AI-Driven Autonomous Optimization

Advanced Analytics of process data cannot be done without domain knowledge, Yokogawa can provide best practices of the industry with its domain knowledge of more than a century, and thus provide something more than only Data Analytics, we say it AI-Driven Autonomous Optimization. The offerings are not only highly efficient, realistic and actionable, but also more insightful to provide quality stabilization solutions, that can help companies stabilize and continuously improve the quality of their products.
Practically, any powerful AI algorithm alone is not sufficient to realize the true potential of customer data. Advanced analytics methods with domain and business knowledge are key to unleash the latent potential in plant data. 

Yokogawa Futrure Plant: More AI and Less Reliance on Human Skills

Figure_Yokogawa Future Plant: More AI And Less Reliance on Human Skills

Image Zoom


Smart And Agile Decision Making Without Experts

Take for example a condition monitoring dashboard which Yokogawa developed to optimize the running/decoking cycles in an olefin plant with 90 t/h average production rate. The earned benefits estimated at 170,000 USD/year. 

Advanced Monitoring Dashboard

Customer relied on TLE, TMT, HC Feed, in addition to other factors to determine the End of Run. Even with the same operator, the criteria for determination was not consistent, that has led to some furnaces being run too long (risk of overloading) or too short (opportunity loss).

Customer rely on one single true indicator to make informed decision on which furnace need to be prioritized and schedule for decoking.

Solutions Benefits
(1) Assuming we could increase plant productivity by 2 days each year. 
      Assuming 90 t/h average ethylene production rate, 40 USD/ton average margin.
      The estimated recurring USD value is (90 t/h * 24 h * 2 d) * 40 USD/ton
      = 170,000 USD/year

(2) Additional benefits by reducing furnace switchover one time/year.

Image Zoom

Below table summarizes our areas of expertise.

Summary for our areas of expertise

Who Should Buy this AI-Driven Autonomous Optimization Solution?

•    If your top management has set goals to utilize Artificial Intelligence; but you do not know how to start applying it to field to solve every day’s problems or to improve your process.
•    You were assigned the job of analyzing large amount or complicated process data, however you realized that there are many conditions that one need to take into consideration before dealing with such data.
•    R&D department has developed a new anomality detection system, but you need to test it before deployment with field data. 

    * Company names and product names described on this page are trademarks or registered trademarks of their respective companies.


Use cases

Case_1; Load Estimation on a Refrigeration Cycle


Petrochemical, Olefin Plant


Ethylene, Propylene, Butene-1


C2/C3 refrigeration cycle

Customer Issue

Identify heat exchangers that cause overload on refrigerant compressor:
Especially during summer, overload on compressor was caused by insufficient cooling on one of the heat exchangers. If such overload happens, customer can only reduce the feed and lose production. However, due to complications caused by delays in the heat exchange process, customer used to reduce feed too much to avoid compressor overload.
By identifying which heat exchanger causes compressor overload at an early time, customer will be able to avoid production loss.

Yokogawa Solutions

・Yokogawa Data Analysts analyzed process data and identified heat exchangers that contributed most to the compressor overload

・Yokogawa Data Scientists used process domain knowledge to express cooling demand on each heat exchanger.

・Yokogawa process experts designed operator-friendly dashboard visualization to help an operator avoid or mitigate the load on compressor


Case_2; Condition Monitoring of Heat Exchangers


Petrochemical, Olefin Plant


Ethylene, Propylene, Butene-1


Quench Tower Cooling of Cracked Gas


High quench tower overhead temperature during summer leading to feed reduction & Quench loop/exchangers frequent fouling:
Customer performed a series of changes which eventually lead to improvement of quench tower performance.
However, customer did not know which change was most relevant and contributed most to improvement in the performance.
By knowing and understanding which heat exchanger had the largest impact on quench tower temperature, customer can maintain the temperature with minimum maintenance cost.

Yokogawa Solutions

・Yokogawa Data Analysts identified few heat exchangers that contributed directly to improvement in the performance.

・Yokogawa Data Scientists created a machine learning model to monitor the health condition of the most relevant heat exchanger.


Möchten Sie weitere Informationen über unsere Mitarbeiter, Technologien und Lösungen?