Expert System for the Design of Inherently Safer Processes. 1. Route

routes and illustrate it using an acetic acid manufacture case study. ..... 1 flash point not defined nonflammable. 0. Ind. Eng. Chem. Res., Vol. 41, ...
0 downloads 0 Views 292KB Size
6698

Ind. Eng. Chem. Res. 2002, 41, 6698-6710

Expert System for the Design of Inherently Safer Processes. 1. Route Selection Stage Chidambaram Palaniappan, Rajagopalan Srinivasan,* and Reginald Tan Laboratory for Intelligent Applications in Chemical Engineering, Department of Chemical and Environmental Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore 119260, Singapore

In the past, the design and engineering of process plants has been driven by factors related to economics followed by operability, reliability, maintainability, and safety with an emphasis on engineering budget and time. Decisions concerning safety during the evaluation phase focused on risk reduction instead of reducing or eliminating hazards. Increased public concern on safety issues and stringent environmental standards have led plant designers to consider inherently safer and environmentally friendlier processes. Opportunities for developing such processes are highest in the early stages of design. Constraints such as time and lack of inherent safety analysis tools have been cited as hurdles to the development and implementation of an inherently safer design. We address the need for inherent safety support tools in this two-part series by developing a systematic methodology for automating inherent safety analysis in the route selection and flowsheet development stages of process design. In the first part, we describe how the knowledge required for inherent safety analysis during route selection can be formalized in the form of expert rules. We also present a new inherent safety index for ranking of process routes and a graphical method for analyzing reaction networks. We illustrate the methodology by using it to compare three routes for phenol manufacture. In the second part, we describe the methodology for inherently safer flowsheet design and the implementation of an expert system that automates the methodology. 1. Introduction Several well-publicized disasters such as Flixborough, Seveso, and Bhopal have increased the public concern for safety, health, and environmental impact of industrial processes. In addition, the risks posed by chemical industries to life, property, and environment have significantly increased in recent years for many reasons: increased population density near industrial complexes, larger size of operation, higher complexity, and use of extreme operating conditions. This has led to the development and use of better hazard identification and analysis techniques such as failure modes and effects analysis, quantitative risk analysis, hazard and operability analysis, etc., which help reduce the frequency and consequences of accidents. These techniques are design-specific and are based on experiences with similar processes. Moreover, they are applied when the design is fixed and hence result in tactical risk management strategies such as controls, alarms, and operating procedures, which require periodic revision and maintenance but are still liable to fail. Such safety analyses at later stages complicate the design, lead to modification chains, and prompt additional costs. Estimates show that, in the oil and chemical industries, 15-30% of capital cost is now spent on safety and pollution prevention measures.1 In recent years, intense competition, demand for consistently high product quality, reduced lead time, shorter product life cycle, and stringent safety and environmental regulations have challenged process designers to develop processes that are inherently safer, environmentally friendlier, sim* To whom all correspondence should be addressed. Email: [email protected]. Tel: +65 68748041. Fax: +65 67791936.

pler, and more cost-effective.2 To achieve these, the process design must simultaneously satisfy economic, safety, environment, and social objectives at early stages of process development when opportunities to impact the design are the highest. An inherently safer process avoids or reduces hazards instead of controlling them. It relies on naturally occurring phenomena and robust design and eliminates or greatly reduces the need for instrumentation or administrative controls, thereby reducing the costs related to safety and environmental protection. Such a process can be designed by applying inherent safety principles such as intensification, substitution, attenuation, limitation of effects, simplification, etc., throughout the design process, from conception until completion. These principles help to avoid or reduce hazards by using safer materials and operating conditions and minimizing inventory and result in a simpler and friendlier plant. The benefits realized by industrial application of inherent safety concepts have been discussed by several authors3-5 and include improved public image, reduced life-cycle cost, improved productivity, increased reliability, reduced company liabilities, and improved safety and environmental performance. Despite the obvious importance of an inherently safer design, there has only been limited work on intelligent tools that support inherent safety analysis. We address this important need in this two-part series by developing (1) a systematic methodology for identifying hazards and inherently safer alternatives to minimize or reduce hazards at product specification, route selection, and flowsheet development stages of process development and (2) an intelligent design support tool envisioned to assist plant designers in the development of inherently safer chemical processes.

