Chemical Process Safety Research Articles

Why do major chemical accidents still happen in China: Analysis from a process safety management perspective

As an important consideration in the chemical industry, chemical process safety has received notable attention in China. However, catastrophic chemical accidents still occur. To better understand why accidents continue to occur, this paper presented a diagnostic analysis of 14 major chemical accidents in China from 2012 to 2022 based on VOSviewer software. The authors analysed the correlation between the accident causation and their relationship with the safety management elements. The study observed that inferior process safety culture, intentional violation (rule-breaking) of procedure, inadequate safety training, and illegal operations were the most frequent causes of accidents. These causes highlighted the prominent gaps in PSM in China in the process safety culture, compliance with standards, the conduct of operations, process safety competency, and training & performance assurance. The results based on co-occurrence analysis indicated a strong correlation between these gaps in PSM. Enterprises should pay attention to collaborative management among them. These deficiencies in the enterprise’s PSM system showed that the essential defects in China’s chemical industry are a poor safety culture, inadequate accident investigation, inadequate training, and a lack of chemical safety personnel. The study recommended that the chemical industry establish superior process safety culture and competency for all personnel, monitor leading and lagging process safety indicators, apply inherent safety, and practice advanced safety management concepts. We hope that the findings can provide China’s perspectives and strengths for global chemical safety.

ENPS 101—Back to school for process safety

AbstractThis paper will provide an overview of the recently issued CCPS Process Safety for Engineers: An Introduction, Second Edition and how to use it. The objective of the text is to support teaching and learning of risk‐based process safety fundamentals as defined by the Center for Chemical Process Safety (CCPS). It also addresses the Accreditation Board for Engineering and Technology (ABET) Program Criteria for Chemical Engineering, which requires that chemical engineering curricula include consideration of the hazards relevant to chemical processes. This book can be used in meeting that requirement either as a curriculum for an undergraduate‐level process safety course or to provide process safety content into common chemical engineering courses. The team creating the book included professors who teach process safety. The book is founded on the CCPS risk‐based process safety elements and points to numerous tools to support the instructor in teaching and the engineer in conducting process safety fundamentals. Each chapter begins with a detailed process safety incident case study and includes what a new engineer might do, problem sets, numerous references, and a generic presentation file for use by professors.

An institutional modernization project in chemical engineering education in Brazil: Developing broader competencies for societal challenges

Contemporary societal challenges put in evidence the need to improve the hard and soft skills of chemical engineering students. To promote a more student-centered approach, active-based learning, and improved assessment strategies, the Brazilian government approved the so-called New National Curriculum Guidelines (NCG) for engineering courses. To comply with those guidelines, the Department of Chemical Engineering of the Polytechnic School of the University of São Paulo (USP) is currently developing an educational modernization process sponsored by the Fulbright Commission in Brazil, called Special Program for Modernization of Undergraduate Education (PMG). The project is based on three pillars of modernization: content (what), form (how), and infrastructure (where). This paper describes initiatives in each of those pillars: content and format changes in Chemical Reaction Engineering and Process Safety courses and the creation of new spaces for a student-centered approach (an innovative classroom layout and a makerspace). By gathering two concrete classroom experiences guided by a broader institutional educational policies (the PMG project and the NCG), this paper highlights that slight changes can lead to great improvements in the learning process, leading to more engagement in the development of hard skills while favoring improvements in soft skills, such as communication, team-based work, and critical thinking.

Power oscillation calibration and its application in power compensation reaction calorimetry

Thermodynamic parameters, including heat transfer coefficient and specific heat capacity of chemicals, are key information for the application of reaction calorimetry in chemical process safety and optimization. Conventional methods for parameter calibration used in commercial reaction calorimeters exhibit limited efficiency and accuracy. In particular, heat dissipation and calibrated parameters vary with temperature during calibration. The impact of these issues on measurement results has not been carefully considered so far. Therefore, this study presents a power oscillation calibration method, for estimating the heat dissipation coefficient at the reference temperature. By generating temperature oscillation and averaging the results of multiple measurement stages, the parameters can be determined, while the impact of parameter variation with temperature is inhibited. The comparison experiments confirm advantages of the proposed method. Finally, the application of the method in power compensation calorimetry is also demonstrated by experiments.

