System resilience denotes the capacity to uphold desired system performance in the face of disruptions. Evaluating the resilience of a process system necessitates a thorough consideration of the intricate interplay between its components and the pivotal role of process parameters in reflecting the repercussions of disruptions on the system. This paper introduces an integrated methodology that takes into account component interactions and leverages process data for the resilience assessment of a process system. The proposed methodology comprises four key components: system structure analysis, disruption impacts analysis, process simulation, and resilience assessment. Firstly, the system structure is meticulously scrutinized using a P-graph model. This analysis encompasses the assessment of the significance and interplay of components, as well as the evaluation of how component failures affect the system's overall processes. Secondly, a Markov model is devised to examine the state transition process of components and quantifies the maintenance time needed for failed components. Subsequently, a simulation model is formulated to acquire real-time process parameters in the presence of disruptive events. Finally, the system's performance response function (PRF) is derived from the normalization of these process parameters. Building upon this foundation, a resilience assessment is conducted with a focus on the PRF. To illustrate the effectiveness of this methodology, an LNG terminal system is employed as an exemplar.

@en

Chemical process systems are becoming more automated and complex, which leads to increased interaction and interdependence between the human and technical elements of process systems. This urges the need for updating the safety assessment method by treating “safety” as an emergent property of a system. Uncertainty comes together with complexity. To enhance system ability of dealing with uncertain disruptions, this paper proposes a quantitative resilience assessment method by modeling the failure propagation (initiated by a disruption) across the functional units of a system. The Functional Resonance Analysis Method (FRAM) is utilized to model the system operation to represent the relationship among its function units and to consider the interactions among human-technical factors. Then, a Cascading Failure Propagation Model (CFPM) is developed to quantify the fault propagation process and reflect the system functionality changes over time for resilience assessment. The proposed method is applied to a propane-feeding control system. The results show that it can help practitioners understand the process of fault propagation and risk increase, identify potential ways to design a more resilient system to respond to uncertain disruptions/attacks, and provide a real-time dynamic resilience profile to support decision-making.

@en

Recent years have seen the increasing complexity of engineered systems. Complexity and uncertainty also exist in engineered systems’ interactions with human operators, managers, and the organization. Resilience, focusing on a system's ability to anticipate, absorb, adapt to, and recover from disruptive situations, can provide an umbrella concept that covers reliability and risk-based thinking to ensure these complex systems' safety. This paper discusses the quantitative aspects of the notion of resilience. Like the quantitative risk assessment framework, a generic framework should be developed for quantitative resilience assessment. This paper proposes a framework based on a triplet resilience definition consisting of disruption, functionality, and performance. Uncertainty treatment is also considered. The proposed framework aims to answer the question of “resilience of what to what” and how it can be quantitively assessed.@en

Artificial Intelligence (AI) is a scientific subject investigating and developing theories, methods, technologies, and application systems to simulate, extend, and expand human intelligence. Research in AI includes robotics, language recognition, image recognition, natural language processing, and expert systems. As a comprehensive frontier technology, machine Learning (ML), an essential part of AI, has drawn widespread attention. This chapter discusses the application of ML in process safety and asset integrity management (AIM). It gives a brief literature review of the state-of-the-art of AI in process safety and AIM and describes the use of ML approaches in probabilistic risk assessment. The chapter also presents a conceptual model for big-data-driven AIM. Failure mode and effect analysis is used for damage mode identification and cause and effect characterization. Random forest regressor is an ensemble algorithm that comprises a set of decision trees built independently and with a different structure.@en

Chemical process industries are complex environments prone to accidents with potentially drastic consequences. This ever-growing sector is becoming increasingly complex to satisfy the needs of global energy markets and consumer supply chains. Parallel to this growth and increased complexity, the number of accidents where conventional safety practices have failed to prevent them rises. Many relate such shortcomings in the preventive approach to the unattainability of imagining all failure scenarios. Moreover, conventional safety's reliance on linear causal chains impairs its ability to comprehend the hazardous conditions arising from human, societal, and organizational factors' interaction with the technical system. A further complicating factor adding to the ambiguity of these interactions is the digitization of chemical process industries. Rapid technology integration, application of novel tools, and associated methods could increase the possibility of deviating from normal operating conditions and result in hazardous conditions. Many believe resilience-oriented concepts, if not a paradigm shift in safety, could improve the shortcomings of conventional safety practices. Resilience's approach to safety recognizes that all expected deviations and unknown hazards cannot be prevented; therefore, it increases system readiness and strengthens the capacity to absorb, adapt, and recover from adverse events to prevent catastrophic failures. This chapter aims to provide inclusive yet brief insights into resilience, why the chemical process industries should strive to become resilient, how resiliency is achieved, and how it may be measured by reviewing the state-of-the-art published literature concerned with resilience assessment.@en