10.1021/ie020175c CCC: $22.00 © 2002 American Chemical Society Published on Web 11/27/2002

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6699

The organization of this paper is as follows: In section 2, an overview of conventional process design and safety analysis is presented. The status of inherent safety assessment is reviewed in section 3. In section 4, we propose a design methodology for performing inherent safety analysis during product specification, preliminary route screening, and detailed route evaluation stages. We also propose a new index for comparing process routes and illustrate it using an acetic acid manufacture case study. A graphical approach to analyzing the reactions in a route is then presented. The complete methodology is illustrated in section 5 using three routes for phenol manufacture. 2. Process Design and Safety Process design is a complex activity carried out by a team of people from different disciplines in stages over a period. Design at each stage can be seen as a combination of synthesis, analysis, and evaluation of process alternatives. A process goes through various stages of evolution, which includes research, process development, design and construction, operations, maintenance, modifications, and finally decommissioning. The key decision points are (1) product specification, (2) synthesis route selection, (3) flowsheet development, (4) conceptual design, (5) detailed engineering, and (6) construction and commissioning. During the product specification stage, feasibility of production of a chemical is assessed and a search for new products that serve the same function is carried out through a market survey and research and development. Once a product is selected, the routes available for the manufacture of the product are evaluated. The goal during the synthesis route stage is to discover new process chemistry and to evaluate available process routes to manufacture the product. After exploratory research, feasible routes are further developed through preliminary technical and economic models. During the flowsheet development stage, information about desired product purity, production rate, process chemistry, catalysts, and solvents is used to develop alternate flowsheets for the process which includes feed preparation, reaction-separation, recycle, recovery, and purification. These are developed and evaluated for feasibility. During this stage, unit operations, operating conditions, heat-transfer fluids, solvents, control, and operational philosophy are selected. Once a process flowsheet is chosen, process simulation is carried out to refine the operating conditions in order to optimize the process yield and operating costs and investigate the need for recycles as well as the effect of impurities on the process and product quality. During the conceptual design stage, various decisions including process equipment selection and approximate sizing, materials of construction, inventory, storage and utility requirements, reliability and availability aspects of utilities and critical equipments, mode of transportation of chemicals, effective containment of process materials, minimization of fugitive emissions, and plant layout are made. The operational regime within which the process can be operated safely is defined, and the effect of deviations is studied during this stage. During the detailed engineering stage, major and minor equipments are sized more accurately and the infrastructure is designed. Traditional process design addresses specific issues one at a time, and revisions are made to the design as and when constraints and opportunities are recognized.

Frequent economic assessments using raw material cost, product price, and process yield as metrics are made at each stage to compare the overall attractiveness of the alternatives. Safety plays only a passive role and acts as a verification criterion for the final design. This results in a tactical approach to safety where active and procedural measures are used to control hazards by detecting potentially hazardous deviations and taking corrective actions using controls and safety interlocks before deviations result in an incident. The other focus in this approach is on minimizing the impact of an accident using procedures, administrative controls, emergency response, and other management approaches. In recent years, chemical industries have been challenged to adopt ideas of life-cycle design and concurrent engineering, thereby integrating different objectives such as environment, safety, economics, and operability. These objectives are inextricably linked together; for example, an economical option might be hazardous or an ecofriendly alternative difficult to operate. Hence, tradeoffs are necessary. To develop a process that is better overall, it is imperative to identify the possible design alternatives and understand the tradeoffs associated with them in the context of the whole process. The degrees of freedom to satisfy design objectives and opportunities for the identification and development of process alternatives for reaction paths, solvents, catalysts, etc., are high during the early design stages. Several authors6,7 have also emphasized the need to shift from compliance and remediation of safety issues to integrating safety with core plant and management processes and thereby improving business metrics such as yield, resource usage, operating costs, etc., and meeting company objectives such as “sustainable technology”, “product stewardship”, and “responsible care”. Recently, CCPS has published a book that presents a unified approach called MERITT8 (maximizing EHS returns by integrating tools and talents) for enhancing process development through effective integration of environmental, health, safety, and changing business imperatives. This approach draws upon components such as pollution prevention, inherent safety, green chemistry, and related paradigms. The shift from traditional sequential design to concurrent design has enabled the adoption of a strategic approach for hazard elimination or reduction in the place of the traditional tactical risk reduction one. Here, the strategic approach refers to inherent and passive measures that eliminate hazards by using safer materials and process conditions or minimize hazards through changes in process and equipment design features. This approach can be implemented early in the facility development and has a wideranging impact on the process design. As one strategic approach, Kletz9 introduced several principles for achieving inherently safer plants, namely, intensification, substitution, attenuation, and limitation of effects. He has also mentioned simplification, avoidance of knockon effects, making incorrect assembly impossible or difficult, making status clear, tolerance, computer control, and ease of control as a means to make the process user-friendly. The application of these principles would result in elimination or reduction in the use of hazardous materials and unsafe operations, minimization of inventory, moderation of operating conditions, and an overall simpler plant. As the design evolves along its life cycle, opportunities for safer design decrease and incorporation of inherently

