Abstract

China is currently the world’s largest exporter of firecrackers, with many production enterprises. As a high-risk industry that produces dangerous explosive goods, there have been frequent major accidents causing serious threats to people’s Fireworks Safety of lives and property in recent years. In this situation, carrying out a reasonable safety evaluation of firecracker production enterprises is extremely important. However, there is still no safety evaluation method specifically targeted at this industry, and there is an urgent need for a more targeted (consumer fireworks requires this), scientific, and reasonable safety evaluation method.

Currently, firecracker enterprises usually adopt a qualitative evaluation method, which cannot quantify the risks of the entire production process or production line, cannot analyze hazardous and harmful factors in-depth, and cannot achieve the purpose of safety evaluation. At the same time, as an efficient and concise semi-quantitative evaluation method, LOPA uses mathematical formulas to quantify risks based on qualitative evaluation, solving problems that existing evaluation methods cannot solve.

This article identifies the hazardous factors in the production of firecrackers and analyzes the characteristics of accidents and disasters that occur during the production process. The LOPA analysis is used to determine the accident scenarios to be studied. It is found that there are certain limitations to LOPA’s application in handling these types of complex accident scenarios due to its strict definition of independent protection layers.

To address the typical accident scenario of “multiple initiating events leading to the same consequence,” selective strategies such as the summing and maximum value methods are proposed to expand the effectiveness of LOPA’s application in this type of accident scenario. Furthermore, to address the situation where the passive protection layer is not fully effective in this scenario, the consequence reduction coefficient σ is introduced, and a scheme based on the correction of accident consequences is proposed to calculate accident frequency closer to the actual value, ultimately improving the accuracy of LOPA analysis.

Using the example of a firecrackers production enterprise, an improved LOPA is used to analyze the scenario of an explosion accident caused by the explosive dust reaching its explosion limit when workers mix the medicine. The risk level of this accident scenario is determined to be very high, exceeding the acceptable range for the enterprise. The conclusion is consistent with the BZA-1 evaluation results. Specific values for scenario frequency and consequences are provided, and the operation, maintenance, and related training focus is determined for various protection layers. It provides a scientific basis for implementing safety measures and solutions in producing firecrackers and is a new exploration in risk assessment.

Keywords: fireworks, LOPA, safety evaluation, accident scene.

Chapter 1: Introduction

1.1 Research Background and Significance of the Topic

Firecrackers are a type of product that is dangerous and explosive, making their production process extremely hazardous. In recent years, accidents related to firecrackers have occurred continuously, causing serious consequences, including loss of life and significant property damage and creating adverse social impacts [1]. Meanwhile, for producers of firecrackers, finding an effective safety assessment method is urgent to ensure safe production and establish a theoretical foundation for the work of safety regulatory departments (emergency department treated).

• Risk evaluation method for firecracker enterprises: drawbacks and solutions.
• Improving LOPA evaluation method for better risk assessment in firecracker production.
• Case study: using improved LOPA to evaluate a high-risk accident scenario in a firecrackers factory.
• Reliable risk assessment for firecracker production: the improved LOPA safety evaluation method.
• A new exploration of safety evaluation methods for firecrackers production.

1.1.1 Importance of the Development of Fireworks and Firecracker Production

Since ancient times, China has had a tradition of setting off firecrackers. With the continuous growth of the economy, various large-scale events and celebration activities have increased, which has given rise to China’s huge firecracker consumer market [2]. China is currently the world’s largest exporter of firecrackers. According to incomplete data, China’s annual production of firecrackers is around 45 million boxes, ranking first in the world, accounting for 75% of the world’s production. China’s firecrackers are exported to various parts of the world, generating approximately $400 million in foreign exchange income annually, accounting for more than 90% of the world’s total [2].

In China, many regions are engaged in producing firecrackers, including familiar areas such as Liuyang and Liling in Hunan Province, Wanzai and Pingxiang in Jiangxi Province. Numerous related enterprises exist in Jianhu of Jiangsu Province and Anping of Hebei Province [3].

According to statistics, approximately 7,000 production enterprises and nearly 140,000 sales enterprises are engaged in fireworks-related production in China, with a total of 1.5 million production employees. Production-oriented enterprises are mainly concentrated in areas south of Hunan and east of Jiangxi. The total output has reached nearly 20 million boxes, accounting for over 80% of the annual production. China’s major firework export provinces are Hunan, Jiangsu, and Guangdong. The main export destinations are the United States, Canada, Europe, the Middle East, and Southeast Asia. At the same time, China occupies most of the global trade volume of fireworks, which is worth billions of US dollars [5]. With the rapid development of international trade, China’s firecracker exports will also usher in unprecedented development.

China’s firecracker industry effectively utilizes surplus labor value, driving employment opportunities in impoverished areas. Relevant management departments and emergency department treated in our country are also continuously improving related laws and regulations, strengthening safety supervision, emphasizing safety training, and enhancing employee safety awareness, thereby ensuring safe production within the enterprise.

1.1.2 Fireworks Production Safety Accidents

According to the latest reports, since October 2014, multiple firecracker explosions and fires have occurred, with three incidents happening in Guizhou, Liaoning, and Henan provinces. The accidents were caused by illegal fireworks production that did not meet national industry standards, resulting in 20 deaths and one injury (severe burns, fireworks received severe burns, due to fuse violations in few hours), as well as significant economic losses and severe social impacts.

On the evening of October 6, 2014, in Zunyi City, Guizhou Province, an abandoned and dilapidated house exploded due to illegal firecracker production by local villagers in Lanjiang Village, Meijiang Town, Meitan County, resulting in the deaths of seven people. After investigating (Hazard investigation board) the cause of the accident, strict accountability was demanded, with the illegal explosives producers facing criminal penalties and several leading cadres facing disciplinary action.

On November 6, 2014, an accident occurred in Liujiagou Village, Chaoyang City, Liaoning Province. Villagers Yu and his relatives were illegally producing firecrackers in an idle factory, resulting in an explosion that killed all 7 people.

On December 7, 2014, in Zhujiaying Village, Gaozhuang Town, the integrated urban-rural demonstration zone of Anyang City, Henan Province, Zhu illegally processed firecrackers in his own house, resulting in an explosion of illegally produced firecrackers that killed 6 family members and seriously injured 1 person (fireworks related injuries).

From these accidents, it can be concluded that there are many dangerous and harmful factors in the production of firecrackers. However, the demand for firecrackers is increasing yearly, and the whole lighting fireworks related production industry is also constantly growing. Therefore, the top priority is to seek a reasonable safety evaluation and guide companies to improve in a safer direction.

1.1.3 Significance of the Proposed Topic

The safety evaluation of firecrackers in China is still very backward. The commonly used method is still the traditional qualitative evaluation, with few targeted methods available. However, this method largely relies on the professional experience accumulated by the evaluators over the years to make judgments. The evaluation structure is also qualitative. The evaluation conclusion of the enterprise’s on-site environment, production process, and personnel management issues determines whether the enterprise has reached the corresponding safety standards.

In recent years, LOPA has gradually become the focus of academic research and is widely used in many enterprises. Based on qualitative analysis, it uses certain mathematical formulas to calculate and evaluate different accident scenarios, providing quantitative results with numerical values, including accident frequency, protection layer failure probability of preventing accidents, consequence level, and other factors [8], achieving more accurate risk measurement.

To reduce accidents, lower potential accident risks for enterprises, minimize losses, and ultimately achieve standardized production in the firecrackers industry, this paper discusses the problems encountered in evaluating firecrackers. Concerning relevant laws, regulations, standards, and norms established by the state for this industry, it applies the methods and principles of safety system engineering to analyze potential unsafe factors thoroughly, uses Layer of Protection Analysis (LOPA) to evaluate the risks of enterprises, and proposes improvement suggestions to reduce similar accidents from occurring, ensuring the safety of workers and enterprises, and providing a basis for regulatory authorities.

1.2 Current Status of Domestic and Foreign Research

1.2.1 Current Status of Foreign Research

LOPA was first proposed in 1993 by the Center for prohibited Chemicals Process Safety (CCPS) in the United States. In 2001, the “Simplified Process Risk Management Using LOPA” application guide was announced [9].

Subsequently, further technical services related to LOPA application details were provided in the Netherlands and an applied science research institute. The industrial department of the research institute conducted an in-depth analysis of the technical and economic benefits of the evaluation method. It supplemented the key project tool kit with the Safety Quality Factor (SQF). The SQF defines the concept of independent protective layers (IPL), including human factors [10].

The LOPA theory was first used in 2005 to address the issue of oil and gas emissions in production. Snorre Sklet conducted the research at the Norwegian University of Science and Technology. The LOPA theory defines initial events, human errors, process measures, technical failures, external design issues and uses quantitative calculations of hydrocarbon release frequency for protection layer determination. Surprisingly, the protection layer and risk control measures proved effective [11].