Chemical process systems (CPSs) involve complex dynamic processes. Besides, the emergent and uncertain hazards and disruptions cannot be identified entirely and prevented by conventional methods. In those situations, resilience for CPSs plays an essential role in absorbing, adapting to disruptions, and restoring from damages. Systemic modeling plays a vital role in assessing resilience. A system-based analysis model, system-theoretic accident model, and process (STAMP) can provide a robust framework. This paper develops a comprehensive methodology to systematically model and assess system resilience. The STAMP is employed to model and analyze the system safety of a process system. A new method of dynamic resilience assessment is then proposed to quantify the resilience of the system. The proposed method is applied to the diesel oil hydrogenation system. The results show that it quantifies the resilience of complex process systems considering human and organizational factors in a dynamic manner.@en

Digital technologies have been reshaping how the process industries operate. The extensive use of physical and information digitalization called by ambitious revolution to redesign process plants significantly transforms the process safety landscape in the process industries. Digitalization depends on the reliable use of data. This becomes a new focus of process safety in digitalized process systems. Digitalization brings the opportunity to generate and collect digital operational data and reduce human operation for effective process monitoring and control for safety assurance. The system's capability of processing massive data becomes essential for process safety. Meanwhile, digitalization increases the complexity of human-computer interaction and consequently leads to new research and practical problems. With standalone processes connected to the Internet of Things, process plants become more attractive to terrorists. This inevitably invites cybersecurity concerns to process safety solutions. This chapter provides a brief overview of these benefits and issues.@en

Resilience is an emergent property of a system, which changes with various internal and external factors. Resilience is also a hidden property of a system that cannot be observed. Thus, experiments should be performed for a given system to measure its resilience. However, physical experiments are practically impossible. Inspired by the tensile test for the stress-strain curve in Material Science, this paper proposes a virtual experiment for measuring system resilience and applies it to a chemical process system. The physical parameters of system resilience of a process system are mapped to those of material resilience. A process system is viewed as a 'specimen' in this experiment. The system performance variation caused by disruptions is seen as the displacement of the specimen caused by the applied load. In absorption phase, the decrease speed of system performance is determined by the failure rate of components under disruptive condition. Response time, including fault diagnosis time and resource allocation time, is used to represent adaptation ability. Restoration ability depends on repair rate of components. For simplicity purpose, the proposed method is applied to resilience assessment of a release prevention barrier system used in the Chevron Richmond refinery crude unit and its associated upstream process.@en

Chemical process systems involve complex dynamic processes, and the state of the system often fluctuates during the production process. To ensure the continuation of production, these fluctuations are often ignored or processed online instead of shutting down the unit. However, the interdependence between components in the system is strong, and small fluctuations or faults will be propagated to downstream nodes in turn if the fluctuation is omitted or processed online. A large number of accident investigations prove that the system risk increments as the failure propagates. This may eventually cause the entire system to collapse, causing severe casualties, property losses, and environmental damage. However, little attention has been paid to this type of risk. To measure the dynamic risk profile considering the fluctuation of the production process, this paper proposes a new risk assessment model that integrates the system-theoretic accident model and process (STAMP) and the failure propagation model. Firstly, the STAMP is used to model and analyze the system safety of a process system. An approach is then developed to quantify the risk accumulation of the model based on the failure propagation model. The process of the Chevron Richmond refinery crude unit and its associated upstream process is used to demonstrate the application of the proposed approach.@en

Due to the rapid development of technology, process systems become dynamic, automated, and complex, resulting in the strong interdependence and interaction among components and ensuring system safety by conventional methods a challenge. Compared with traditional risk assessment methods, resilience assessment is a more appropriate method for ensuring the safety of process systems under uncertain disruptions. Resilience refers to absorbing and adapting to changing conditions and recovering from disruptions. This paper presents a comprehensive assessment model that combines the catastrophe theory (CT) with the dynamic Bayesian network (DBN) to measure dynamic resilience. Firstly, the CT is employed to quantify the intensity of disruptions. Subsequently, the performance response function (PRF) of the system is determined by DBN. A resilience metric is then introduced to measure system resilience under uncertain disruptions. The method is demonstrated through a release prevention barrier system.@en