6700

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

safer features becomes difficult and expensive. Although it is possible to apply inherent safety principles at later stages of design, the scope for implementation is greatest in the earliest stages, in particular during route selection and the flowsheet development stage. Decisions made during the early design stages concerning the choice of the synthesis route for product, throughput, and location for manufacture are crucial and fix 80% of the capital cost.10 To make these decisions, management and designers need the knowledge of downstream ramifications of hazards and the potential benefits of alternatives on the final design. However, at this early stage there is usually little information available to identify the nature and magnitude of hazards. The conventional safety analysis tools used to identify, quantify, and control risk and operational problems are inapplicable during the pre-engineering design stages because of this lack of information. Researchers have, therefore, developed other process safety approaches that enable designers to establish a basis for process safety and to make effective decisions at early stages. These are reviewed in the next section in the context of inherent safety assessment. 3. Status of Inherent Safety Assessment Several authors have emphasized the need and importance of considering safety, environmental, and health (SHE) factors at the early stages of design. Kletz9 mentions the importance of conducting a study at the conceptual or business analysis stage in order to decide the route for the product to be manufactured. He also emphasizes the need to involve chemists along with engineers at early stages. Screening of various process routes is crucial and critical for an inherently safer design. A process route requiring high temperature, high pressure, or a large inventory of exothermically reactive chemicals is considered inherently less safe than one that operates at ambient temperature and is characterized by a low inventory of safer chemicals. To rank process routes at early stages, Edwards and Lawrence11 developed a metric called the prototype index for inherent safety (PIIS). PIIS takes into account reaction conditions such as temperature and pressure, properties of materials such as the width of the explosive range, flammability and toxicity, process inventory, and reaction yield. The relation between economics and inherent safety practices is important and necessary for the implementation of inherently safer technologies. An attempt has been made to correlate the inherent safety index with economics of six different methyl methacrylate process routes by Edwards and Lawrence.11 One drawback of PIIS is that it concentrates on reactions and does not consider other process aspects. It is, therefore, most suitable for the route selection stage. Heikkila12 developed an index, called the inherent safety index (ISI), which considers a wider range of factors, such as process safety structure, side reactions, corrosiveness, chemical interactions, and type of equipment. This makes ISI applicable during both route selection and flowsheet development. ISI accounts for inventory based on annual throughput instead of yield. The ISI serves as a yardstick that reflects changes in the magnitude and direction of individual hazards due to changes such as reduction in temperature, pressure, and alternate chemistry made to the process. It also acts as a guide for the design team to set priorities for

further design work. A low value of the index means a high degree of inherent safety. However, the index calculation is not sufficient in itself to make a decision on the choice of a process route. Detailed inherent safety analysis of each process route is needed to identify hazards and alternatives to rectify them. Several manufacturing companies have reported their inherent safety review practices. ICI has reported the use of a six-stage hazard review procedure, which covers the process life cycle from project development to plant operation.13,14 In this approach, the first stage focuses on the identification of hazards that are related to materials, handling, material compatibility, etc., using checklists, while the second stage concentrates on the examination of preliminary flowsheets for the likelihood of occurrence of significant hazards. Union Carbide has published a checklist that can be used during inherent safety reviews.15 Lutz has discussed the need to pursue inherently safer design at every phase of the life cycle and pointed out the need for the long-term culture shift with management support to move rapidly into inherently safer invention, design, development, and process plant capitalization.16 Exxon has developed an inherent safety review process and applied it to different stages in the process life cycle.17 French et al.17 have discussed the timing of the review, team composition, activities involved, and review objectives during each stage. The review process includes a study during routine operations to identify potential improvements for the next plant. The reduction in project investments realized through inherent safety review has been found to have a strong relation with the timing of the study. Preston and Hawksley18 described the incorporation of inherent safety principles into the ICI SHE management system, provided a list of SHE study questions and guide words, and suggested the use of a SHE target diagram during safety review. A review of the status of an inherently safer process design in the U.K. conducted by the Health and Safety Executive stressed the need for methodologies that address safety, health, and environmental issues alongside business and technical objectives of design and operation in an integrated manner.19 This pilot study resulted in the European community cofunded project called INSIDE to review the status of inherently safer process and plant design in European process industries and to develop tools and methodologies for systematic application of inherent safety. The INSIDE project resulted in the development of a paper-based toolkit, called INSET, that facilitates the application of inherent safety principles along the process life cycle. This toolkit provides a set of procedures and forms to assist process designers in the search for inherently safer and environmentally friendly plants and documentation of the information related to SHE decisions.20 Schabel21 and Mansfield22 give a brief description of these tools and explain how they can be applied to identify, assess, and select alternatives at different stages. Malme´n23 and Ellis24 applied the INSET toolkit to case studies from the fine chemicals and polyurethane process, respectively. They identified the difficulties in applying the toolkit as the time involved in index calculation, the need to screen a large number of alternatives, and the necessity for analyzing complex issues at early stages. As a means of solving these problems, they have suggested a computer-based tool that generates specific alternatives, calculates the index, and facilitates faster

