboards) and past (Reports), but the edge is going come from the future, often learning from past to predict the future. The user must be looking to the near future to allow decisions this could minutes to 12 hours into the future with high predictability so he can ask “what ifs” and help make rapid decisions.

 All the areas in the orange provide new insights into the direction of the operations or their situation so they can predict, what you they do now. This requires lots of information, and analytics to predict models, and patterns.

Good example is the partner company Asset Insights from Pattern Discoveries Technologies.:

Taking known events around assets, combining these as context to determine patterns from within the historian data to bring value a repetitive pattern that can now be used to “shift to the To Be” state. So decisions are made faster and in more confidence, with insight into what to do next.

Software like this will grow fast around the core big data of such items as historians, alarms, and asset, energy systems.