In 2007, to address the issue of hazardous chemical transport pipelines, Professor Markowski and his colleagues from Lodz University of Technology proposed the concept of fuzzy logic [12]. In 2008, they introduced the “tie pattern” based on fuzzy logic. They used Fault Tree Analysis (FTA) to determine the frequency of initiating events and the probability of failure in another type of protection layer. They also used Event Tree Analysis (ETA) and then employed the QRA model to estimate and quantitatively evaluate the risks of accidents [13].

“Exsys-LOPA” resulted from a combination of LOPA analysis and expert scoring by multiple experts in 2010. They were integrated into an expert system that includes event-specific processes and databases related to the consequences of protection layers for all types of accident scenarios. The model uses LOPA and directly calls the model to prevent specific patterns of development. This process identifies the site and helps to expedite the hazard analysis process. This method is also important for the hazard analysis process. In 2011, the “Explosion-based Protection Analysis (ExLOPA)” was proposed based on fuzzy logic to analyze explosion protection layers. This method employs the probability method with fuzzy effect parameters and probabilities for ignition frequency and exposure to explosive mixtures. Data obtained using this method was verified and found to be more consistent with actual data [14].

In 2008, Texas A&M University proposed applying the risk assessment LOPA Bayesian method to liquefied natural gas terminals [15]. This method uses OREAD data to determine initial condition events, including Bayesian probability and failure values, to optimize both and ensure their results are consistent with the actual field situation.

In 2009, the Energy Flow/Barrier Analysis (EBA) model established by the Central University of Technology and Bernard Shahrokhi of the University of Kurdistan in Iran integrated the analysis of ETA and EBA to provide free trade agreements [16]. Its role is effective for individual objects.

In 2010, Dow Chemical Company was the first to consider using artificial protection layer analysis technology to determine the inspection plan based on human analysis, drawing curves based on the structure, and ultimately simplifying the calculation process. On this basis, the HAZOP/LOPA method summarized four useful steps: first, determine the SIL; second, verify whether the total risk of the process is within acceptable standards; next, compare the overall process and on-site risks; and finally, optimize measures to mitigate risks [17].

In 2011, when the SIS-TECH working group, including Sammers, in the United States, studied protective layer analysis to determine the consequences of improper classification, the demand for exaggerated consequences was expected to reduce risk, and it was estimated that underestimating consequences would lead to insufficient risk demand [18]. They proposed a semi-quantitative analysis of the consequences to assist researchers in accurately determining the hazard level of the consequences, thereby increasing the reliability of the analysis results.

1.2.2 Current Status of Domestic Research

Starting in 2007, the LOPA research was initiated by the Qingdao Institute of Safety Engineering. In 2008, the concept of the computer was introduced by the Beijing University of Chemical Technology to implement the computer-assisted HAZOP research method. The research was based on SDG-HAZOP, and the advantages and feasibility of the computer-assisted method were summarized [19].

In 2010, Zhou Rongyi and others from the Hunan Province Key Laboratory of Hunan University of Science and Technology began conducting Hazard and Operability Analysis (HAZOP) analysis to explore safety technology in coal mining. They applied it to an explosive gas-pressure vessel. Combined with the LOPA method, the defects of the HAZOP analysis were improved, and the technical analysis results were more comprehensive.

China started its safety evaluation relatively late. It wasn’t until the early 1980s that the country’s safety system was developed, based on foreign evaluation methods introduced and then shaped by the absorption of analytical methods and industry-specific knowledge from abroad. Then, China started to apply program security analysis and evaluation methods. In 1986, the original Ministry of Labor and Personnel put forward the standards for hazard classification, which also involved the metallurgical and mechanical industries. It wasn’t until 1987 that the concept of safety evaluation in the mechanical industry was first proposed by the original Ministry of Mechanical and Electronic Industries and promulgated on New Year’s Day of the following year, which became the first “safety evaluation standard for mechanical factories.” Before that, the labor and environmental protection agency and department of the chemical research institute had integrated Dow Chemical Company’s fire and explosion hazard index method into the existing method to develop a method for classifying chemical hazards. After a long time, a main hazard identification and evaluation method that combined quantitative and qualitative analysis was proposed, which provided good technical support for identifying, evaluating, and classifying major disaster monitoring and control in China. However, compared with developed countries, there is still a significant gap in China’s theory and technology of safety evaluation methods. [20]

Currently, our country still does not have a complete safety assessment system. The choice of evaluation methods should be based on the characteristics of the evaluation objectives, risk types, system complexity, scope of harm, and other factors. However, since joining the World Trade Organization, our country has obtained international standards and has higher and more updated requirements for safety assessment and the development of intermediary organizations.

In 1989, the first national standardization organization was established, which was significant for the times and achieved the standardization of the firecracker industry. Four mandatory national standards and three national standards were introduced, as well as seven recommended standards for the light industry and two mandatory inspection standards. [21]

In the early 1980s, the concept of system safety engineering was just widely accepted. The safety assessment in the fireworks industry also absorbed many foreign safety evaluation methods, which greatly improved the evaluation level. In addition to safety assessment, it also involves science and technology, as well as ethics, management, law, psychology, and other social science knowledge. [22] Moreover, safety assessment indicators are related to many factors, such as the level of production technology and enterprise safety management, which have a close connection. Therefore, each evaluation method has certain applicability and limitations.

1.3 Main Content of the Research Topic

1.   Firstly, this article introduces the current production management status of firecrackers and then discusses the research progress of risk assessment for firecrackers, introducing LOPA to solve the existing problems. The research direction of the topic is obtained by studying the domestic and foreign research status of LOPA technology and selecting advanced methods.

2.   This section outlines the basic concepts of LOPA and compares them with other safety assessment methods, pointing out the advantages and disadvantages of LOPA analysis. The characteristics of firecrackers and their manufacturing process are also introduced. The potential safety hazards in using raw materials and the production process are studied and analyzed, laying the foundation for future risk assessment of firecrackers.

3.   Based on the characteristics of accidents in firecracker companies, complex accident scenarios with multiple initiating events leading to the same consequences are analyzed. Two solutions, the summing method, and the maximum value method, are introduced, and an effective selection scheme for these methods is proposed, expanding the application range of LOPA. The different characteristics of active and passive protection layers in accident defense are analyzed, and the effectiveness of passive protection layers is emphasized. The consequence correction factor is also defined to address situations where the passive protection layer is not fully effective.

4.   An analysis of a specific firecracker factory is conducted to verify the feasibility of the improved LOPA evaluation method. The results of the LOPA analysis are compared with the BZA-1 evaluation method, verifying the feasibility and advanced nature of the improved LOPA application.

Chapter 2: LOPA Evaluation Factor Analysis

In recent years, there have been frequent fireworks and firecracker accidents that have caused serious consequences and property losses, which have attracted widespread attention. To reduce the losses caused by accidents, it is possible first to study the characteristics of the hazards in the firecracker production industry to identify the potential accident (hot enough to melt and unused fireworks) characteristics. Then, targeted risk analysis methods can be selected, which are significant for the safety production of China’s firecracker industry.

A layer of Protection Analysis (LOPA) is a risk analysis technique developed from event tree analysis. This method is simple, fast, and logically clear, and is therefore widely used and has become a research hotspot. It is a semi-quantitative tool for identifying and evaluating risks and bridges qualitative and quantitative analysis. The main purpose of LOPA is to determine whether there are sufficient protection layers to prevent firework misuse and safety accidents (whether the risk can be tolerated). Therefore, this method can achieve safe production while ensuring compliance with the risk standards in the firecracker industry.

This article provides insights into the risk evaluation method for firecracker enterprises. It identifies the limitations of certain evaluation methods and proposes the LOPA method to overcome these drawbacks. The article outlines the improvements made to the LOPA method and highlights its benefits, such as reliable risk assessment and guiding factories on which protective measures to focus on. The improved LOPA method is applied to a common accident scenario, and the results indicate a very high-risk level, leading to the addition of an independent protection layer. This research provides scientific evidence for the safety management and risk analysis of firecracker production enterprises, making it a valuable resource for those in the industry.

2.1 Overview of LOPA

2.1.1 Definition of LOPA

LOPA is a simplified risk assessment tool based on qualitative hazard assessment information. The risk of a scenario can be determined by judging the frequency of the initial event and the severity of the consequences and then approximating the severity of the consequences based on the magnitude of the Independent Protection Layer (IPL) failure rate [23].

Similar to other risk analysis methods, the core goal of LOPA is to determine whether enough protective layers can be found in the production process to prevent accidental incidents, that is, to determine whether the risk can be tolerated. As shown in Figure 2-1, multiple types of protective layers may exist in a given scenario. Similarly, depending on the complexity of the production process and the severity of potential accident consequences, a scenario may sometimes require one or more protective layers.

In actual production processes, for a specific scenario, it is only necessary to ensure that the protective layers can operate successfully to avoid safety accidents. However, since there are no protective layers that are effective for all situations, a sufficient number of protective layers must be provided to reduce the risk of accidents and meet the risk tolerance standards of the production process.