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6701

documentation. Rushton et al.25 listed easy access to safety-related information like flammability, toxicity, etc., and a structure that represents key elements such as the use of a hazardous intermediate, operation under extreme conditions, etc., as the desired capabilities in a support tool. He has also emphasized the need for a computer aid that will perform comprehensive inherent safety analysis and evaluation of the combinatorial number of design alternatives at each key decision point in the process life cycle. Journet26 has developed a prototype knowledge-based system that identifies hazards at the initial stages of design. This prototype system follows Douglas’ hierarchical design approach27 and uses the Mond index as a means to quantify the hazard level of the process. The key benefits of automation are a substantial reduction in time and effort, enhanced decision-making, improved documentation, and a good understanding of the process. The system focuses primarily on the flowsheet development stage and is limited to a few unit operations. Ashford and Zwetsloot28 developed a five-phase methodology to identify complementary managerial and inherently safer options through integration of a company’s SHE management system with managerial, organizational, and human aspects of the process. The viability of this methodology has been demonstrated through field experience gained from four different firms. The success of the methodology depends largely on the identification of inherent safety issues and the options to rectify them. Several authors29-32 have discussed inherently safer options and factors to be considered in designing storage tanks, heat exchangers, distillation equipment, piping, reactor systems, and rotating machinery, selecting materials of construction, and operating the plant with reduced risks. CCPS33 provides a set of guidelines in the form of checklists that can be used for selecting design bases for process safety systems. The checklists include inherent, active, passive, and procedural approaches to minimize potential hazards associated with various unit operations. Kletz9 also provides inherent safety checklists that can be used at different stages. This knowledge, available in the form of checklists in the literature, is used as a knowledge base for the identification of hazards and alternatives in our methodology. Several authors have stressed the need to promote inherent safety via dissemination of governmental policy statements, publications, directives, and standards.19,28 The existing literature on inherent safety has largely focused on dissemination of the approach. Despite the growing interest and obvious importance of inherently safer design, its adoption into practice has been slow. The main barriers to adoption of inherently safer approaches are the time and cost pressures on the project that do not allow for systematic study of alternatives during the early stages, conservatism in design and management, lack of awareness, and lack of support tools.34 In addition, existing design tools do not support such an analysis. There is a need to develop a systematic methodology that can be used to perform inherent safety analysis throughout the process design life cycle. An automated tool for inherent safety assessment at the early design stages is also desirable. These two needs form the basis for this paper. In the next section, we describe a systematic methodology to develop inherently safer processes that is also amenable to automation.

Table 1. Scores for Flammable Nature and Subindex26 condition flash point > 0 °C and boiling point e 35 °C flash point < 21 °C flash point e 55 °C flash point > 55 °C flash point not defined

flammable nature

score (Nf)

very flammable

4

easily flammable flammable liquid combustible nonflammable

3 2 1 0

4. Methodology for the Inherent Safety Evaluation of Synthesis Routes A systematic approach to an inherently safer process design would allow for identifying hazards of chemicals and reactions from easily available physical, chemical, and toxicological properties. It would also enable the use of heuristics for inferring hazards that are not immediately visible but that could occur downstream in the design life cycle. As additional information becomes available in the different stages of design, the methodology would induce application of additional stage-specific heuristics that uncover the hazards in the process. The approach would also guide the designer in identifying alternatives that preempt the identified hazards. Furthermore, the systematic nature of the methodology would enable its automation as an intelligent software tool that can be used throughout the design process. We have developed such a methodology that uses common chemical engineering and inherent safety principles for inherent safety analysis in the product specification, process route selection, and flowsheet development stages. The hazard identification and alternative generation principles needed for the first two stages share strong similarities and are described below. The methodology for the flowsheet stage and the implementation of an expert system for inherent safety analysis is described in part 2 of this series. 4.1. Product Specification Stage. The objective of inherent safety analysis during the product specification stage is to identify safety issues associated with the product and to challenge the design team to consider alternatives to the product, means of transportation, and handling systems. The inherent hazards related to storage, handling, and transportation of materials and incompatibility of materials are identified based on material physical, toxicological, chemical, and reactive properties such as flammability limits, LD50, stability, reactivity, etc. A material is concluded to be hazardous if it is flammable, explosive, toxic carcinogenic, teratogenic, mutagenic, water reactive, unstable, pyrophoric, or corrosive. Thus, flammable propane, corrosive concentrated sodium hydroxide, and carcinogenic benzene fall under the category of hazardous materials. The toxic nature of the material is determined based on its LD50, the flammable nature based on flash point and boiling point.26 The conditions for determining the flammable nature are shown in Table 1. Similarly, a material is deemed unstable if the material is sensitive to shock, impact, friction, or heat, if it has a self-polymerization tendency, or if it is incompatible with water or air. The physical and chemical properties of the material such as its vapor density, boiling point, freezing point, and particle size are considered when evaluating the impact of the hazard. The vapor density of a material determines its atmospheric dispersion characteristics in the atmosphere. Based on this, the possibility of accumulation on the ground or the need for special

