Digital twins and AI support for urea plants

Since 2016, Toyo Engineering Corporation (TOYO) has developed DX-PLANT™, a digital transformation platform for industrial plants. Through this platform, TOYO has introduced digital solutions supporting engineering, operations, maintenance, and business functions. For urea plants, these solutions include process performance monitoring, anomaly visualisation, and simulator-based operational analysis, drawing on TOYO’s process engineering expertise.

TOYO’s DX-PLANT™ platform and related digital solutions have been presented in previous issues of Nitrogen+Syngas, including the November–December 2020 issue, which outlined the development of TOYO’s digital services.¹

A core technology within this framework is PMOS™ (Plant Monitoring & Optimisation System), TOYO’s digital twin solution for urea plants. By combining real plant operating data with plant-wide process simulation, PMOS™ delivers engineering insight into plant performance and operating conditions. More recently, TOYO has extended the practical application of these technologies through technical advisory services for operating urea plants. Since August 2024, TOYO has provided technical advisory services for the urea plant operated by Yacimientos Petrolíferos Fiscales Bolivianos (YPFB) using DX-PLANT™. This service combines digital technologies with TOYO’s expertise as a urea process licensor, creating an environment in which operational data, simulation results, performance indicators, and engineering evaluations are continuously generated and accumulated. As digital twin applications produce growing volumes of plant data, simulation results, performance indicators, and technical reports, the next challenge is to use this information more effectively.

At the same time, advances in artificial intelligence (AI), particularly generative AI, are creating new opportunities to support information summarisation, report preparation, and the application of accumulated engineering knowledge. Building on its digital twin platform, TOYO has begun applying AI technologies to enhance technical advisory services and operational evaluation activities.

This article describes TOYO’s progression from digital twin-based monitoring to AI-assisted operational support for urea plants, and considers future opportunities to integrate digital twin technologies, engineering expertise, and AI-enabled information utilisation.

PMOS™ as a digital twin for urea plants

PMOS™ is TOYO’s digital twin solution for urea plants, integrating real plant operating data with a plant-wide process simulation model. As described in a previous publication,¹ the system continuously replicates plant operating conditions and estimates process information that cannot be obtained directly from conventional distributed control system (DCS) measurements (Fig. 1).

By combining operating data with process simulation, PMOS™ provides engineering information including:

• material balances;
• stream compositions;
• process conversion;
• utility consumption;
• equipment-performance indicators.

These outputs allow both plant owners and TOYO engineers to assess plant conditions from a process-engineering perspective, rather than relying solely on measured operating variables.

The availability of this information supports a more comprehensive understanding of plant performance and operating conditions. Operational data, simulation results, performance indicators, and engineering evaluations are continuously generated and accumulated, creating a digital representation of plant operation. This digital twin environment provides a foundation for advanced operational analysis and future digital applications.

From digital twin monitoring to future optimisation

Although PMOS™ was initially developed as a digital twin platform for monitoring and performance assessment, the information it generates may also support future operational optimisation.

TOYO has previously developed advanced process control (APC) applications for urea plants. APC uses key process indicators, such as the N/C ratio (NH3 /CO2 molar ratio), together with process measurements including pressure, temperature, and flow rate. Based on optimisation calculations, the system determines appropriate control actions to maintain optimal operating conditions and improve overall plant efficiency (Fig. 2).

TOYO is studying how the process variables and performance indicators generated by PMOS™ can be used more effectively in future operational-support applications. One potential approach is to integrate PMOS™ outputs with APC.

Conventional APC applications are primarily based on direct process measurements. While these variables are effective in maintaining stable operation, they may not always provide a complete view of overall process performance. PMOS™ supplements these measurements with engineering information on process conditions and material balances in areas where direct measurement is unavailable. It also provides plant-wide indicators, including process conversion, utility consumption, and equipment-performance trends.

This broader process perspective could complement the conventional measurements used by APC. In turn, APC may be able to identify changes in process performance at an earlier stage and support control actions that reflect overall plant condition, rather than individual measurements in isolation.

Further development and validation are required. However, combining digital twin technologies with advanced-control approaches such as APC may create new opportunities to improve plant stability, reduce energy consumption, and further optimise urea plant operations.

AI-assisted reporting and knowledge utilisation

As digital twin monitoring and technical advisory activities continue, substantial volumes of operational data, simulation results, plant-event records, and engineering evaluations accumulate over time. Using this information efficiently has become an increasingly important challenge.

While digital twin technologies offer detailed insight into plant conditions, the growing volume of operational and engineering information can increase the time required for review, interpretation, and reporting. Building on TOYO’s previous work in AI-assisted reporting, generative AI has now been applied in technical advisory services for operating urea plants.

