A New Approach to HAZOP of Complex Chemical Processes

By Fabienne-Fariba Salimi, Ali Akbar Safavi, Leonhard Urbas and Frederic Salimi

A New Approach to HAZOP of Complex Chemical Processes provides practical methods to identify and categorize chemical process complexities systematically. The book follows a holistic assessment of risks and required safeguards which enables readers to define the boundaries of HAZOP 4.0 accurately. The book is written by authors who have decades of experience in advanced process monitoring and artificial intelligence to support HAZOP teams with a holistic dynamic simulation and multivariable monitoring of the complex systems, and to assess historical failure and accident data and information using artificial intelligence techniques in a user-friendly way.

• Presents complexity assessment and management to the conventional HAZOP ?
• Provides multivariable monitoring to dynamic simulation for a holistic hazard identification and process safeguards requirements ?
• Describes AI to support the HAZOP team with code-based requirements and historical failure and accident data ?
• Explains AI to find the dynamic behavior of process based on empirical data without the models with simplification assumptions

• Language: English
• Publisher: Elsevier Science
• Release date: May 3, 2023
• ISBN: 9780323906845

Fabienne-Fariba Salimi

Fabienne-Fariba Salimi has over 30 years of experience as a chemical process engineer. She has experience in Management and Chemical Process Safety engineering in oil and gas industries both offshore and onshore installations. She has particular expertise is dynamic simulation of chemical processes and accident consequence modelling in quantitative risk analysis. Her main qualifications were obtained in Iran and France and she is member of the Institute of Chemical Engineers, American Institute of Chemical Engineers, International Society of Automation, and Association of the French System Engineers. She is the co-developer of ADEPP and has been the Project Manager of ADEPP Academy since 1994.

Book preview

Section 1

Basic concepts and methodologies

Outline

Chapter 1 Introduction

Chapter 2 Hazard identification techniques

Chapter 3 HAZOP techniques*

Chapter 1

Introduction

Abstract

HAZOP is an effective hazard identification method. It is applied for all types of processes regardless of their level of complexity. This chapter aims to highlight the importance of considering the complexity of risk assessment of the major accident hazards.

Keywords

Hazard identification; complexity assessment; HAZOP 4.0; ISO 31000

1.1 Introduction

1.1.1 What do we mean by HAZOP complexity?

Since 1963 when the basis for Hazard & Operability was laid by ICI to perform the critical examination of the design of a new phenol plant to date, the HAZOP procedure has been matured and standardized by the major companies and IEC 61882 [1]. However, specialists still thrive to improve two inherent weaknesses of HAZOP which are:

1. Limitations of the heuristic method of HAZOP

Using experience to learn and improve is the formal definition of heuristic. The heuristic method is designed to help people capture a new way to see.

The heuristic systematizes our approach to optimization, but it may limit the spontaneous insight. Learning the step-by-step process can obscure the development of the HAZOP team’s capacity to see the whole.

HAZOP facilitator uses the heuristics to effectively conduct the brainstorming sessions, but if the HAZOP team is not encouraged to identify the hazards with holistic insight, then some of the major operability hazards may remain unforeseen.

2. Subjectivity of the HAZOP team

Subjectivity in a philosophical context has to do with a lack of objective reality.

In HAZOP Context, Subjectivity is a term to explain that judgment of the people about a given subject is biased or influenced by their perceptions, experiences, expectations, and personal or cultural understanding of, and beliefs about that subject rather than the truth and evidences.

Subjectivity is contrasted with the philosophy of objectivity, which is described as a view of truth or reality that is free of any individual’s biases, interpretations, feelings, and imaginings. Subjectivity and objectivity are usually seen as two directly opposing views; therefore an understanding of one usually influences that of the other.

Fig. 1.1 illustrates the risk assessment steps according to the ISO 31000 risk assessment framework [2]. For HAZOP, these steps can be interpreted as follows:

1. Establishing context is the first step of any risk assessment. HAZOP context is established by dividing the overall process system into the process subsystems called nodes. Boundaries of nodes are defined based on engineering judgment considering the complexity level of the configuration of the mechanical, instrumental components, and operational modes.

2. Operability hazards of each node are identified using the HAZOP guidewords checklist. HAZOP guidewords are formulated as the deviation of a process parameter, such as, High Pressure. System thinking calls for a life cycle vision. Therefore the HAZOP checklist should encompass all the operability hazards during normal operation, start-up, shutdown, maintenance, and any other relevant operational phases.

3. In conventional HAZOP, the causes and consequences of HAZOP scenarios are analyzed qualitatively. The heuristic approach calls for focusing on the single jeopardy causes. Generic or project-specific Failure Mode & Effect Analysis of the complex items could be applied as supporting study to identify the credible causes of process upsets.

4. Finally, the process safeguards in place are reviewed to ensure that the risk of HAZOP scenarios is reduced to as low as reasonably practicable (ALARP). Very often more than one independent protection layer (IPL) is implemented.

5. HAZOP worksheets are designed to clearly communicate the credible operability hazards during different normal operations modes, start-up, shutdown, degraded mode of operation, simultaneous operation, combined operations, maintenance/isolation/inspection, etc. to address all the operability, availability, reliability, and maintainability issues which have an impact on piping and instrumentation drawings.