6702

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

measures for identifying leakage can be identified. The design alternatives at this stage, in general, relate to the use of safer materials or modification of hazardous material into a less hazardous form. The burgeoning field of green chemistry offers general guidelines for designing safer chemicals and modifying hazardous chemicals while preserving the efficacy of function through masking or replacement of functional groups, identifying alternate functional groups, and minimizing bio-availability by changing the shape, size, structure, dilution, and altering properties.35 Such molecular design is beyond the scope of the current paper. However, specific alternatives such as the use of bleaching powder instead of chlorine, the use of toluene instead of benzene as a solvent, handling of solids in the form of pellets or granules, processing with solution, or wetting instead of fine powders can be derived with the help of heuristics. 4.2. Route Selection Stage. The process route selection is the heart of the design process and determines the inherent hazards associated with the reactor in addition to the number of downstream separation units and the need for recycle. The objective during this stage is to (1) identify the hazards and processing problems that are associated with reactions and chemicals involved in the process route and (2) rank the available process routes. During this stage, raw materials, byproducts, intermediates, catalysts, and the reaction conditions are selected. To evaluate the inherently safe nature of a route, reactions can be classified into intended and unintended reactions. Main reactions and side reactions fall under the category of intended reactions and are defined by the process chemistry. Main reactions are those in which raw materials or intermediates are transformed to products or intermediates. In contrast, side reactions result in the formation of byproducts from raw materials, products, or intermediates. Unintended reactions occur during abnormal process conditions such as utility failure, disproportionate reactants or missing ingredients, contact between incompatible materials, etc. Decomposition due to thermal runaway is an example of an unintended reaction. Reaction hazards can be identified by focusing on intended and unintended reactions occurring in the process. Hazardous reactions are characterized by high temperature or pressure, large heat release, unstable materials, or chemicals that are sensitive to air, water, rust, or oil. Material properties have to be evaluated to ensure that chemicals that are benign at ambient conditions do not become hazardous at process conditions. This will result in the definition of a safe operating regime for the reaction. Inherent safety analysis during this stage can be performed at two levels, preliminary and detailed, based on the amount of information available. 4.3. Preliminary Route Screening Stage. During preliminary route screening, information about the main reactions involved in the process route is sufficient for analysis and screening. The information needed includes the reactants, products, intermediates, byproducts (materials formed and consumed), temperature and pressure conditions, yield, heat of reaction, phase of reaction, catalysts, solvents used, and unit processes involved. Examples of unit processes include chlorination, nitration, oxidation, carbonylation, polymerization, etc. Using these, the value of an ISI for the route can be calculated.

4.3.1. Index Calculation. Process routes available can be ranked using inherent safety indices. Tyler36 has mentioned the requirements for an index to be used for assessing inherent hazards. It must be simple to use, applicable to both continuous and batch processes at all design stages, and capable of ranking individual hazards with apparent fundamental causes. The index should also promote simplified processes without penalizing novel technology. Finally, a normally acceptable range, which aids comparison with other processes, should be available. The existing indices, PIIS and ISI, satisfy most of these criteria. PIIS focuses on the main reactions and considers factors such as temperature, pressure, reaction yield, heat of reaction, flammability, toxicity, explosiveness, and inventory. It can, therefore, be used only during the chemistry selection stage. ISI accounts for side reactions, corrosion, inventory (both inside and outside battery limits), type of equipment, and process structure in addition to the factors considered by PIIS. ISI can, therefore, be applied for the flowsheet stage as well. However, ISI requires the inventory in the process, which is often not available at this early stage of design. Both of these indices do not have well-established levels of acceptability because they have not been applied to many case studies. To overcome these limitations of the existing indices, we propose an inherent safety index that is broadly based on ISI and PIIS but considers additional factors for ranking process routes. In our approach, a process route is compared based on an overall safety index along with three other supplementary indices: worst chemical index, total chemical index, and worst reaction index. 4.3.2. Individual and Overall Chemical Index. Each chemical involved in a route is characterized by an individual chemical index (ICI). ICI is related to the properties of materials involved in the route and is an indication of the hazardous nature of chemicals used in the reaction. ICI for an individual material is calculated by the summation of the indices assigned for flammability (Nf), toxicity (Nt), explosiveness (Ne), and NFPA reactivity rating (Nr). Toxicity is based on the threshold limit value, flammability on the flash point and boiling point, and the explosiveness index on the difference between explosion limits for a material. In contrast to ISI and PIIS, we include the reactivity as a measure of stability.

ICI ) Nf + Nt + Ne + Nr

(1)

The overall chemical index (OCI) for a main reaction is equal to the maximum of ICI for all of the chemicals involved in the reaction.

OCI ) max(ICI)

(2)

The OCI, therefore, resembles Edwards and Lawrence’s11 chemical score and Hekkila’s12 chemical ISI. 4.3.3. Individual and Overall Reaction Index. The individual reaction index (IRI) is calculated by the summation of subindices for temperature, pressure, yield, and heat of reaction. The temperature subindex (Rt) is a direct measure of the heat energy available at release. Similarly, the pressure subindex (Rp) is a measure of the energy available to cause a release. The degree of hazard increases when the temperature, pressure, or both are far above or below the ambient conditions. Scores for calculating Rt are shown in Table 2. Inventory in the process has a large effect on the

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6703 Table 2. Scores for the Temperature Subindex12 process temp (°C) 600

2 3 4

degree of hazard. During the early design stages, information on the inventory is usually lacking and the reaction yield can be used as a measure of the inventory because higher yield results in the reduction in size of reactors and recycles. The yield subindex (Ry) is thus a measure of the capacity of the process and the residence time in vessels. The scores for the Ry range from 0 to 10 with Ry given a value of 0 when the yield is 100%. If the process yield is less than 100% and greater than 90%, an index value of 1 is assigned. Similarly higher index values are assigned for lower yield. This is similar to the scoring pattern developed by Edwards and Lawrence.11 The thermal nature of the reaction is concluded based on values of the heat of reaction. The heat of reaction subindex (Rh) is a measure of the energy available from the reaction. This subindex is calculated for both main and side reactions. A large amount of heat release may result in higher temperature and dangerous runaway reactions.