For certain specific scenarios, LOPA requires a consistent standard of judgment to determine whether there are enough independent protection layers in the production process to control the risk of accidents. When the risks in the scenario cannot be eliminated, additional independent protection layers must be added. LOPA can also be used to evaluate safer design scenarios that include inherent safety. LOPA does not enforce the need for additional independent protection layers or prescribe the design approach. Still, this method can effectively measure various options that can mitigate risks. LOPA is not a strictly quantitative risk assessment method, but it can assess protective layers for accident scenarios quickly and effectively.

Fig.2-1: The protective layer in accident scene

2.1.2 Uses of LOPA

LOPA provides a method for risk analysis professionals to validate risk in specific scenarios repeatedly. Typically, accident scenarios can be identified in qualitative hazard assessment, such as process hazard analysis, change management evaluation, or design review. LOPA analysis can be employed when a production method with an unacceptable consequence has been selected, or certain reasons may result in certain consequences. The LOPA analysis method can approximate the magnitude of scenario risk.

The LOPA method can only evaluate accident scenarios with single cause-effect pairs. Once the cause-effect pairs are determined, analysts can use the LOPA method to determine which management control measures (i.e., protective measures) and engineering satisfy the definition of independent protection layers and then evaluate the risk status of the scenario. Based on the results of the assessment, analysts can determine the likelihood of risk occurrence and identify which additional risk mitigation measures need to be added to achieve an acceptable risk level. The LOPA analysis process for a particular scenario can also reveal measures that may exist in other scenarios.

Another approach to studying LOPA is to examine its relationship with quantitative risk analysis methods. In this research method, a LOPA scenario is represented by a particular path through an event tree (usually the path that could result in the most severe consequence). An example of an event tree for an initial event is shown in Figure 2-2. The event tree can clearly identify all possible outcomes of the initial event [24]. The CCPS books on quantitative risk assessment for chemical production processes, “Guidelines for Chemical Process Quantitative Risk Analysis” and “Guidelines for Hazard Evaluation Procedures and Application Examples,” provide methods for using event trees and other risk assessment methods for evaluation [25]. The individuals and teams involved in LOPA analysis must restrict risk assessment to a single consequence and a single cause paired with it, which is the initial event. During the application of LOPA, the goal of the analysts is to identify all cause-effect pairs that exceed the company’s risk tolerance standard.

Fig. 2-2: Analysis and comparison of LOPA and event tree

Analysts sometimes need to select the cause-effect pairs that represent the highest risk scenario from many similar scenarios based on LOPA analysis and process experience. This selection process can be quite difficult.

During the actual application of LOPA, the LOPA analyst cannot rely solely on the event tree to select the scenarios, and the determination of scenarios is usually based on qualitative analysis. Therefore, LOPA is a method that lies between qualitative and quantitative analysis, and if the analyst believes that LOPA can effectively assess risk, they can use this method for analysis. LOPA analysis aims to identify scenarios that represent the highest risk, and the specific selection method will be discussed in detail in the next chapter.

2.1.3 Comparison of LOPA with Other Safety Assessment Methods

LOPA, a semi-quantitative risk assessment tool, evolved from the event tree analysis method and can be seen as a bridge between qualitative and quantitative analysis. Firstly, the risk level without independent protection layers should be analyzed. Then according to the risk tolerance criteria of different companies, the weakening degree of various independent protection layers to risk should be evaluated. The basic feature of this method is that risk-specific assessments can be made according to accident scenarios. LOPA has a better foundation for risk decision-making than other methods, such as the analytic hierarchy process and fuzzy comprehensive evaluation. Compared with qualitative analysis methods, LOPA can accurately judge risks and provide accurate calculation values for scenario frequency and possible consequences. Figure 2-3 compares the characteristics of LOPA with other risk assessment methods.

1.   Qualitative Analysis Method

The key to the qualitative analysis method is the evaluator’s own experience and observational and analytical abilities. This method involves a qualitative assessment of the consequences of the production system’s processes, surrounding environmental conditions, personnel quality, equipment, facility management, etc., to identify the potential for accidents. The most common qualitative methods are HAZOP and safety checklists. The qualitative analysis method is simple and quick and can be applied to evaluate simple accident scenarios. Still, it relies heavily on the researcher’s personal experience and therefore has a strong subjectivity, and the accuracy of the evaluation results is not very high.

1.   Simplified Quantitative Analysis Method

The simplified quantitative analysis method roughly estimates the risk level by assigning scores to process parameters. Methods for assigning scores include F&EI, chemical exposure index (CEI), etc., with LOPA also included. The index evaluation method requires a single project as the evaluation object, gives a score range for each category, and calculates the total score. Assuming that the value obtained by multiplying the frequency by consequences is the risk value to be calculated, the risk level is determined based on the risk matrix obtained from the frequency level and consequence category. The simplified quantitative analysis method introduces a simple quantification process, which is conducive to improving the accuracy of the evaluation results. It avoids the influence of evaluators’ experience, therefore having better objectivity. The disadvantage of this evaluation method is that it is time-consuming. The simplified quantitative analysis method has good universality and is more suitable for evaluating complex accident scenarios.

1.   Quantitative Analysis Method

The quantitative analysis method uses models to precisely calculate the size of the risk, with representative methods including event trees and fault trees. The quantitative analysis method greatly improves the accuracy of the evaluation results. However, due to the tedious evaluation process and the high demands on evaluators, the practical operability of the method is not strong, and it is only used for risk analysis and evaluation of complex accident scenarios.

By comparing and analyzing the various risk assessment methods mentioned above, it can be seen that the LOPA method takes less time in the assessment process than quantitative analysis. It can effectively study and evaluate events that may result in serious consequences or high-frequency events and can identify the root causes of the events. The LOPA method combines the advantages of qualitative and quantitative analysis, is easy to understand and operate, and has a high degree of objectivity. Therefore, it is more suitable for evaluating complex accident scenarios.

In system safety theory, safety assessment refers to the qualitative and quantitative evaluation of safety hazards that exist in the system or are triggered by external conditions, combined with a specific situation analysis to determine the probability of system failure and the severity of the consequences after failure. However, compared with corresponding standards, the degree of system danger is evaluated to derive corresponding improvement measures to reduce accident rates and achieve safe system operation.

Safety assessment can reduce or even prevent safety accidents and effectively promote the implementation of safety management work. Appropriate measures can be identified through reasonable safety assessments to avoid risks, which helps improve the safety awareness and safety qualities of production personnel and safety technicians. On the other hand, safety assessment can provide a healthy working environment for production personnel and help to accumulate reliable data, providing a theoretical basis for government safety supervision work.

2.2 Hazard Factor Analysis

2.2.1 Fireworks Production Process

The main raw material for firecrackers is gunpowder, which is processed through a specific production process to produce special effects when ignited [27] to prevent pyrotechnic materials overload. The difference between fireworks and firecrackers lies in their motion and display effects (public displays). Fireworks emphasize visual effects, producing colors during the ignition process and sound and motion effects. Firecrackers emphasize auditory effects and are made by packing gunpowder in tightly rolled paper tubes, producing sound and flashes during ignition [28]. The production processes for fireworks and firecrackers are shown in Figures 2-4, 2-5, 2-6, 2-7, and 2-8.

Fig.2-4: Black powder production flow chart

Fig.2-5: Pyrotechnics production flow chart

Fig.2-6: Firecracker production flow chart

Fig.2-7: Fireworks production flow chart

Fig.2-8: Lead production mainly flow chart

2.2.2 Hazard Identification Technical Indicators for Fireworks and Firecrackers Production

Identification of Hazardous Substances. The nature of safety hazard factors within the system is analyzed and studied to determine the risk assessment object of firecracker accidents. During the process, the evaluation unit is determined as the unit of evaluation, which refers to a hazardous system. The types and quantities of hazardous substances within the unit are determined, and then the amount of storage is confirmed to be within the safe storage standards.

Suppose the quantity of hazardous substances within the evaluation unit is equal to or exceeds the limit specified in the “Identification of Major Hazardous Sources” GB18218-2000. In that case, the hazard source is considered a major hazard source [29]. For potassium chlorate:

The limit for the production site is 2 tons, and the limit for the storage area is 20 tons.

If there are multiple types of hazardous substances within the evaluation unit, they are judged according to a formula 2-1.

In the formula, q1, q2,…,qn represent the number of hazardous substances in units of tons; Q1, Q2,…, and QN represent the critical quantities of each hazardous substance in units of tons.

When the calculated result is greater than or equal to 1, the hazard source is defined as a major hazard source. Based on the content and properties of hazardous substances within the evaluation unit, the hazardous factors should be divided into three sub-units, and each sub-unit should be evaluated separately [30].

Identifying safety hazards within the evaluation unit includes analyzing unsafe human factors and the level of safety management, the condition of stored items (pyrotechnic compositions for particular risk), hazardous substances, equipment during production, and unsafe environmental factors within the production environment [31].

Severity assessment of tested fireworks products and trash fire safety accidents involves selecting different evaluation methods for hazards with different characteristics and calculating potential casualties and property losses. If the casualties and losses meet the “major” accident criteria, the hazard source can be identified as a major hazard source.