6. Finally, all HAZOP actions shall be implemented and closed out. Effective close-out of the HAZOP actions is among the most important key performance indicators of the completeness of the process design and management of change procedures.

Figure 1.1 ISO 31000 risk assessment framework.

This approach is applied to all processes regardless of the degree of their complexity. However, for the complex processes, double jeopardies and domino effects could be credible major accident scenarios. Examples of the complex processes are given in Table 1.1.

Table 1.1

It should be noted that double Jeopardy is not a discussion on the simultaneous failure of the multiple protection layers. In fact, we implement the IPLs because we are aware of the possibility of simultaneous failure of the IPLs due to random, systemic, and common cause failures. For example, if both control and high-pressure trip of an isolatable section fail to protect against the overpressure, the pressure relief valve discharges the excess pressure to the safe location to prevent loss of containment and subsequent fire and explosion in the process area.

Domino effect in the context of HAZOP means that consequences of a deviation within a given node can be cumulated or propagated to the downstream and/or upstream nodes. The complex physical chemistry and reaction between the process fluids and the complex configuration and connectivity of the unit operations may cause domino effect.

Amplitude of deviation and response time of process could also be crucial in the evaluation of the consequences of a complex HAZOP scenario. For example, a conventional high-pressure trip cannot be a safeguard against the surge pressure following the water hammer in a hydraulic system.

For example, the surge pressures (water hammer) generated by the sudden changes in the velocity of flow in a system can be caused by the operation of valves and pumps or by expulsion of air from the piping system. The order of magnitude of the surge pressure is much higher than the gradual pressure build-up due to the causes such as pressure control failures. Therefore the surge pressures should be considered in a separate deviation scenario. A surge analysis should be performed to minimize the risk of the surge pressures by inherently safer design and proper operating procedures.

Neglecting the complexity of the HAZOP scenarios could lead to unforeseen major accidents. Major accidents of the BP Texas Refinery in 2005 are an example of double jeopardies. Fig. 1.2 illustrates the sequence of events in the major accident of the BP refinery [3].

Figure 1.2 Sequence of events resulted in major accident of BP Texas refinery in 2005 [3].

On March 23, 2005, a BP Texas City Refinery distillation tower experienced an overpressure event that caused a geyser-like release of highly flammable liquids and gases from a blowdown vent stack. Vapor clouds ignited, killing 15 workers and injuring 170 others. The accident also resulted in significant economic losses and was one of the most serious workplace disasters in the past two decades. The total cost of deaths and injuries, damage to refinery equipment, and lost production was estimated to be over $2 billion.

Subsequent investigative reports pointed to a strong cost-cutting focus by BP senior management that resulted in a lack of adequate training and supervision of filling and operating the distillation tower. Fundamental procedural errors led to overfilling the distillation tower, overheating, liquid release, and the subsequent explosion. Unit supervisors were absent during critical parts of the start-up, and unit operators failed to take effective action to control deviation from the process or to sound evacuation alarms after the pressure relief valves opened. The BP safety and quality assurance inspection and monitoring processes were absent and/or ineffective as a barrier to this failure chain. In addition, there was inadequate local, State, and Federal government safety oversight.

To control a distillation column, at least five control loops on the column should be implemented and optimized simultaneously. In the conventional HAZOP, the HAZOP rule set of One failure at the time assumption means that the process upsets and safeguards of the top and bottom of the column shall be considered separately while the distillation column is operating under the thermodynamic equilibrium which means that the deviation of bottom affects the top of column and vice versa. Neglecting the self-regulating thermodynamic of the distillation process could lead to underestimation or overestimation of the operability hazards and required process safeguards.

In BP major accident, liquid was pumped into the tower for almost 3 hours without any liquid being removed or any action taken to achieve the lower liquid level mandated by the start-up procedure. The distillation tower liquid level detection system was not designed to measure levels above a maximum height of 10 feet, providing no insight into off-nominal operational scenarios. The tower liquid level reached an estimated height of 138 feet immediately prior to the overpressure event. Therefore when the relief valves were opened, the liquid is discharged to the blowdown drum for a long time.

In other words, BP’s major accident was the outcome of multiple failures and domino effects. Table 1.2 summarizes the causes which are contributed to other major accidents [4].

Table 1.2

Causes of major accidents [4].

Source: From: COMAH Guidance: Case studies (hse.gov.uk).

HAZOP is the backbone of managing the process hazards. It provides the crucial insights for optimizing the process safeguarding strategy and operating procedures. Complex HAZOP scenarios are not numerous, but if they are overlooked or oversimplified, a major accident can occur.

The purpose of this book is to propose a novel method for complexity assessment and management of HAZOP studies. In this context the following subjects are covered:

1. Complexity of the HAZOP nodes (CLs)—In Chapter 6, we define complexity of nodes in terms of Number of equipment, Diversity, , and Nonlinearity of the unit operations and process fluids.

It should be noted that in the HAZOP context, a linear relationship between deviations and consequences means that there is a direct relationship between cause and effect which can be encoded in beautifully compact equations. The line is often understood as the shortest or most direct path from one point to another.

Inspired by risk graphs and calibration tables proposed by IEC 61511 [5] for safety integrity level (SIL) assessment, the authors propose a complexity calibration table to rank complexity characteristics of the HAZOP node. Complexity level of node is calculated and ranked as simple, complicated, complex, chaotic, and disorder according to the complexity criteria of the Cynefin