IRI ) Rt + Rp + Ry + Rh

(3)

The overall reaction index (ORI) for a process route is calculated by summing the IRI for each main reaction.

ORI )

∑IRI

(4)

This definition of ORI for a process route penalizes a process with a large number of main reactions. Therefore, a route with a smaller number of reactors (corresponding to fewer main reactions) is preferred over one with many. In addition, oftentimes, the hazards associated with a side reaction can dwarf those from a main reaction. Thus, it is important to include the heat of side reactions in the analysis. For this, we redefine ORI to include the maximum of the heat of reaction index of side reactions (Rhs) in the process route.

ORI )

∑IRI + max(Rhs)

(5)

In addition to the calculation of OCI and ORI, hazardous chemicals and hazardous reactions in the process can be identified by ranking the chemicals and reactions in a process route based on ICI and IRI, respectively. This will help in focusing on the reduction of inventory of hazardous chemicals and the elimination of hazardous reactions. 4.3.4. Hazardous Chemical and Reaction Index. The hazardous chemical index (HCI) is the maximum of the ICI of all of the chemicals in the process. Similarly, the hazardous reaction index (HRI) is the maximum of the IRI of all of the main reactions in the process.

HCI ) max(ICI)

(6)

HRI ) max(IRI)

(7)

4.3.5. Overall Safety Index. The overall safety index (OSI) accounts for the hazards due to the chemicals in a process route and their reactions. OSI is the sum of

OSI )

∑OCI + ORI

(8)

The method of calculating OSI only provides a snapshot of the worst chemical in the route and does not differentiate between routes based on the number of chemicals or reactions and their hazardous properties. We overcome this limitation by introducing three additional supplementary indices. 4.3.6. Supplementary Indices. OCI underestimates the hazardous nature of the routes if different chemicals have similar toxicity, flammability, reactivity, or explosive ranges. For example, consider a route with a highly toxic chemical, A, and a highly flammable chemical, B. This route would be rated based either on A or on B but not both and would result in the hazard of the process being underestimated. To account for such situations, we have introduced an additional index, called the worst chemical index (WCI), which is the summation of maximum values of the flammability, toxicity, reactivity, and explosiveness subindices of all of the materials involved in that reaction step.

WCI ) max(Nr) + max(Nf) + max(Nt) + max(Ne) (9) Similarly, the worst reaction index (WRI) is also calculated for the process route. WRI is the sum of the maximum of the individual subindices of temperature, pressure, yield, and heat of reaction of all of the reactions involved in the process.

WRI ) max(Rt) + max(Rp) + max(Ry) + max(Rh) + max(Rhs) (10) For processes involving only one main reaction, the values of IRI, HRI, and WRI would be equal. The total chemical index (TCI) is a measure of the number of hazardous chemicals involved in the route. That is, a route with just one highly toxic chemical is safer compared to another route with several such toxic chemicals. TCI is calculated as the sum of the ICI of all of the chemicals involved in the process route.

TCI )

∑ICI

(11)

It should be noted that the WCI and TCI are calculated for all of the chemicals in the process route. The OSI along with the three supplementary indices are used to rank process routes as follows. Routes are first ordered according to OSI. For cases where two competing routes have similar OSIs, the supplementary indices are then compared. While the relative weightage of the three supplementary indices is subjective, we have weighted TCI, WRI, and WCI in that order. 4.3.7. Index Calculation Illustration. The use of supplementary indices is illustrated using four routes for acetic acid manufacture.37 The reactions involved, overall safety index, and supplementary indices for each process route are shown in Table 3. The reader is referred to work by Palaniappan37 for details of the index calculation. The methane oxidation is a two-step process, while the others are single-step processes. By comparison of the OSI values, the methane oxidation process can be concluded to be the most hazardous one.