2.2.3 Hazard Analysis of Raw Materials

The inherent hazard of firecrackers refers to the safety hazards brought by the production of raw materials during the manufacturing process, including chemical raw materials and semi-finished products.

The production raw materials used in the production of firecrackers are all chemical raw materials, including oxidants, binders, colorants, reducers, etc. [32]. Among them, commonly used oxidants include potassium chlorate, strontium carbonate, etc. These oxidants are highly oxidizing and chemically active and belong to flammable and explosive substances. They can decompose in the air and release a large amount of heat. Even without sufficient oxygen, they can maintain a burning state. They can easily mix with combustibles to form explosive mixtures if they encounter high temperatures, impact, friction, vibration, and other conditions. Sulfur, charcoal, magnesium-aluminum alloy, and other reducers may also be used in the production of firecrackers [33]. These reducers can undergo a chemical reaction and release hydrogen and heat when they absorb moisture and are also flammable and explosive substances [34]. Moreover, the activity of these substances increases when they come into contact with oxidants, and they are prone to explode when external forces impact them. The burning and explosion characteristics of pyrotechnic formulations are shown in Table 2-1.

Table 2-1: Pyrotechnic Explosion Characteristics Table

In short, from chemical raw materials to finished or semi-finished consumer fireworks products, there is a possibility of acute poisoning, corrosive injuries, fireworks related injuries and chemical burns, in addition to the dangers of fire and explosion that we are familiar with [35].

 2.2.4 Hazard Analysis of Production Processes

1.   Hazard Analysis of Process Layout

The factory process layout should comply with the provisions of the “Safety Code for the Design of Fireworks and Firecracker Factories” (GB50161-92), staying away from densely populated areas and public places to avoid safety accidents caused by mutual influence between public places and factories [36].

The production area needs to be divided into zones. For buildings protected by protective embankments, internal distances should be strictly regulated, and the amount of stored materials should not exceed the limit. Operations should be carried out according to the regulations; otherwise, safety accidents may occur.

For hazardous processes such as bright bead granulation, the workshops should be arranged in specific areas to avoid affecting other buildings and causing greater economic losses in the event of a safety accident.

To ensure that the width and slope of roads within the production area are strictly designed following regulations, it is necessary to ensure that the roads are level, have proper slope, and have a sufficient turning radius for vehicles to avoid accidents caused by small gaps between vehicles during transportation. In addition, during loading and unloading, it is important to ensure a safe distance between vehicles and potential hazards.

1.   Analysis of Hazards and Operability

When selecting and crushing raw materials, it is important to ensure that potentially hazardous materials are separated and ground using specialized machines to prevent accidents during production. The electric motors used in these machines must be explosion-proof, and the motors must be turned off when materials are being loaded or unloaded. Proper ventilation must also be maintained during the inspection process.

In mixing and preparing gunpowder, it is important to ensure that the raw materials meet national standards and that there are no expired or damp materials. Otherwise, this can result in spontaneous combustion or even explosions. The relevant formula must be strictly followed if prohibited substances are added to the gunpowder, as unauthorized changes can cause explosions. The equipment must be immediately cleaned if the dust concentration exceeds the standard. Waste material in the sedimentation tank must be processed regularly to prevent combustion or explosions caused by external ignition sources.

During loading, building, and assembling fireworks, it is important to follow the “Labor Safety Technical Regulations for Fireworks and Firecrackers” (GB11652-89) and strictly adhere to the prescribed standards for taking medicine. For the remaining limited quantities of drugs, semi-finished products should be returned to the warehouse as soon as possible. It is important to ensure that the number of personnel working in the workshop does not exceed the specified limit and that all operations are carried out strictly with rules and regulations. The prescribed doses must be strictly followed for effect, whistle, and launch drugs, and excess storage must be avoided to prevent serious accidents.

1.   Hazard analysis of the working environment

A lightning strike is one of the important factors affecting the production of firecrackers. It can cause the combustion of the materials and lead to a fire. In this high-risk industry, strict lightning protection design must be implemented, and the appropriate lightning protection measures and grounding resistance should be selected based on the specific conditions of the enterprise. It is also necessary to ensure sufficient safety clearance to prevent direct or induced lightning strikes. Lightning protection and static electricity prevention devices should be installed in the production area. Regular inspections are necessary for the already installed lightning protection devices to ensure their normal operation.

The hazard of dust explosions is that producing firecrackers generates a large amount of dust particles. If the workers inhale it for a long time, it may cause lung diseases. In severe cases, it may result in the loss of certain respiratory abilities, leading to pneumoconiosis. Currently, there is no effective treatment for this disease. In addition, an explosive environment with high dust content is easily formed, and it is highly likely to cause explosion accidents when exposed to open flames.

The potential hazards of explosive dust mainly include the slow explosion speed, low pressure, but long burning time with the production of large amounts of energy during the combustion process, resulting in significant destructive power. When dust explodes, it splatters while burning, causing severe carbonization of combustibles, which can easily burn operators. Another danger of dust explosions is that a single explosion can easily trigger secondary explosions in the surrounding dust, causing further harm. The incomplete combustion of dust can also generate toxic gases.

2.3 Common Methods for Safety Assessment of Fireworks and Firecrackers

There are many methods for safety assessment of firecrackers. Here, we briefly introduce several of the most widely used and researched methods:

1.   Pre-hazard analysis method

The pre-hazard analysis method involves analyzing the potential hazards, categories, and conditions that may arise from materials, equipment, and process during the project development phase. This method evaluates safety risks before the project begins and identifies potential safety hazards during production. The pre-hazard analysis method can timely discover potential risk factors, determine the hazard level of the system, and provide a basis for developing protective measures, effectively preventing accidents from occurring [39].

1.   Safety checklist method

As the most basic and convenient method for hazard assessment in system safety engineering, the safety checklist (SCL) method is widely used in safety assessment work. Experienced safety inspectors analyze and evaluate the object of assessment and then compile the required inspection items and requirements into a safety checklist. During the safety assessment process, the inspection is carried out following the items listed in the safety checklist. The degree of hazard and the level of destruction that may be caused can be divided into four levels. The specific method for dividing hazard levels is shown in Tables 2-2 and 2-3.

Table 2-2: The levels of danger

Table 2-3: The grade of the possibility of accident

3. The Hazard Evaluation Method for Working Conditions

The Hazard Evaluation Method for Working Conditions, also known as the LEC method, is a simple and practical safety evaluation method that can be used to evaluate work performed in hazardous environments. This evaluation method defines the hazardousness of the work as the product of three hazardous factors: L, E, and C. Here, L represents the likelihood of an accident occurring, E represents the amount of time the operator is exposed to the hazardous environment, and C represents the potential consequences of an accident. Therefore, the hazardousness of a job can be expressed as D = L × E × C, where D represents the hazard level of the job. As the value of D increases, the hazard level of the job also increases [39]. The specific score settings for this evaluation method are shown in Table 2-4 below.

Table 2-4: The Percentile of the possibility of the incident of accident or risk

The frequency of personnel being exposed to potentially dangerous environments is included in the key considerations. As the number of exposures increases and exposure time lengthens, the likelihood of being at risk of harm also increases. There are relevant regulations abroad, such as the exposure frequency value should not exceed 10, and the minimum acceptable value should be 0.5. There may also be a score of 0 for this, but in reality, it is impossible to avoid exposure completely. The frequency score of operators exposed to hazardous environments is shown in Table 2-5.

Table 2-5: The Percentile of the frequent exposure to a potentially dangerous environment

The possible consequences of accidents or dangerous events are also the main focus of analysis. For different safety incidents, the consequences can vary greatly. K.J. Graham and C.F. Kinney have established that the score for minor injuries is 1, while accidents that may cause many deaths are scored 100. The scores for other accidents range from 1 to 100 [41], with specific scores shown in Table 2-6.

The risk score can be calculated using the above and the formula. According to the classification criteria shown in Table 2-7, the risk level can be determined based on the calculated score.

Table 2-6: The Percentile of the result of the incident of accident or risk

Table 2-7: The level of risk

1.   Method of simulation analysis of accident consequences The method of simulation analysis of accident consequences is mainly used to analyze the consequences of fires, explosions, and other accidents. It evaluates using mathematical models established under idealized assumptions and provides some reference for identifying hazards.

2.   Accident Tree Analysis Method The Accident Tree Analysis method treats possible or occurred accidents as top events and uses logical symbols to connect various causal events. Based on the tree diagram, it evaluates the various factors that could lead to accidents. Qualitative analysis using this method requires the preparation of an accident tree, determining basic events through the accident tree, obtaining the minimal cut set and minimal path set, and ultimately obtaining the importance of basic events.

3.   BZA-1 Method The BZA-1 Method is a hazard assessment method suitable for producing pyrotechnic products. It evaluates the danger and harm caused by domestic pyrotechnic product manufacturers during production by fully referencing domestic and foreign evaluation methods and organizing technical personnel to study and develop a hazard degree evaluation method applicable to pyrotechnic product manufacturers. The characteristics of this method are that it identifies hazard sources and evaluates hazard levels for independent systems and units in hazardous areas. Based on this, it adopts necessary safety countermeasures and provides reference opinions for decision-making by enterprise management departments, ultimately controlling its hazards within acceptable ranges.