Attention-based long short-term memory fully convolutional network for chemical process fault diagnosis

A correct and timely fault diagnosis is important for improving the safety and reliability of chemical processes. With the advancement of big data technology, data-driven fault diagnosis methods are being extensively used and still have considerable potential. In recent years, methods based on deep neural networks have made significant breakthroughs, and fault diagnosis methods for industrial processes based on deep learning have attracted considerable research attention. Therefore, we propose a fusion deep-learning algorithm based on a fully convolutional neural network (FCN) to extract features and build models to correctly diagnose all types of faults. We use long short-term memory (LSTM) units to expand our proposed FCN so that our proposed deep learning model can better extract the time-domain features of chemical process data. We also introduce the attention mechanism into the model, aimed at highlighting the importance of features, which is significant for the fault diagnosis of chemical processes with many features. When applied to the benchmark Tennessee Eastman process, our proposed model exhibits impressive performance, demonstrating the effectiveness of the attention-based LSTM FCN in chemical process fault diagnosis.

An analysis of α-epichlorohydrin-water runaways

A curiosity began after plotting adiabatic reaction calorimeter (ARC) data from a homework problem placed in a textbook entitled Chemical Process Safety by Daniel A. Crowl and Joseph F. Louvar, 2011. The data were for a 50/50 mixture (by mass) of α − epichlorohydrin and water mixture. Two exotherms appeared. The first exotherm involved hydration and polymerization of α-epichlorohydrin. A second exotherm appears as well related to the decomposition of reactants and products. Modeling exotherm 1 as a direct hydration of α-epichlorohydrin to glycerin and HCl underpredicted the observed exotherm. Thus, what are the mystery reactions involved that released more energy than predicted by simply modeling hydration to glycerin? This paper reports the answers found to this mystery. A combination of 13 intermediate reactions is proposed.

Uncertainty Modeling of a Chemical System with a Flexible Node by Mapping the Fault Tree into the Response Surface Method

This paper elaborates three novel contributions in the field of chemical process safety. The first contribution is the identification and classification of chemical system variabilities into seven broad categories, namely, media, equipment, component, operator, procedural, management, and external (MECOPME). The identified variabilities lead to epistemic and aleatory types of uncertainties in the probabilistic safety analysis. To deal with the uncertainties caused due to the variabilities, a concept of the flexible node is proposed, which demands a failure probability in the flexible range of a lower level to a higher level instead of a fixed static probability. Since the existing techniques are not robust enough to handle the probability range, the classical fault tree is mapped into a statistically more reliable approach of the response surface method (RSM). The unique idea of using RSM in the failure analysis is demonstrated over the fault tree of an overtemperature scenario in a semipilot scale setup for the hydrogenation process and successfully evaluated over an industrial accident of the release prevention barrier scenario. The contour and surface plots of RSM reveal more information than the traditional approach of minimal cut sets. The statistical markers of RSM are a better substitute for the improvement index for sensitivity analysis. The proposed approach deals with chemical system variabilities and the lack of knowledge of exact occurrence probabilities more effectively.

Evolutions of the LOPA method to a fully quantified method

AbstractLOPA (Layer Of Protection Analysis) is a risk analysis method developed to highlight various layers of protection from the process design to the emergency response. In its basic conception, each barrier is associated with a Risk Reduction Factor (RRF) corresponding to the inverse of the probability of failure (e.g., RRF = 10 corresponds to a probability of failure on demand [PFD] = 0.1). In its most common form, the LOPA method uses orders of magnitude for the probability of failure and frequency values generally inspired by those reported by the center for chemical process safety books. Frequency of accidents then consists of multiplying the frequency of the initiating event by the PFD allocated to each barrier. The LOPA method is an attractive approach because, at first sight, it is relatively easy to implement and can be used by engineers who are not necessarily specialists in industrial risk analysis. In reality, this approach can quickly lead to an erroneous estimate of accident frequency. This article presents how LOPA can be applied in a quantified way and the expected benefits. A practical application example related to tank overflow will highlight the advantages of using a fully quantified method.