6704

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Table 3. Comparison of Acetic Acid Process Routes37 process methane oxidation

halcon vapor-phase oxidation acetaldehyde oxidation low-pressure carbonylation

reactions

IRI

ICI

OCI

ORI

OSI

TCI

HRI

WRI

HCI

WCI

step 1: methane + oxygen f methanol + carbon monoxide + water step 2: methanol + carbon monoxide f acetic acid ethylene + oxygen f acetic acid

9

7

17

14

31

32

9

9

10

12

5

10

11

8

8

11

19

16

11

11

8

12

6 7

12 10

12 10

6 7

18 17

20 25

6 7

6 7

12 10

13 12

acetaldehyde + oxygen f acetic acid methanol + carbon monoxide f acetic acid

Table 4. Heuristics for Inherent Safety Analysis during the Route Selection Stage applicable item raw material/ intermediate/ catalyst/solvent

solvent reaction

condition hazardous material flammable material corrosive material toxic/mutagenic/carcinogenic/ teratogenic material

reactive/polymerizable/pyrophoric/ peroxide forming/water reactive/ thermally unstable hazardous solvent liquid-phase catalyst temperature index > 1 or pressure index > 1 heat of reaction index > 1 process yield index > 0 and catalytic reaction process yield index > 0 and noncatalytic reaction liquid-phase reaction

safety issues use of hazardous material in the reaction possibility of fire and explosion corrosion of equipment possibility of toxic vapor cloud formation

possibility of unintended reactions use of hazardous solvent increases in the exposure potential use of extreme operating conditions release of a large amount of energy process yield is low

large inventory of hazardous chemicals in the reaction system

The OSIs for the remaining three processes are close to each other and hence do not indicate a clear ranking among the three. The OSI for the low-pressure carbonylation process is lower than that of the Halcon vapor-phase oxidation process, while the low-pressure carbonylation process’ TCI is higher. The higher TCI for the low-pressure carbonylation process is due to the use of a large number of hazardous chemicals. Because the WCIs of these two processes are the same and because the information on the inventory of each hazardous chemical in the process is not available during this stage, their HRIs are compared. The HRI comparison reveals that the low-pressure carbonylation process is inherently safer. Similarly, processes can be ranked based on the worst possible combination of reaction conditions using WRI. The safety indices do not take into account safety issues related to factors such as the phase of the reaction, use of catalyst and solvent, phase generation, byproduct formation, auto-ignition, decomposition, and type of unit process involved in the chemistry. These issues are not easy to quantify into an index but can be identified using heuristics and addressed using inherent safety principles. In the next section, we describe the heuristics for issue identification and alternative generation in the route selection stage. 4.3.2. Heuristics for Identifying Hazardous Issues. Heuristics can be used to identify hazardous issues and inherently safer alternatives to minimize

alternative look for a process chemistry that uses safer material; replace the material with a safer material; change the material with a safer material by either changing or masking the material structure, form, or the phase or by dilution reduce the reactivity of the material by using stabilizers or inhibitors use of a solvent-less process use a solid- or polymer-supported catalyst look for process chemistry with safer operating conditions look for process chemistry that releases less energy look for a higher yield process use of an alternate catalyst that will improve yield use of a catalytic process that will improve yield look for a vapor-phase process chemistry

them. A part of heuristics used during the preliminary route selection stage is shown in Table 4. The phase of the reaction influences the dispersion characteristics in the case of a leak in the reactor system. For liquid-phase reactions, the use of a vapor phase can be recommended because the mass of material released for the same volume is smaller compared to that of the liquid phase under the same conditions. Waste products from reactions need to be reduced in order to reduce wastage of raw materials and decrease the reactor size required for the same throughput. If a material produced in the reaction is a byproduct, then “improve process selectivity” can be proposed as an alternative. The use of catalysts may increase the yield and allow reactions to be carried out at safer conditions. Therefore, if a process is noncatalytic and the yield is low, then the use of a catalytic process is proposed. In the case of catalytic processes with low yield, a change of the catalyst to improve the yield is suggested. Catalysts can be a hazard by themselves. For example, commonly used heavy-metal catalysts are potentially poisonous. The hazards related to catalysts and solvents used in the reaction are evaluated similarly to other materials. Heuristics are also developed for identifying hazards associated with common unit processes. For example, for an oxidation reaction, the possibility of explosion due to handling of oxygen and flammable materials and the possibility of thermal runaway due to exothermic reaction can be highlighted. The assumption underlying the

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6705 Table 5. Heuristics for Reaction Network Analysis condition

safety issues

use of a hazardous intermediate in process chemistry

increased inventory of chemicals in the process

byproduct formation as a result of a main reaction

byproduct formation results in a need for more separation units, use of more raw material, and waste of energy in the process

noncatalyzed main reaction side reactions reactants for side reactions are raw materials, intermediate, or utility reactants for side reactions are products, intermediate, or coproducts and no reactants > 1

