Resilient Operations for Industrial Processes

Industrial plants are becoming increasingly autonomous, transforming the role of human operators from direct process control to higher-level supervision and strategic decision-making. In the future, operators will oversee multiple facilities, validate automated decisions, and intervene in rare or complex situations where automation reaches its limits. However, despite technological advances, human interventions will remain indispensable in critical, uncertain, or safety-related conditions.

This shift introduces a key challenge: the automation paradox. As systems become more autonomous, human operators are needed less frequently, but when they are, it will be in high-stakes and time-critical situations. These events often occur under abnormal process conditions, requiring rapid, precise decisions from operators who may no longer have continuous hands-on experience. Another key challenge for operators is to continuously monitor and understand the different automation states, analyse the data and models they rely on, and adjust them based on their expert knowledge of the controlled process. This can result in information overload, cognitive overload, and increased recovery time during system failures which will directly affect plant resilience, safety, and efficiency.
From a user and customer perspective, i.e. industrial operators, control room engineers, and plant owners, the current systems lack adequate mechanisms to support and guide human intervention when automated process control is independently unable to resolve the criticality. Today’s automation frameworks provide detailed data and control interfaces, but they rarely integrate situational explanations, adaptive assistance, or human–machine interaction strategies tailored to operators’ needs. Consequently, operational downtime and recovery costs increase while operator confidence and trust in autonomy decline.

The RESOPS project addresses this critical gap by investigating how resolutions during critical situations should be designed to support resilient operations in highly autonomous environments. Specifically, the project will explore:

• How human operators interact and collaborate with AI-driven Intelligence and automated process control systems under critical situations, and
• How these interactions and transitions of control can be designed to mitigate skill degradation and sustain human and machine competence and trust over time.

To ensure safe, efficient, and trusted operation, the next generation of industrial systems must be designed around resilient human–AI-Automation collaboration.