Fault Diagnosis of Unseen Modes in Chemical Processes Based on Labeling and Class Progressive Adversarial Learning

While deep neural network (DNN)-based fault diagnosis methods can monitor the faults that occur on the operating modes, they cannot perform well on the modes that are never experienced before. This limitation makes it challenging to ensure the production safety of chemical processes. In this article, this issue is formulated as domain generalization (DG), which aims to learn a universal fault diagnosis model from historical operating modes that can generalize well to unseen modes. Many existing DG approaches focus on learning a domain-invariant representation by aligning marginal distributions between domains, which ignore the conditional relationships and label information. Recently, some researches start to reduce the discrepancy of the class conditional distributions across domains, while the theoretical justifications of that are still missing. To address this issue, the theoretical analysis of DG is developed to investigate how to minimize the unseen domains’ risk, which is a theoretical guarantee that the DG methods can generalize well over unseen domains. This theory reveals that the unseen domain error can be bounded by the shift of the label and class conditional distributions across source domains. Then, this result motivates a novel labeling and class progressive adversarial learning (LCPAL) algorithm for fault diagnosis, which simultaneously controls the errors weighted based on label information and aligns the class conditional distributions between different historical operating modes, as well as reducing the adverse effect of the domain-specific feature. Empirical results on both the numerical example and the Tennessee Eastman process (TEP) demonstrate the effectiveness of the LCPAL approach.

Differences in perception of the importance of process safety indicators between experts in Iran and the West

Introduction: The importance of safety in high-risk industries such as oil and gas facilities has been reported previously. Process safety performance indicators can provide insight into improving the safety of process industries. This paper aims to rank the process safety indicators (metrics) by Fuzzy Best-Worst Method (FBWM) using the data gathered through a survey. Method: The study uses a structured approach considering the UK Health and Safety Executive (HSE), the Center for Chemical Process Safety (CCPS), and the IOGP (International Association of Oil and Gas Producers) recommendations and guidelines to generate an aggregate set of indicators. It calculates the level of importance of each indicator based on the opinions of experts from Iran and some Western countries. Results: The findings of the study demonstrate that some lagging indicators such as the number of times processes do not proceed as planned due to insufficient staff competence and the number of unexpected disruptions of the process due to failure in instrumentation and alarms are important in process industries in both Iran and Western countries. Western experts identified process safety incident severity rate as an important lagging indicator, whereas Iranian experts considered this as relatively unimportant. In addition, leading indicators such as sufficient process safety training and competency, the desired function of instrumentation and alarms, and proper management of fatigue risk play an important role in enhancing the safety performance of process industries. Experts in Iran viewed permit to work as an important leading indicator, while experts in the West focused on fatigue risk management. Practical Applications: The methodology used in the current study gives a good view to managers and safety professionals in regard to the most important indicators of process safety and allows them to focus more on important process safety indicators.

A conceptual framework integrating numerical simulation with system theory based method for quantitative explosion process hazard analysis