2.4 Characteristics of Accidents in Fireworks and Firecracker Production

There are many fire and explosion hazards in firecracker production. For example, if operating procedures are violated during the filling of explosives, such as tapping, collision, and friction, these actions could lead to explosions. During drug screening, if iron or other metal tools are used in violation of operating procedures, sparks generated by collisions could cause explosions. When preparing medications, the mixture of drugs has high sensitivity and danger. If operating procedures are not followed, combustion and explosion accidents will likely occur. Open flames, high temperatures, heat, and lightning are also important causes of accidents. In the entire processing process, a certain amount of static electricity will be generated if a person walks wearing fiber clothing (quickly ignite clothing) and rubber-soled shoes. In this case, if the person comes into contact with firecrackers that do not contain static electricity, it may trigger a fire and explosion accident [40]. In addition to fire and explosion hazards, there are also risks of dust, lightning, and mechanical injury in the entire production process of firecrackers.

Therefore, it is important to analyze and study the accident characteristics of firecrackers and to use the accident tree analysis method to draw a tree diagram of the accidents in firecracker production, as shown in Figure 2-6. The causes and events in the accident tree for fireworks and firecrackers are shown in Table 2-8, and the basic events are shown in Table 2-9.

Fig.2-6: The analysis of major production accidents of the fireworks

Table 2-8: Main reason events

Table 2-9: Basic events

Through analysis and research of the accident tree for firecracker accidents, the main characteristics of such accidents are as follows:

1.   There are many flammable and explosive materials. Many raw materials used in production facilities are flammable and explosive substances.

2.   There are many accident-prone processes. The long production process of firecrackers involves mostly processes involving drugs, which are prone to explosion accidents. Processes such as drug preparation, loading, transfer warehouse handling, drug sieving, and fuse cutting can all lead to explosion and combustion accidents.

3.   Firecrackers explode before they burn. Accidents in firecracker production facilities always involve explosions before burning. After removing the explosion source, fireworks complete other combustible materials may burn at a certain distance, leading to fire accidents.

4.   Burning occurs before an explosion. The production facility for firecrackers may experience a fire due to certain factors, which can ignite fireworks and explosive materials easily and form explosion accidents.

5.   Explosion and burning occur simultaneously. Due to the complex structure of the materials used in the production of firecrackers, the accident process can be quite complex. Accidents may be caused by firecrackers that lead to fires or fireworks related injuries or by fires that lead to explosions. Both occur simultaneously during the accident.

6.   Burning is rapid, and explosions are violent. The consequences are unimaginable if an accident occurs in the production and storage facilities for firecrackers. It is difficult to take effective measures to control the situation once an accident occurs because burning light fireworks is rapid, and explosions are violent.

7.   The probability of personnel injury is very high. Due to the sudden and severe consequences of firecracker accidents, workers have very little time to take measures, resulting in a high rate of personnel injury in accidents.

8.   The products of explosion and burning are toxic and harmful. Most of the production materials for firecrackers are chemicals, and the combustion and explosion of chemicals produce toxic and harmful smoke, which endangers the safety of surrounding personnel.

2.5 Summary

In this chapter, we introduced the semi-quantitative assessment method LOPA, discussed its application, and compared it with other evaluation methods. We also analyzed the production process of firecrackers, identified potential hazards, and summarized the characteristics of accidents and fires based on past incidents. As a result, we have identified the main risks associated with firecracker accidents.

Chapter 3: Improvements of LOPA in Risk Assessment

The LOPA (Layers of Protection Analysis) technique has the characteristics of a simple operational process and a rigorous logical condition. It can be used for semi-quantitative risk assessment of systems and has made great progress in recent years. However, during the research process, it was found that when analyzing complex accident scenarios using LOPA, its strict requirements for independent protection layers, in some cases, limit its range of application to some extent.

This chapter focuses on complex accident scenarios caused by “multiple initiating events leading to the same consequences” and “whether passive protection layers are effective.” LOPA is improved for these two common accident scenarios in the fireworks industry, and targeted solutions are proposed. The flowchart of the LOPA risk assessment is shown in Figure 3-1.

Fig.3-1: Flowchart of risk assessment in LOPA

3.1 Accident Scenarios with Multiple Initiating Events Leading to the Same Consequences

3.1.1 Background of the Study

An accident scenario comprises a series of events, from the initiating event to the consequence of the accident. The two main factors in the LOPA application process are the initiating event that triggers the potential causes of the accident consequence and the consequence, which is the final result when all layers of protection fail. In addition, a complete accident scenario also requires conditional events (necessary conditions for the initiating event to lead to the consequence), IPL failure, and consequence modifiers (such as fire, explosion, or personnel injury as the criteria for determining the consequences). The specific composition of an accident scenario is shown in Figure 3-2.

Fig.3-2: The key elements of the accident scene

LOPA is usually used to evaluate relatively simple accident scenarios, such as when a single initiating event leads to a single consequence. However, in practical risk assessment, the relationship between events can be complex and cannot be solved by traditional methods.

In the production process of firecrackers, there are some typical complex accident scenarios, such as the most common scenario where multiple initiating events lead to the same consequence (complex accident scenario). If the traditional method of applying a single initiating event leading to a single consequence is mechanically used, it will result in inaccurate risk assessment results. Further research and analysis, as well as statistical data, are needed to improve the accuracy of LOPA. We can consider complex accidents as a parallel combination of many simple accidents, as shown in Figure 3-3.

Fig.3-3: The accident scene of “multi-initiating events trigger a single accident consequence”

3.1.2 Introduction of Traditional Methods

Through extensive research of references, some traditional methods have been identified. For example, CCPS proposed that complex accident scenarios can be decomposed into two or more simple ones. Each simple accident scenario can be evaluated separately, as shown in formulas 3-1 and 3-2. This approach can determine the final accident frequency (FC) for such complex accident scenarios.

In the above equation, FC represents the frequency of consequence C resulting from multiple initiating events. FCI is the frequency of occurrence of the accident caused by the i-th initiating event, fIi is the frequency of occurrence of the i-th initiating event, and PFDij is the probability of the failure of the j-th protection layer to prevent the initiating event I from causing consequence C. This method is called the summation method.

The summation method was used by both Bingham in 2008 and Day et al. in 2010 in their research. However, for certain industries, such as the firecracker production industry studied in this article, CCPS has proposed using the maximum value selection method, as simply adding up the probabilities of the causal events is not appropriate. Instead, the maximum probability of occurrence for each event should be selected and used as the overall probability of occurrence for the entire event, as shown in equation 3-3.

However, when introducing these two methods to handle “multiple initiating events leading to the same consequence,” there is a problem with how to choose. Further research can also be conducted on the application conditions and processing procedures of other risk assessment techniques, complementing and improving the success rate of LOPA technology.

3.1.3 Improvements to Traditional Methods

From a fundamental perspective, “multiple initiating events leading to the same consequence” is equivalent to at least one initiating event occurring simultaneously with the same consequence in multiple initiating events. Correspondingly, the cf is the accident frequency with at least one simple accident scenario occurrence.

From a statistical perspective, the probability of event occurrence and event frequency jointly measure events and serve as two statistical characteristics to evaluate the likelihood of event occurrence. The theoretical value of probability can reflect the precision of the likelihood of event occurrence. As an experimental value, the frequency may have some randomness and requires a large number of statistics to be obtained from statistical data. Risk estimation calculations are carried out in units of incidents/year. Frequency objectively reflects the probability size, and as the number of statistics increases, the frequency will gradually approach a stable value and eventually reach probability. The frequency value can also represent the size of the probability value.

As shown in Equation 3:

If there is an exclusive relationship between A1, A2, …, An, the following Equation 3-5 can be obtained.

By deriving from Formula 3-4, we can obtain Formula 3-6, which defines the desired value range. Since frequency values can represent the magnitude of probability values, the CF value is between the value obtained by the summation method and the value obtained by the maximum value method, which leads to Formula 3-7.

The process is relatively complex and time-consuming if we want to calculate CF further. However, LOPA is a semi-quantitative analysis method, and obtaining accurate calculations is a challenge for LOPA. To strike a balance between accuracy and efficiency, the summation and maximum value methods are proposed as a compromise. If we only use the summation method, the CF value obtained may exceed the true risk value. Conversely, only using the maximum value method may lead to underestimating the risk value. So, how do you choose between the summation and maximum value methods? As shown in Formula 3-7, the summation method is more accurate if the CF value is closer to the maximum range value in Formula 3-7. If the CF value is closer to 1/2 max(cf f_i), then the maximum value method is more appropriate.

From the perspective of probability theory, the summation method requires certain conditions, such as the exclusion relationship between various simple accident scenarios. In contrast, the maximum value method is only suitable for simple accident scenarios with certain correlation relationships. Based on the above principles, the following three principles should be followed to judge how to choose between the summation method and the maximum value method:

1.   Compare the accident frequency values of simple accident scenarios. Suppose the frequency of accidents caused by a certain initiating event is much higher than other initiating events. In that case, the maximum value method should be chosen, and the CF value can be approximately equal to the maximum value.

