Artificial intelligence for refiners

First define the problem, then artificial intelligence can be the way forward to higher operational efficiencies.

JANE REN
Atomiton

Viewed : 10247

Article Summary

Artificial intelligence (AI) is an umbrella term that covers many things. By its name, it is using computer programs to do what intelligent humans could do, and often doing it even better. Artificial intelligence is also called cognitive science, which is the most popular computer science course now in universities.

At its core, AI facilitates the ability of machines to learn from experience, adjust to new inputs and perform human-like tasks. Most examples encompass deep learning and natural language processing. Using these technologies, computers can be trained to accomplish specific tasks by processing large amounts of data and recognising patterns in the data.

In the 1950s, the foundation of AI research covered problem solving and symbolic methods before the US Department of Defence took interest in this type of work with the Defence Advanced Research Projects Agency (DARPA) projects in the 1970s. They had some early success with DARPA, producing intelligent personal assistants in 2003, long before Siri, Alexa or Cortana were household names. This early work paved the way for the automation and formal reasoning that we see in computers today, including decision support systems and smart search systems that can be designed to complement and augment human abilities.

Five attributes of AI
The cognitive tasks of AI can be divided into five categories: perception, learning, forecasting, reasoning, and coordinating. With perception, AI can understand the environment with sensing, and detect and recognise occurrences; is that smell a fuel leak? From this it can learn by synthesising that information into knowledge; this could be learning the relationship between temperature set points and distillate yield.

You begin to extract value from this data by being able to forecast with high precision and simulate outcomes such as the diesel fuel demand at your terminal next week. When it comes to solving logical problems, or reasoning, AI can make decisions or suggest the best solutions; given what I know, what is the optimal distribution of my products at different terminal sites?

Finally, coordinating, or what economists call ‘playing games’, is behaving with the assumption of other autonomous agents and responding rationally.

Despite the expanding range of problems AI can solve, there is one thing which no AI program has been able to replace humans in: defining the problem itself.

Given the obvious benefits that can be derived from adopting AI, what are the challenges that downstream oil and gas companies face when they embark on a programme? One of the biggest mistakes that companies make is that they embark on AI without first defining the problem. They collect lots of data, but do not know what to do with it, since they do not know what problem they are trying to solve by collecting all this data. Other industries have made this mistake before. A familiar example is the auto industry, which spent fortunes on telemetry programmes, collected terabytes of data, and have not yet figured out what to do with them. That fervour did not save the auto industry from its 2008 crisis, nor is it protecting the industry from disruptive forces.

Machine learning
One of the most popular subcategories today is machine learning. In fact, machine learning has become so popular that many people equate machine learning to AI. In fact, machine learning is only a small part of AI.

Machine learning is popular because it overcomes scientific unknowns through large quantities of historical data, and hence has made fortunes for companies that in the past found their data too complex to interpret (Google, Amazon, Facebook). Machine learning is based on pattern recognition, and machine learning methods consider all data as either inputs (features) or outputs (prediction). Multiple inputs are fed into an algorithm that produces an output. If the output does not match the actual data, the algorithm is tweaked to do better next time. This is called training in machine learning.

The IT world has been using machine learning for many years, in Amazon’s recommendation engine, Gmail’s spam filters, and Google’s search engine ranking algorithm.

Because machine learning relies on large quantities of data about the same subject, it is better at very focused problems and parameters, such as what is the relationship between vibration and engine failure?

Machine learning behaves poorly when the problem is a system problem with more complexity, such as a refining process or a logistics supply chain for oil that has many moving parts, which prevents repeating patterns.

It can also struggle when most of the information is domain specific, such as the pressure setting on the steam boiler that has a certain relationship with the steam energy generated and subsequently the processes in the distillation column. Such domain-specific information from the data cannot be utilised unless an engineer or data scientist has spent time to structure and correlate the data to correctly represent the relationship between them; this is something that machine learning cannot replace. The cost of this manual work is often ignored when companies want to train their data. They end up not having meaningful conclusions.

Another problem occurs when time and sequence are important. Most machine learning programs do not incorporate time based patterns. For example, the best way to predict the loading queue at the terminal in the next hour is to count the current queue length. Fuel demand estimates at a retail fuel station require information such as which month of the year and which day of the week it is in order to predict more accurately.
This is where time series come in. The central point that differentiates time series problems from most other statistical problems is that in a time series observations are not mutually independent. Rather a single chance event may affect all later data points.

Yet, existing time series technology alone does not solve all the new problems either. Enterprises are trying to aggregate and store all data in time series format, which understands time, but misses all domain correlations. This correlation across the domain of operations is critical for gaining contextual intelligence. Even though historian has been a familiar technology to first use, it is not sufficient.

Companies should consider the nature of the problems before they invest. You need the right AI tool based on the problem you have defined. Be sure to define the problem first, so that you can select the right tool. Do not make the auto industry’s mistake.

AI is already making inroads into the downstream sector. The oil and gas industry is known for its long supply chain and complex activities. To manage cost and ensure quality in today’s oil price market, supply chain transparency and asset performance offer the biggest returns on investment. Digitised supply chains gain insight from across terminals to better forecast demand and plan supply, while smart assets improve asset productivity and reduce energy costs.