Explosion is a significant threat to chemical process safety, which often occurs suddenly with a wide range of impact, resulting in catastrophic disasters. Predictions of explosion hazards are required for regional safety assessment and disaster resilience enhancement of chemical industrial parks (CIPs). This paper, therefore, aims to develop a conceptual framework for systematically analyzing explosion evolutions and quantifying the potential hazards by combining system-theoretic process analysis (STPA) method with numerical simulation. The proposed framework consists of 5 steps: (i) establishment of hierarchical safe control structures (SCSs) of important chemical processing zones in the CIP, (ii) computational fluid dynamics (CFD) modeling for potential explosion evolutions in each zone by changing the examined parameters randomly, (iii) development of a convolutional neural network (CNN) prediction model through constant self-learning of CFD pressure field data, (iv) comprehensive assessment of blast damage by incorporating the outputs of the above numerical models into existing evaluation methods, (v) identification of unsafety control actions and causes, and safety constraints for the improvement of SCSs. Provided with monitoring data, the developed analysis architecture can predict explosion process hazards and recommend appropriate safety strategies in real time. This would service the multi-level requirements for explosion prevention and protection, supporting better-informed decision-making. The paper describes the concepts and implementation process of the method as a first step.

Fault monitoring for chemical processes using neighborhood embedding discriminative analysis

The importance of chemical process safety and the availability of abundant samples keep popularizing the wider application of data-driven fault monitoring techniques. With a goal of efficiently discovering the inconsistency between the online monitored sample and the normal samples, a novel fault monitoring algorithm called neighborhood embedding discriminative analysis (NEDA) is proposed, which can adaptively provide different latent feature generating mechanisms for different monitored samples so that the inherited inconsistency could be uncovered in a timely manner. Instead of extracting representative features from a dataset only given from the normal operating condition, the objective function designed for the NEDA algorithm additionally takes the online monitoring sample into account, and then timely generates but only one projecting vector to point out the specific inconsistency for the corresponding monitored sample. The NEDA algorithm aims to figure out a discriminative projection so that the neighborhood embedding error (NEE) corresponding to the online monitored sample could be maximized, while the NEE associated with the normal samples is minimized. Furthermore, the corresponding NEE for the monitored sample of current interest is employed as the indicator for fault monitoring purposes. As demonstrated through comparisons, the salient performance achieved by the proposed NEDA-based fault monitoring method in monitoring static as well as dynamic processes can be always guaranteed.

Safety Criteria for Solid-Liquid Heterogeneous Systems in Semibatch Reactors.

An exothermic reaction in a semibatch reactor can potentially cause thermal runaway due to evolved energy accumulation or a secondary reaction. This research aims to propose safety criteria for solid–liquid reactions in semibatch reactors. Simulation modeling was carried out to build thermal runaway criteria for solid–liquid reactions in semibatch reactors. A new model for the energy and mass balance of solid–liquid reactions was successfully established. Criteria for the safety boundary diagram and the temperature diagram were ameliorated for solid–liquid reactions. The results showed that the dissolution heat has a great influence on the thermal behavior of the reaction. Experiments to neutralize citric acid and sodium hydroxide were carried out to determine the critical parameters for the neutralization reaction using the temperature diagram criterion. The proposed criteria would be reasonably expected to provide some guidance for chemical process optimization and safety design for engineering.

Layout optimization for safety evaluation on LNG-fueled ship under an accidental fuel release using mixed-integer nonlinear programming

The Liquefied Natural Gas (LNG)-fueled ship is an alternative solution to deal with the Emission Control Area (ECA) issued by the International Maritime Organization (IMO). LNG fuel has eco-green characteristics since sulfur oxide and particulate matter are eliminated. The equipment sitting on the Fuel Gas Supply System (FGSS) layout needs a good arrangement to prevent the domino effect. This study’s scopes are to establish the feasible FGSS layout based on the equipment position and conduct an accidental LNG release that adopts the Computational Fluid Dynamics (CFD). Mixed-Integer Nonlinear Programming (MINLP) is provided as an optimization tool. The MINLP framework is written in the Python programming language. The minimum separation distance, deck size, and non-overlapping functions of equipment are included as the MINLP constraint. The separation distances are provided based on the Center of Chemical Process Safety (CCPS) recommendation. Also, several options of the separation distance are added that consist of 5.0, 6.0, and 7.0 m to observe the congestion effect. The total pipe length is considered as the MINLP objective that can be used to estimate the pipe cost. The proposed MINLP framework is designated to generate a feasible layout that has an adequate separation distance and minimum cost of the pipe. CFD results show that the increase of separation distance can reduce the gas accumulation effectively which eventually reduces the severity of other consequences. Finally, the feasible layout is proposed based on a series of several consequences under an accidental LNG release, such as gas dispersion, fire, and vapor cloud explosion.