In the technical advisory service for YPFB’s urea plant, operational information accumulated through daily reports is automatically summarised to support the preparation of weekly technical reports. AI assists engineers by organising operating information, identifying key changes, and preparing report drafts for engineering review. This enables engineers to devote less time to routine information processing and more time to technical evaluation and engineering analysis.

Practical experience has also demonstrated benefits beyond report preparation. AI-assisted analysis can rapidly identify historical operating records associated with non-routine conditions that may warrant attention from a stress corrosion cracking (SCC) risk perspective. Identifying such conditions traditionally requires experienced process engineers to review large volumes of daily reports and operational records – an increasingly demanding task as information volumes grow.

By automatically extracting and highlighting potentially relevant events, AI helps engineers focus on issues requiring further technical evaluation. Rather than replacing engineering judgement, AI acts as a productivity tool that supports faster and more consistent review of operational information.

AI can also assist with trend identification by linking operational events with PMOS™ performance indicators and historical engineering knowledge. This can help engineers evaluate plant conditions more efficiently and identify emerging operational issues at an earlier stage.

In parallel, TOYO is developing retrieval-augmented generation (RAG)-based solutions. In this approach, AI retrieves relevant technical information before generating responses or draft comments, combining operational information, PMOS™ analysis results, and accumulated engineering knowledge.

Over many years of urea process licensing, plant design, commissioning support, technical services, operational reviews, troubleshooting, and improvement studies, TOYO has accumulated extensive knowledge assets. RAG-based AI can retrieve relevant information from technical reports, troubleshooting records, operating manuals, and previous advisory activities, helping engineers prepare technically grounded recommendations and advisory comments.

This capability can reduce the time required to locate relevant historical information and support more consistent use of knowledge accumulated across previous projects and operating experience. It therefore has the potential to improve the consistency of technical support and strengthen the use of organisational knowledge across advisory activities. Fig. 3 illustrates the concept of AI-assisted reporting and knowledge utilisation.

Operational data, PMOS™ results, historical reports, and troubleshooting records are processed by AI to support summarisation and knowledge retrieval. The resulting outputs are reviewed by engineers and used to support technical reporting and the identification of improvement opportunities.

The combination of AI-assisted reporting and knowledge utilisation represents an important step in the evolution of technical advisory services. Through this approach, TOYO aims to enhance both the efficiency and quality of technical support while enabling faster identification of operational risks and improvement opportunities. These developments illustrate how AI can function as a practical engineering-support tool in plant advisory activities, complementing rather than replacing the knowledge and judgement of experienced process engineers.

Future perspective: From monitoring to optimisation

TOYO’s digitalisation strategy for urea plants can be viewed as a stepwise progression from data collection and monitoring to operational visibility enabled by digital twin technologies, AI-assisted engineering and operational support, and ultimately more integrated operational intelligence.

Today, digital twin technologies such as PMOS™ provide process-oriented information that improves visibility of plant conditions. At the same time, AI technologies help engineers organise operational information, use accumulated knowledge, and prepare technical recommendations more efficiently.

Looking ahead, TOYO aims to expand the integration of plant operating data, PMOS™ analysis results, AI-assisted information utilisation, and accumulated process expertise. This integration can support more advanced technical advisory services, including more systematic reviews of operating conditions, improved use of historical knowledge, and more effective identification of improvement opportunities.

Over the longer term, the combination of digital twin technology, AI-assisted information utilisation, and advanced-control technologies such as APC could create further opportunities for operational optimisation. Together, these technologies can contribute to more stable, efficient, and optimised urea plant operation.

Conclusion

This article has described TOYO’s evolution from digital twin-based monitoring to AI-assisted operational support for urea plants.

PMOS™ provides the digital twin foundation by combining real operating data with plant-wide process simulation, improving visibility of plant conditions and performance. The engineering information generated by PMOS™ supports plant evaluation and may create opportunities for future integration with advanced operational-support and optimisation technologies.

Building on this foundation, TOYO is applying generative AI in technical advisory services to support reporting, operational-information review, and knowledge utilisation. In the technical advisory service for YPFB’s urea plant, AI-assisted summarisation of daily operational information supports preparation of weekly technical reports.

By integrating digital twin technologies, AI-assisted engineering support, accumulated process expertise, and future optimisation technologies, TOYO aims to further enhance the safety, reliability, and efficiency of urea plant operation.

Reference

“TOYO adds new services and digital solutions to DX-PLANT,” Nitrogen+Syngas, No. 368, pp. 55–57 (Nov–Dec 2020).