InTech

MAY-JUN 2019

Issue link: http://intechdigitalxp.isa.org/i/1124215

Contents of this Issue

Navigation

Page 23 of 55

24 INTECH MAY/JUNE 2019 WWW.ISA.ORG FACTORY AUTOMATION Analytics are necessary for foresight and oversight, and we also consider them beneficial for hindsight and insight. We are strong believers in using first principles-based ana - lytics tools in conjunction with emerging correlation-based analytics (also known as statistical or stochastic analytics) for situational awareness—a so-called "ensemble approach." First principles tools bring rigor due to their built-in under - standing of physics, chemistry, and dynamics, but at the cost of complexity and relatively high computation time. Corre- lation-based analytics suffer from lower fidelity without any guarantee of feasibility, but with the advantage of simplicity and speed of solution. DECISION MAKING In the same way that we recommend an ensemble approach to situational aware- ness, we also believe that decision mak- ing should be grounded on first princi- ples in conjunction with correlation-based tools as necessary. Decision making is about looking for answers. In an operat- ing plant seeking to improve performance, there are three main kinds of answer that can be sought: (a) forecasting ("what next?")—a judgment of what is likely to happen in the future based on knowledge of the past; (b) prediction ("what if?")— an estimate of what will happen in the future based on chang- es that could be made in the present; and (c) optimization ("what is best?")—an approach that answers the question, "Of all possible changes that can be made, which has the best eco- nomic outcome?" In the energy and chemical industry, there are many complex decisions to be made due to the vast number of variables that can be controlled and the large quantity of disturbances and constraints. Correlation-based decision tools are useful when accuracy is not as important as feasibility and when the answer lies within an already-experienced operating window. However, READINESS Frost & Sullivan estimated that process industries use less than 5 percent of the data that is collected—95 percent of the data is either siloed (used selectively), dark (unused), or not consistently in use. Problems of assigning context to data and poor quality have also been identified. To be ready for digitalization, the impediments to data utilization must be addressed: (a) data readiness (data suf- ficiency, data trust, data propagation, and data governance), (b) infrastructure readiness (physical infrastructure; secu- rity, privacy, and confidentiality; software infrastructure and cloud infrastructure), (c) consumption readiness, and (d) people readiness. A wise approach is to perform a readiness assessment and to tackle any readiness issues before starting (or perhaps in parallel with) a digitalization initiative. SITUATIONAL AWARENESS To take the right actions to improve a plant's operation, it is important to un- derstand the potential for improve- ment. Situational awareness is therefore a crucial step—knowing how the plant is and has been performing in absolute terms ("hindsight"), understanding where it has capacity for improvement versus its constraints and optimal capability ("insight"), predicting responses to changes ("foresight"), and assessing the success and value of such changes ("oversight"). Tools associated with hindsight and insight are largely visual—dashboards, BI tools, spreadsheets. These gain sig- nificant value when they align with goals, targets, and con- straints. Therefore, to present decision makers with valid in- formation in dashboards, for example, the right tools must be applied to each situation being analyzed. READINESS SITUATIONAL AWARENESS DECISION MAKING OPERATIONAL EXECUTION VALUE SUSTAINMENT Data Infrastructure Consumption People Hindsight Insight Foresight Oversight Forecasting ("What next?") Prediction ("What if?") Optimization ("What is best?") Best practices Advice-based action Closed-loop control Procedural automation Closed-loop optimization Goal monitoring Economic stewardship Knowledge management Management of change Technology and capability refresh Digitalization journey All digitalization initiatives go through this series of five steps.

Articles in this issue

Links on this page

Archives of this issue

view archives of InTech - MAY-JUN 2019