2.   Determine the correlation between initiating events. The smaller the correlation between the initiating events, the smaller the probability of simultaneous occurrence. Therefore, the summation method is more suitable. Conversely, the maximum value method is more suitable.

3.   Whether simple accident scenarios share the same protective layer, for simple accident scenarios, if many protective layers are commonly used, it can be concluded that there is a correlation between them. In this case, the maximum value method is more appropriate.

The larger the number of shared protective layers, the less likely the scenarios will be mutually exclusive. The CF value will be closer to 1/2 max(cf f_i), so the maximum value method should be used.

In summary, the conditions for using the summation method and the maximum value method are summarized in Table 3-1.

Table 3-1: Conditions for using the Summation Method and Maximum Enforcement

The arrangement of the evaluation order is also an important issue. Considering the above factors, the maximum value method is relatively more reasonable and effective even if there is a relatively strong correlation between the initiating events and multiple protection layers commonly used. In this process, there may be difficulty in determining the degree of connection between events, and there are no unified standards or regulations that can be applied. Likewise, there is no unified standard for the number of protective layers that conform to the conditions of the maximum value method. Therefore, the first step in the selective diagram should be to compare the frequency values. Compared to determining the relevance of the initiating events, it is easier to determine the number of shared protection layers. The second step is to judge the number of shared protection layers, and the last step is to determine the correlation between the initiating events.

The selection diagram for the summation method, maximum value method, and quantitative risk evaluation method is shown in Figure 3-4, which requires the use of factors mentioned above to be judged according to a certain selection order.

Fig.3-4: Method select diagram

3.2 Passive Independent Protection Layers

Passive independent protection layers refer to protection layers that do not require any active actions to reduce risks. Table 3-2 provides examples of independent protection layers that reduce the frequency of high-consequence events through passive means [42]. The table also presents the typical range of PFD values for various IPLs and the adopted PFD values. If the process or mechanical design, construction, installation, and maintenance of independent protection layers are correct, their function can be achieved. Examples of such devices include firewalls, explosion-proof end caps or compartments, durable materials, barriers, or flame arresters, which are intended to prevent adverse consequences from occurring (such as large-scale leaks and diffusion, shock wave damage to protected equipment and buildings, fires, and explosions in containers or pipelines, and the penetration of flame or explosion waves through pipeline systems). If these passive systems are well-designed, they can serve as highly reliable and independent protection layers, greatly reducing the frequency of events with potentially serious consequences. However, risk assessments should also be carried out for accidents with low severity, such as equipment damage caused by shock waves.

Fire-resistant coatings are a way to reduce equipment heat input. For example, when the design specification of a safety valve exceeds the base value when a boiling liquid expanding vapor explosion (BLEVE) occurs or to prevent uncontrolled exothermic reactions caused by external heat input. Fire-resistant coatings can reduce the scale of releases or provide more response time for system depressurization and firefighting. Suppose fire-resistant coatings are used as independent protection layers. In that case, the fire-resistant material must be proven to effectively prevent consequences (such as BLEVE) or provide sufficient time for other actions. At the same time, it should also be ensured that the fire-resistant coating can remain intact when directly exposed to a fire and that the spray water from the nozzle (stored in a bucket of water) of the hose does not cause it to fall off (mortar launch malfunction).

Table 3-2: The grade of the possibility of accident

Other passive independent protection layers, such as flame arresters or explosion suppressors, use simple physical principles, and therefore, they are susceptible to contamination, blockage, corrosion, unexpected conditions, potential maintenance errors, and so on. These factors must be considered when determining the PFD values of these devices.

Passive independent protection layers designed to prevent consequences, such as insect barriers or blast walls, have relatively lower PFD values in WPA analysis. Still, appropriate PFD values must be carefully determined. If process design functions (such as special materials and inspections) can prevent consequences in some companies, they can also be considered independent protection layers. This approach allows organizations to evaluate the risk differences between equipment designed using different standards. For this method, inherently safer design functions require appropriate inspections and maintenance (audits) and determination of PFD values to ensure that process changes do not alter the PFD.

In many companies, eliminating scenarios rather than mitigating the consequences of the worst-case scenario is taken for inherently safer design functions. For example, if the designed equipment can withstand an internal explosion, all scenarios where the container is ruptured due to an internal explosion can be eliminated. Specific scenarios are not referenced when using this approach, so no protection layers are set. However, inherently safer design functions require appropriate inspections and maintenance (audits) to ensure that the effectiveness of the inherently safer design functions is not changed during the change process.

 3.2.1 Problem Background and traditional treatment methods

3.2.2 Improvement of traditional methods

1.   Consequence mitigation factor σ When the passive protection layer is not completely effective, the effective part of the protection layer also has a certain mitigation effect on the consequences of the accident. If this effect is not considered, the risk will be exaggerated. Among them, the consequence mitigation factor σ represents the degree of impact of the protection layer on the consequences of the accident. If there is no protective layer, the severity of the consequences of the accident will be C1, and if there is a protection layer, the severity of the consequences of the accident will be C2, so it can be judged that C1-C2 has reduced the consequences. The consequence mitigation factor is expressed as formula 3-8.

From equation 3-8, it can be seen that the consequence correction factor value ranges between 0 and 1. The smaller this factor value, the less the protective layer affects the numerical value of the accident consequences. When the protective layer is effective, the value is 0; when it is ineffective, the value is 1.

2. Passive protective layers are not completely effective.

Fig.3-5: The accident scene when the passive protective layer has an incomplete effect

In an accident scenario without any protective measures, the consequence would be C. If all protective layers are successfully activated, they are not completely effective. There are M passive protective layers, namely PL, PL2, …, PLy, then the consequence of the accident would be reduced to o, C, o, C, …, omC (0< <1, representing the i-th consequence correction factor, i = 1, 2, …, m). The independent protective layers are IPL, IPL2, …, IPLN, and there are N of them. Figure 3-5 shows that the active protective layers do not apply to accident scenarios when they are ineffective.

If the risks of the same accident scenario are accumulated, multiple possibilities may arise. If any IPL, IPL2, …, IPLN is successfully activated, the accident consequence will not occur. Therefore, when calculating the risk of the accident scenario, only consider the case where IPL, IPL2, …, and IPLN are all ineffective. When PL, PL2, …, PLy, IPL, IPL2, …, IPLN is all ineffective, the risk of the accident scenario is fc = f’ x DpLi x … x PFD x IPFDi x PLi x C, where i ranges from 1 to M. When PL, PL2, …, PLy, IPL, IPL2, …, IPLN are all ineffective, but PL is successfully activated, the consequence of the accident would be reduced to oC, and the risk of the accident scenario is calculated using Equation 3-9.

When PL3, …, PLM, IPL, IPL2, …, IPLN are all ineffective, but PL1 and PL are successfully activated, the consequence of the accident would be reduced to oo, C, and the risk of the accident scenario is calculated using Equation 3-10.

When IPL, IPL2, …, IPLN are all ineffective, but PL, PL2, …, PLy are successfully activated, the consequence of the accident would be reduced to oo, …, omC, and the risk of the accident scenario is calculated using Equation 3-11.

Therefore, the total risk of the accident scene is the sum of the risks of each scenario, which is expressed as equation 3-12.

Based on the above equation, if we treat the consequences of a simple scenario as a constant, denoted as C, then the frequency of the entire accident scene can be expressed as equation 3-13.

In which h takes 0 or 1, respectively, but the consequences and frequency of the accident are different and are only approximations, not actual values.

In general, there should not be two passive protection layers existing simultaneously. When there is only one passive protection layer, the risk value of the accident scene can be determined as follows:

The accident frequency is,

LOPA requires high efficiency in evaluation, and some simplification work can be done to improve accuracy. The expression value is mainly based on comparing the failure probability PDFpL of the passive protection layer and the consequence correction coefficient σ. Specifically, there are two situations: When PDFpL >>, the impact of the consequence reduction factor on risk can be ignored, and the value of the expression is taken.

When PDFpL <<, the passive protection layer contributes less to the risk, and the value of the expression is taken.

Regarding practical issues, the problem of not all passive protection layers being effective, the consequence correction coefficient was subsequently introduced to expand its various potential applications. However, there are still unresolved issues that need further research. The calculation process above assumes that the probability of risk occurrence depends on the probability of the accident and its consequences, as well as the correctness of the formula itself. Previously, the above equation has been used in fields such as CCPS, and some researchers believe that risk can be viewed as a parameter function of the frequency of accidents and their consequences. This statement is inaccurate, but the result is approximately equal to the true value of the risk, so it does not affect the final analysis process significantly, and semi-quantitative risk analysis can be performed.