Chemical process safety education in China: An overview and the way forward

The chemical process industry (CPI) in China is developing rapidly with installations becoming more complicated and integrated to meet people’s rising demands for chemical-related products. However, the fast-growing CPI has caused catastrophic consequences and bad social influence due to accidents occurred in the last decades, which has threatened its sustainable development. As one of the solutions, the Chinese government is promoting chemical process safety education to train interdisciplinary graduates who understand both chemical process and loss prevention, who are skilled in technology, and how to manage risk. In this paper, we reviewed the development of chemical process safety education in China by researching syllabuses of accredited undergraduate Chemical Engineering and Safety Engineering majors in higher education institutions, discussed the associated shortcomings by analyzing the current discipline construction of the newly established major Chemical Safety Engineering, including education methodologies, resources, faculties, curriculum provision, and professional accreditation. Based on the analysis results, suggestions were provided to encourage institutions to strengthen chemical process safety education, thereby inherently reducing human errors and consequently improving the safety of the entire CPI.

Analysis of Dangerous Factors and Preventive Measures in Chemical Process Safety Design

Analysis of Dangerous Factors and Preventive Measures in Chemical Process Safety Design

Chinese international process safety research: Collaborations, research trends, and intellectual basis

This article presents a bibliometric analysis and mapping of the Chinese process safety research, focusing on the contributions made in core process safety journals and on the influences of international collaborations and knowledge sources on the developments of this research domain. Collaboration networks, term co-occurrence networks, and co-citation network were analyzed to identify trends, patterns, and the knowledge distribution of the Chinese research on process safety. Work to data has been clustered mainly on safety of chemical processes, fire and explosion, and risk management and accidents. Chinese research contributions are concentrated in only few journals, while the corresponding intellectual base draws on the wider literature focused on understanding and modeling phenomena, and on the broader risk research literature, although to a lesser extent. While various foreign authors are highly cited by Chinese authors, only very few direct collaborations with international scholars are identified. The results are used as a basis for a discussion on future research directions and developments for the community. Increased focus on uncertainty treatment and handling of black swan events, risk evaluation and economic aspects of safety decisions, interorganizational risk management, road and maritime transport of hazardous substances, risk perception and communication, and integrated safety and security assessment, are highlighted as fruitful directions for future scholarship. It is hoped that the insights obtained from this work can facilitate new and consolidated collaborations, as well as further invigorate the Chinese process safety domain, ultimately contributing to improved safety performance of process industries in China and elsewhere.

A theoretical framework for data-driven artificial intelligence decision making for enhancing the asset integrity management system in the oil & gas sector

Asset integrity and reliability is one of the 20 elements of Process Safety Management (PSM) as defined by the Center for Chemical Process Safety (CCPS). We combine expert knowledge and data analytics (Artificial Intelligence, Machine Learning, and Keyword Analysis) to create a reaction network for Asset Integrity Management (AIM) and provide a theoretical and practical basis for handling uncertainty in large data sets such as company incident databases. The purpose of the current study is to control and minimize the total number of incidents that occur within an oil and gas operation by applying a multidisciplinary approach to explore and develop AIM. This systematic approach can improve AIM to better understand PSM as a whole and the underlying dynamics ever-present in the system. In this study, AIM is divided into 2 major groups – asset and human factors – and then, in order to get more detailed results, each group is divided into 9 and 5 subcategories, respectively. To analyze the relationships between the different factors of AIM, two score-based (Tabu and Hill Climbing) and one hybrid (Max-Min Hill-Climbing) Bayesian networks are used to develop one final viable solution. The findings of these techniques point towards the same results for reducing incident rates. Four factors related to assets, including construction, testing, inspection, and maintenance, account for more than half the incidents (54.78%). Additionally, there must be a greater emphasis placed on the impact of human factors as they are directly (23.58%) and indirectly (11.10%) responsible for accidents as well as other technological malfunctions. By focusing on AIM which is a key element of PSM, it will be possible to gain a better understanding of one of the most significant and problematic sources of risk in process safety.

