Yokogawa Electric Corporation is taking industrial automation a step closer to autonomy with the launch of CENTUM R7.02, its latest integrated production control system, which comes with artificial intelligence (AI) capabilities as standard.
Released on November 19, the upgraded system introduces two key AI functions — AI Control and AI Estimation — designed to improve plant stability, boost productivity, reduce operational workloads and support faster, data-driven decision-making.
The launch also marks a significant change in the company's flagship control system's identity. As CENTUM enters a new era of autonomous operations, Yokogawa is dropping the 'VP' from its former product name, CENTUM VP. The 'VP' stood for the company's 'VigilantPlant' concept.
CENTUM R7.02 is part of Yokogawa's OpreX Control and Safety System portfolio and builds on the company's decades-long experience in industrial process control.
The new AI Control function allows AI control models generated through autonomous control AI to run directly on CENTUM as a standard feature. Meanwhile, AI Estimation uses existing process data to estimate product quality indicators and operating conditions, helping plant operators make more informed decisions.
The development comes at a time when manufacturers are grappling with increasingly complex facilities, tighter environmental regulations, labour shortages and the challenge of transferring critical technical expertise. At the same time, they face mounting pressure to improve productivity and energy efficiency without compromising profitability.
Yokogawa believes AI-powered autonomous operations can help address these challenges.
First introduced in 1975 as the world's first distributed control system (DCS), CENTUM has supported stable operations and improved efficiency at more than 30,000 plants worldwide.
The latest upgrade builds on Yokogawa's growing track record in applying AI to industrial process control, particularly in areas where conventional automation struggles and operators must make frequent manual adjustments.
Its autonomous control AI can manage processes involving multiple competing objectives and respond directly to changing operating conditions, even when disturbances cause significant fluctuations in plant processes.
A major milestone came in 2022, when Yokogawa's AI solution achieved autonomous control of a chemical plant for 35 consecutive days — a world first. The plant has since maintained stable operations for more than four years, while delivering energy savings and reducing its environmental impact.
CENTUM R7.02 builds on this experience, bringing AI capabilities into the company's established control system platform.
The first major addition, AI Control, allows AI-generated control models to operate within CENTUM's highly reliable system environment, which is designed to ensure long-term operational continuity.
The models can be downloaded and deployed through the CENTUM engineering environment. Plant operators can then monitor their performance using the familiar CENTUM interface, including graphics, trends, alarms and faceplates.
Before deployment, the AI models undergo repeated learning cycles using a plant simulator that replicates actual process behaviour. This helps identify the optimal control model. Customers then evaluate its reliability and validity before it is introduced into plant operations alongside functions designed to ensure safety.
The second addition, AI Estimation, is aimed at improving operational visibility and decision-making.
The AI-powered soft sensor function estimates critical process values, including component concentrations, using existing data such as temperature, flow rates and pressure.
These indicators are essential for monitoring product quality and process conditions but are often difficult to measure continuously in real time. Some require laboratory testing.
AI Estimation uses machine-learning models to predict these values and identify how individual process variables contribute to the results. Operators can access the estimates and related information through the CENTUM monitoring environment, helping them assess plant conditions, maintain product quality and optimise operations.
Mitsuhiro Yamamoto, corporate officer and head of the Systems Business Division at Yokogawa Electric remarked, “CENTUM R7.02, with its two new AI capabilities, represents the first step toward CENTUM Future, which we announced last year on the 50th anniversary of CENTUM as we seek to advance plant operations from automation to autonomy.
"Combining advanced AI-based control with CENTUM’s proven reliability, stability, and long-term continuity will help customers achieve stable operations while reducing operational workloads. We will continue to contribute to the growth of our customers’ businesses, enhancement of their corporate value, and achievement of their sustainability goals.”
Yokogawa defines autonomous control AI as artificial intelligence capable of independently determining optimal control methods, with sufficient robustness to handle, to a certain extent, situations it has not previously encountered.
One such technology is Factorial Kernel Dynamic Policy Programming (FKDPP), a reinforcement-learning AI algorithm jointly developed by Yokogawa Electric and the Nara Institute of Science and Technology (NAIST). The technology received the Prime Minister's Prize at the 2023 Japan Industrial Technology Awards.
The company's AI control technology is designed to directly adjust manipulated variables in industrial processes, with its world-first achievement in chemical plant operations documented in a Yokogawa Electric survey conducted in February 2022.
CENTUM R7.02 targets a wide range of process industries, including refining and petrochemicals, chemicals, power, gas, pulp and paper, pharmaceuticals, food, iron and steel, metals, water and wastewater, and cement.
Its applications include plant and manufacturing equipment monitoring, as well as automatic process control.
With AI now embedded as a standard capability, Yokogawa is positioning CENTUM R7.02 as a key step in its broader ambition to redefine industrial control — moving beyond conventional automation towards plants capable of increasingly autonomous, stable and efficient operations.