Making Artificial Intelligence into Augmented Intelligence: a Digitalization Approach for Ethylene Plants

Creating a conceptual mantra, a monomania, makes life a lot simpler. It permits one to eliminate any interfering complexity from the intrusion of other facts or contradictory ideas or even common sense. The one idea becomes talismanic. The words in which the idea is encased immediately delivers an inner meaning, outward significance, plan of action and morality. The fact that the idea may not comport with reality is beside the point.”

In today’s environment as fear of falling behind and not being competitive is ever present; the concept of what digitalization can bring may be in this same category of a conceptual mantra that is so attractive that common sense is not being applied. Analyzing data for the sake of analyzing data or proving that something can be modeled does not mean it is needed. Random application of machine learning tools to Ethylene plants is occurring and the promise of digitalization is being tarnished as value is not captured.

Understanding the problems that are occurring and the financial impact of these problems are the first steps. Then determining the analytics needed to understand the problem and its potential solutions should be considered. There is tremendous potential in digitization and digitalization. The purpose should be however to be able to provide the insight and augmented intelligence needed for engineers and operators to improve operations. The goal should be to have the insight needed to take additional appropriate action.

Ethylene plants tend to have considerable room for mathematical optimization through both fundamental models and machine learning models. Ingenero applies digitalization to Ethylene plants to enable improved operations. Ingenero customizes data agglomeration, analytics, and solution visualization for augmenting the intelligence of the operators and engineers to address the specific problems encountered that limit production. Examples of appropriate application of machine learning and result visualization for increasing throughput, yield and equipment availability will be shown for both ethylene furnaces and the separation train. The value is defined upfront and then artificial intelligence is applied with the intent of augmenting intelligence for value capture.