In summary, this chapter explores the unique advantages of LOPA technology in safety evaluation. It studies complex accident scenarios where multiple initiating events lead to the same consequence in LOPA and introduces two solutions: the maximum value method and the summing method. It standardizes the scope and application methods of the maximum value and summing methods. This chapter also examines the advantages and disadvantages of two protection layers used in LOPA technology in risk assessment and finds effectiveness issues with passive protection layers. The role of passive protection layers in safety risk assessment is introduced. The consequence control coefficient is used as a parameter to propose what safety measures should be taken when passive protection layers fail, thereby strengthening the applicability of LOPA technology.

Chapter 4: Case Study of Application of Improved LOPA Evaluation Method

A firecrackers factory was selected as a case study to verify the feasibility of the improved LOPA evaluation method. The factory produces firecrackers products as well as fuses. The amount of gunpowder in a single firework is 0.05g, and according to the provisions of GB10631-2004, the firework products produced by the factory are classified as level C. The annual production of fireworks is about 80,000 pieces, and the annual production of fuses is about 2,500 pieces.

4.1 LOPA Analysis

4.1.1 HAZOP Analysis

The main raw materials used are aluminum powder and pyrotechnic agents for the firecrackers produced by the factory. When mixing aluminum powder with pyrotechnic agents, a certain proportion is required to be followed to produce a substance containing gunpowder. This substance is highly flammable and explosive, with the risk of combustion and explosion. If it encounters impact and collision, it will be ignited and burned at a certain speed, and then it will immediately combust and explode. The combustion process of the nearest layer of gunpowder will release molecular products, which contain a large amount of energy. Suppose there is an impact between molecules and adjacent layers of gunpowder. In that case, the impact energy will be converted into thermal energy, causing a sharp rise in the temperature of adjacent substances, leading to explosion. This type of gunpowder is sensitive to thermal, mechanical, and electrical energy [43]. Under the stimulation of these three types of energy, combustion or explosion is possible.

1.   Sensitivity to Thermal Energy

A heat source ignites the firecrackers used in this paper, which are highly sensitive to thermal energy. Under thermal energy, they are prone to combustion or explosion. The degree of difficulty in combustion or explosion is usually determined by several factors, including flash point, ignition point, combustion flame sensitivity, etc., and external factors affecting the gunpowder.

1.   Sensitivity of Mechanical Energy

The firecrackers used in this article have two sensitivities to mechanical energy: impact and friction. Currently, these two parameters are also the most important indicators to consider in the production process of firecrackers products. Different degrees of combustion or explosion are likely to occur when firecrackers are subjected to mechanical energy. The reason and process of this phenomenon are complex, usually due to the presence of a large amount of oxidant and fuel in adjacent crystals of the free surface of the fireworks. When the mechanical force acts on the surface of a single crystal, the vertical pressure and shear force between the internal materials cause the oxidant and each combustible molecule to generate interaction similar to that between molecules, which breaks the original atomic bonds and undergoes chemical changes (combustion and explosion). The sensitivity of the drug, the determination of the fireworks’ composition elements, and the fireworks’ initial temperature is also affected by mechanical action.

1.   Sensitivity to Electrical Energy

Electricity can cause firecrackers to dropping sparklers burn, or explode; static electricity is also very harmful. If firecrackers are produced and stored with static electricity tools and equipment during a simple accumulation process, the accumulation of static electricity on the object will reach a certain limit. When discharged and contacted with fireworks, this may cause combustion or explosion. The process and production equipment may generate harmful static electricity while producing chemical raw materials and fireworks crushing, grinding, and screening. Moving firecrackers using iron tools mixed with firecrackers can also cause frictional static electricity. Employees wearing synthetic fiber clothing and plastic or rubber shoes may generate static electricity when operating or walking. If static electricity is not grounded in time, it tends to accumulate. At this time, if there is a discharge when contacting fireworks, static electricity may occur, leading to combustion or explosion.

The list of hazardous and harmful factors in the production process is as follows:

1.   Analysis of fire and explosion hazards in the production process of fireworks

(1) During the drug screening stage, if the standard operating procedures are not strictly followed, iron tools can cause fire and explosion accidents.

(2) Improper cutting guidance or operation in an environment above 30°C can lead to combustion and explosion hazards.

(3) During the process of mixing with drug components, if the mixed drugs significantly increase sensitivity and risk, and the standard operating procedures are violated in this process, such as using drugs for steel products, iron products, and personnel wearing clothing made of synthetic fibers that can cause collisions or friction, it may result in an explosion. Using spoons or tools to handle pouring fireworks raw materials may generate harmful fireworks static electricity, which can lead to combustion and explosion.

(4) Collisions and impacts during the loading process may cause explosions if the operating procedures are violated.

(5) During drug transportation, collisions and friction may cause explosion hazards.

(6) Throughout the production process, personnel may carry static electricity if they wear synthetic fiber clothing, plastic, and rubber shoes. If they come into contact with a fireworks device that can cause static discharge, it may result in combustion or explosion.

(7) The factories in various production workshops may have high temperatures, heat, and lightning, which may cause combustion or explosion.

2.   Dust hazards

During the production of fireworks, there is a risk of fire and explosion due to dust hazards. The crushing, screening, preparation, mixing, and loading processes generate a large amount of dust. If there is no good ventilation, the concentration of dust may exceed the limited concentration of national health standards. For example, if factory operators do not have personal protective equipment that meets their needs or lack individual protection, long-term exposure to dust may cause occupational diseases such as shortness of breath, coughing, chest pain, increased phlegm, loss of appetite, and palpitations.

3.   Mechanical Injury

During production, high-speed moving parts of machinery and equipment do not have protective devices and come into direct contact with the human body, which may cause pinch points, collisions, mechanical injuries, and other accidents. The main production equipment of the fireworks factory is a semi-automatic knitting machine. When a worker makes an error and comes into direct contact with the rotating parts, collisions, and pinch points may occur, resulting in accidents.

4.   Lightning Damage

During the production and storage of fireworks, if the production area or storage area is not equipped with lightning protection facilities, if the facilities are ineffective, or if the location of the chemical building is not within the protection range of lightning protection facilities when it is struck by lightning or during thunderstorms, it may cause fires or explosions.

4.1.2 LOPA Analysis

1.   Identify the scenario and determine the consequence level. Select a typical worker mixing drugs in the mixing room, and the drug dust reaches the explosive limit. Non-explosion-proof electrical equipment generates electrostatic sparks, igniting the drug dust and causing a dust explosion, as in the studied scenario. According to Figure 4-1, the consequence level of the accident is judged to be level 5.

Table 4-1: The grade of the consequence of accident

1.   Initial Condition Confirmation: The mixing room’s ventilation and dust removal equipment is not functioning, and the products in each process exceed the standard.

If it is considered that there are two initiating events that lead to dust reaching the explosive limit in this scenario: the failure of the ventilation and dust removal equipment in the mixing room (with a frequency of 0.1 times/year) and the excess products in each process (with a frequency of 0.2 times/year), there are no independent protection layers set up for these two accidents at the scene. In addition, the frequencies of the two initiating events are relatively close, and there is no shared protection layer. Therefore, the initiating events of “failure of ventilation and dust removal equipment in the mixing room” and “excess products in each process” are not closely related. The summing method should be used; that is, the frequency of the complex accident scenario is 0.3 times/year.

If three initiating events lead to dust reaching the explosive limit: the failure of the ventilation and dust removal equipment in the mixing room (with a frequency of 1 time/year), excess products in each process (with a frequency of 0.2 times/year), and the failure of the dust monitoring and alarm device (with a frequency of 0.1 times/year), the protection layer for these events should be personnel control, and the probability of its failure are 0.01. Because when judging the frequency of occurrence of each simple accident scenario, except for the value of excluding the failure of the ventilation and dust removal equipment in the mixing room being relatively large, the other values are relatively small, and there is a shared independent protection layer, and there is a high correlation between the initiating events. Therefore, the maximum value method should be used: the frequency of the complex accident scenario is 1 x 0.01 = 0.01 times/year; that is, the frequency of the initial event fI is 1 x 10/a.

1.   The scenario frequency has two possible results based on the effectiveness of the passive protection layer, as obtained from Chapter 3, by comparing the relationship between PFDpL and.

When PFDpL >>, the impact of the consequence reduction factor on risk can be ignored, and the value of the expression is taken as.

When PFDpL <<, the contribution of the passive protection layer to risk is small, and the value of the expression is taken as.

When calculating the frequency of a computing scenario, it is usually necessary to make adjustments based on the demand for the scenario frequency. It involves selecting the probability of ignition, personnel exposure, and the probability of specific injuries and then modifying the frequency of different consequence scenarios accordingly. Because the failure rate of protective layers is much greater than the correction coefficient, the probability density function (PDF) is equal to.

Fig.4-1: Risk Matrix

4. Determine the scenario risk level and tolerance criteria. Based on a consequence level of 5 and frequency level of 10^-2, and according to the risk matrix shown in Figure 4-1, it is determined that the risk level of the accident is very high and exceeds the acceptable level for the company. Immediate safety measures are needed.

5. Risk decision-making. The analysis team decides to add an independent protection layer (IPL) in the form of an alarm, with a PFD of 1×10^-2, to detect and prevent combustible dust from reaching its explosive limit or exceeding an acceptable concentration range. With this protective layer in place, the event frequency is reduced from 1×10^-2a to 1×10^-4a. Re-evaluating with the risk matrix, the consequence level is 5, and the frequency is 1×10^-4a. The event scenario no longer requires further action, and the risk is effectively controlled.