The abnormal situation with reversal characteristic in chemical processes: Local monitoring and self-recovery

(1)Background. For chemical processes, due to external or internal factors, the system characteristic will not remain unchanged, and the reversal characteristic is a typical case where the system characteristic will change to the opposite characteristic. In this work, the reverse characteristic can be considered to consist of two opposite characteristics, in other words, the relationship between input and output consists of two opposite relationships. As the input increases or decreases, the system characteristics will change, when the input exceeds a certain value, the original system characteristic will even become opposite to the original. Then, for the control loop, the negative feedback becomes positive feedback and the system will be unstable, then, the working condition changes from normal to abnormal, which threatens the safety of chemical processes. Thus, it is urgent to monitor and eliminate this abnormal situation with reversal characteristic. However, traditional monitoring strategies cannot monitor the change of system characteristic and determine when the system characteristic is reversed, and there are still no effective process recovery methods aiming at restoring the system performance from such an abnormal situation. (2)Methods. To provide effective monitoring for the abnormal situation with reversal characteristic, the local monitoring strategy based on the long-short time window was proposed. Then, to avoid accidents and restore the system performance automatically, the first scheme of self-recovery, in which the dual control was used to improve system characteristic and the switching control was used to make the control loop be negative feedback but did not accumulate the changes of input, was proposed. Furthermore, to make up for the detection delay and increase the impact on the system, the changes of input were accumulated to form the second scheme of self-recovery. To verify the effectiveness of the proposed method, the abnormal situation with reversal characteristic was introduced and constructed in the Tennessee Eastman (TE) process. (3)Significant Findings. The case study on the numerical example and the TE process shows that the abnormal situation with reversal characteristic did exist and was well constructed; the long-short time window based local monitoring strategy can effectively reflect the change of the system characteristic and determine when the system characteristic was reversed; both the first and second schemes of self-recovery can eliminate this abnormal situation and restore the system performance automatically, but the second scheme had a better effect because it can reduce the switching times and make the system return to normal faster; and the dual control was an effective tool to improve the system performance. In addition, the discussion of window width showed that when the width of the long time window was twice that of the short time window, the performance of the self-recovery scheme was the best.

Monitoring a chemical reaction using pH measurements: An approach based on multiscale fractal analysis

The monitoring of (bio)-chemical reactions is mandatory to determine the operating conditions that maximize the performance of the processes and guarantee the process safety of chemical and biotechnological processes. However, it is not always possible to have monitoring systems that determine the reactive/product concentration in real-time. Therefore, offline analysis methods are commonly used, which are usually expensive and can generate considerable lag times. In this work, an indirect monitoring approach based on the fractal analysis of fluctuations of pH time series to monitor a chemical saponification reaction is presented. First, an experimental design was implemented to set the validation of the proposed monitoring system, evaluating the temperature and stirring speed effects on the conversion rate. Then, pH time series were analyzed using two multiscale methodologies (i.e., DFA and R/S analysis), which show that the pH series fluctuations exhibit fractal behavior. Also, two distinct regions can be identified, suggesting that the pH signals may contain information on the phenomena that intervene in the reaction process (i.e., transport phenomena). The dynamic scaling exponents show direct correlations with the reagent concentration, obtaining coefficients of determination R2> 0.98 in all experiments carried out. pH measurement is inexpensive, easy to implement, and can be obtained in real-time. The results suggest that multiscale analysis of pH series has a high potential for online monitoring of a class of reactive processes with low economic and computational costs.