6. analysis results. The specific protection layer analysis structure for LOPA is shown in Table 4-2.

Table 4-2: The protective layer analysis results

4.2 Comparison of BZA-1 Results

4.2.1 BZA-1 Method

The BZA-1 evaluation method was developed by technical personnel in the China Ordnance Industry to assess the danger and harm in the production of domestic pyrotechnic products by domestic enterprises. Based on a thorough analysis of domestic and foreign evaluation methods, it is designed to evaluate the degree of danger for pyrotechnic production enterprises. Its key feature is to identify the level of danger of hazardous sources according to the actual situation of the enterprise and to provide reference opinions on the level of danger and harm for both the enterprise and its management department. Based on this evaluation, necessary safety measures can be taken.

The BZA-1 method will be used to evaluate the danger levels of the mixed explosive workshops. The hazard level for the mixed material workshop is classified as Level B, with calculated quantities of A255 kg and A38 kg, respectively. The explosive hazard source evaluation equation for the BZA-1 evaluation method is shown in equation 4-1.

In Equation 4-1, H represents the actual hazard level of an explosion-prone system, generally defined as a relatively independent workshop, warehouse, or workshop: H represents the actual hazard level within the system; H outside represents the actual hazard level outside the system; V represents the material hazard coefficient, that is, the inherent static hazard level of explosives; K represents the controllable hazard coefficient within the system that is not under control; B represents the controllable hazard level within the system; R1i/R0 is the safe distance immediately compliance rate of the building; Ci represents the severity of the impact of the explosion on the i-th building or facility outside the system due to inadequate safety distance.

The distances of the buildings within the pyrotechnic zone comply with the regulations, so R1i/R0i in Equation 4-1 is equal to 1. Therefore, H outside is substituted into Equation 4-1 to obtain Equation 4-2.

According to Equation 4-1, we know that V represents the inherent static hazard level of explosives, which is determined by the comprehensive coefficient α of the sensitivity of the explosive (indicating the degree of difficulty in combustion and the explosion of the explosive under external force, and is a comprehensive reflection of its sensitivity) and the material hazard coefficient β (indicating the destructive capacity of the explosive in combustion and explosion under external influence, expressed as the cubic root of the TNT equivalent of explosives within the system). Its expression is shown in Equation 4-3.

Based on each workshop’s maximum design storage capacity provided in the feasibility study report, we can use the BZA-1 method to obtain the TNT equivalent coefficient and convert it into the actual TNT equivalent value for each unit. We can then use the BZA-1 method to calculate the level of danger for each unit. The maximum storage capacity of the newly built workshops in the pyrotechnic zone, converted into TNT equivalent values, is shown in Table 4-3.

Table 4-3: TNT Equivalent Conversion Table

4.2.2 Calculation of Hazard Level

The calculation of the hazard level of the mixed explosive workshop mainly includes the following five steps:

Determination of α value,

1.   According to the feasibility study report, the calculated amount of the mixed explosive workshop is 55kg + 8kg (aluminum powder + tracer), equivalent to 93.24kg TNT when calculated based on the TNT equivalent coefficient of aluminum powder of 1.48. Using the BZA-1 method to check the table, the material hazard coefficient α is determined to be 4.1.

2.   The determination of β value can be calculated using Equation 4-4.

By using the above calculation, the result of Equation 4-4 can be substituted into Equation 4-3 to obtain the value of V as 18.57.

3. Calculation of the controllable hazard level B is according to the following Equation 4-5:

The calculation of the hazard degree for the mixed explosive workshop mainly includes the following 5 steps.

Determination of α,

1.   According to the feasibility study report, the calculated amount of mixed explosives in the workshop is 55kg + 8kg (aluminum powder + tracer), equivalent to 93.24kg TNT when converted using the TNT equivalent coefficient of aluminum powder of 1.48 according to the BZA-1 method. The physical hazard coefficient α is calculated to be 4.1 by consulting the table.

The determination of β can be calculated according to equation 4-4.

2.   By substituting the result of equation 4-4 into equation 4-3, the value of V can be obtained as 18.57.

Calculation of controllable hazard degree B according to the following equation 4-5.

3.   Where WB is calculated as shown in equation 4-6. The main hazardous factors in the process of mixed explosive manufacturing and processing are temperature, pressing, and friction. Since WB=a:By, according to the conditions and the BZA-1 method, it can be obtained that Y-1 (temperature)2-0 (chemical medium) y3=1.5 (pressing) y4-1 (mechanical action) ys=1.5 (static electricity), so y1+y2+y3+y4+ys=1+1.5+1+1.5=5.

Determination of personnel density or frequency coefficient D in hazardous areas. The capacity of the mixed drug workshop in this design is 22 people (young adults). The tablet pressing process is still operated by workers on site. According to the BZA-1 method, the personnel density coefficient D should be 1.4. The determination of the accident probability index value P of the hazardous source should be 1.3 according to the BZA-1 method. The calculation result of controlling the hazard level B is that B can be 168.99 according to formula 4-5. The value of the system-controllable hazard uncontrolled coefficient K is calculated according to formula 4-7.

According to the actual investigation results, the compliance rate of safety quality and safety management level of personnel in the mixed-pressure drug workshop are as follows: the compliance rate of human factors and equipment safety status is 7.8XS S, the compliance rate of the environmental and chemical safety is 0.9ZS S. According to the formula 4-7, the K value is 0.1996. The danger level H within the system is calculated using formula 4-8.

Also, because Hy = 0, Hmed = H = 52.3. The calculated results show that the danger level of the workshop is at level 39, indicating a certain degree of danger. According to relevant regulations, technical transformation is required for the mixing and pressing workshop to significantly reduce the system’s danger level and improve the intrinsic safety of the workshop.

Chapter 4 Summary

This chapter used the improved LOPA method from Chapter 3 to identify the danger of a malfunctioning fireworks in factory. It validated and analyzed the explosion limit accident scenario caused by the dust of the mixed drug in the workshop. The results were consistent with the BZA-1 evaluation method. Comparing the results, LOPA provided specific values for the scenario frequency and consequences, providing a better foundation for risk decision-making and a more reliable risk assessment.

Conclusion

This article presents the research on the risk evaluation method for firecracker enterprises and draws the following conclusions:

1.   The article identified some drawbacks of certain evaluation methods, such as their inability to deeply analyze accident causes and evaluate risks based on specific accident scenarios. This article identified the hazardous factors in the production of firecrackers and analyzed the characteristics of accidents and disasters. Based on this, the LOPA evaluation method was introduced to overcome the drawback of the qualitative analysis method’s inability to quantify risks in producing and storing firecrackers (celebrate safely).

2.   To address the aforementioned shortcomings, the LOPA analysis was improved by determining the accident scenarios to be studied and identifying the limitation of the strict definition of independent protection layers. For the typical accident scenario of “multiple initiating events leading to the same consequence,” the article proposed selective strategies, such as the summation and maximum value methods, to expand the effectiveness of LOPA’s application in this accident scenario.

3.   To address the situation where passive protection layers are not entirely effective, the article introduced the consequence reduction factor and proposed a scheme based on the modification of accident consequences to increase the accuracy of LOPA analysis.

4.   Using the improved LOPA evaluation method, the article evaluated the risk level of a common accident scenario involving workers mixing explosives at a firecracker factory. The results showed a very high-risk level beyond the acceptable limit of the enterprise. According to the quantitative calculation results, the risk decision was made to add a dust detection alarm as an independent protection layer to ensure risk control within an acceptable range.

5.   The improved LOPA safety evaluation method’s conclusion is consistent with the BZA-1 evaluation results. Compared to qualitative methods, the improved LOPA provides a more reliable risk assessment and determines the specific values of scenario frequency and consequences. Compared to quantitative analysis methods, LOPA is simpler, more efficient, and can guide factories on which protective measures to focus on for operation, maintenance, and related training. This research provides scientific evidence for the safety management and risk analysis of firecracker production enterprises. It is a new exploration of the safety evaluation method for firecrackers production.

Take your firecrackers show to the next level with PyroEquip’s innovative and practical firing equipment. Browse our selection of industry-leading brands or explore our custom-designed hardware for a seamless and impressive display. Contact us today to elevate your fireworks game (enjoy fireworks and celebrate safely in July holiday).

References

[1] Fireworks Security and safety Tips.

[2] Fireworks safety tips.

[3] Eight important fireworks safety tips.

[4] New trends for conducting hazard & operability (HAZOP) studies in continuous chemical processes.

[5] Risk analysis of the French chemical industry.

[6] Improving human factors, incident and accident reporting, and safety management systems in the Seveso industry.

[7] A review of the fire and explosion hazards of particulates.

[8] Lessons learned from past accidents-The integration of human and organizational factors with the technical aspect.

[9] Explosives Safety-An Oxymoron?

[10] Advances in Industrial Safety.