index-based issue identification is that any subindex value greater than 1 is considered hazardous and an alternative has to be explored. For example, if the process yield index is greater than 1, then a low process yield is highlighted as an issue. Issues such as the use of extreme operating conditions, low process yield, release of a large amount of heat, and inherently safer alternatives to minimize them such as “look for safer process chemistry”, “use safer operating conditions”, and “use high yield process” can then be proposed. 4.4. Detailed Route Analysis Using Reaction Networks. In the detailed route analysis stage, the feasible routes chosen in the preliminary screening stage are further evaluated. During detailed evaluation, all possible main reactions, side reactions, and unintended reactions used for preliminary route screening are analyzed. Jacobs and Jansweijer38,39 proposed a knowledge-based approach for deriving reactor strategies from basic reaction information. We have developed a reaction network based approach for inherent safety analysis of a large number of reactions. The reaction network is similar to P-graphs and shows the path through which raw materials are converted to intermediates, products, and byproducts. It provides a means to graphically represent reactions in the form of a network and to analyze them systematically. Reactions are represented by a vertical bar, called a vertex, and materials by a circle, called a node. A reaction and its related materials are connected by directed arcs indicating the relationships between them (reactant or products). A reaction network can represent a single reaction as well as a chain of reactions where intermediates are generated. When a component appears in more than one reaction, the nodes are not duplicated. The reaction network becomes complex when raw materials are depleted by several side reactions, resulting in the formation of several byproducts and branching of reactions. Transformations of such byproducts into raw materials or intermediates through a chain of reactions further complicate the network. The cause-and-effect relationship between the chemicals and reactions represented in the form of the reaction network enables the identification of reaction patterns. Such a network can be generated manually or automatically using computer programs. A hazardous material might be

alternative coupling of reactions which produce the intermediate and the one which consumes into one reaction by changing the reaction conditions and catalyst change the reaction conditions to minimize byproduct formation in the reaction use nonhazardous or less hazardous reactants in excess look for a process chemistry with a higher yield use a catalytic process that will improve yield eliminate or minimize the formation of byproducts limit the use of excess reactants reduce the residence time of products in the reaction system or remove products immediately from the system

Figure 1. Reaction network for the two-reaction route.

Figure 2. Reaction network for material E.

produced from raw materials, intermediates, products, or utilities because of a chain of reactions. The sequence of reactions that produce each material can be identified by backward propagation in the network. This can be illustrated using a simple example. Consider a process route represented by reactions (I) and (II), where A and B are raw materials, C is a product, and D and E are byproducts.

A+BfC+D

(I)

CfE

(II)

The reaction network for these reactions is shown in Figure 1, and the network for material E, generated by backward propagation, is shown in Figure 2. From the material E network, it is clear that E arises from C, which in turn arises from A and B. The identification of such reaction patterns results in a better understanding of process routes and enables chemists to focus on the set of reactions that needs to be eliminated or improved. Side reactions lower the process yield, resulting in the formation of an undesired byproduct and making the need for a separation or processing unit inevitable. Side reactions can occur (1) in parallel with the main reaction from raw materials, (2) in series from the intermediate, product, or byproduct formed through the main reaction, and (3) in series from the byproduct formed because of

6706

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Table 6. Reactions in the Cumene Oxidation Process for Phenol Manufacture reaction conditions reaction name

temp (°C)

pressure (atm)

35-70 35-70 80-120

5-15 5-15 3-7

80-120

3-7

60-100 250-350

0-1 1

60-100 60-100 60-100

3-7 3-7 3-7

reaction

cumene process 1A cumene process 1B cumox process 1A cumox process 1b phenol production 1A cracking R-methylstyrene (RMS) hydrogenation phenol production 1B phenol production 1D phenol production 1E neutralization phenol recovery

benzene + propylene f cumene propylene + cumene f diisopropylbenzene cumene + oxygen f CHP + PDMC + acetophenone + methanol methanol + oxygen f formaldehyde + formic acid CHP f phenol + acetone RMSDimer + OCP + PCP f phenol + tars + RMS RMS + H2f cumene PDMC f RMS + water RMS + phenol f OCP + PCP RMS f RMSDimer phenol + NaOH f sodium phenolate + water sodium phenolate + H2SO4 f phenol + Na2SO4

reaction phase

process yield (%)

∆HR (kJ/g)

liquid liquid liquid

98

-0.815

96

-0.975

liquid

NA

H2SO4

liquid liquid

98 NA

NA NA NA

liquid liquid liquid

NA NA NA

catalyst H3PO4 H3PO4

-2.1

Table 7. Toluene Oxidation and Direct Benzene Oxidation Process Routes for Phenol Manufacture reaction conditions process name

main reactions involved

toluene oxidation process

toluene + air/oxygen f benzoic acid benzoic acid + air + steam f phenol + CO2 phenol f polytar toluene + oxygen f toluene hydroperoxide toluene hydroperoxide f benzaldehyde + water + CO2 + benzyl alcohol benzene + N2O f phenol + N2 benzene + N2O f CO2 + H2O + N2 + C

benzene to phenol

temp (°C)

pressure (atm)

catalyst

reaction phase

process yield (%)

∆HR (kJ/g)

130-160 220-245 220-245

2-3 1-2 1-2

cobalt acetate copper catalyst copper catalyst

liquid liquid liquid

95 88

-8.469 -1.425

300-500 300-500

NA NA

zeolite catalyst zeolite catalyst

vapor vapor

99

-2.757

Table 8. Summary of the Route Index Calculation for the Cumene Oxidation Process37 reaction

reaction type

cumene process 1A cumox Process 1A phenol production 1A cracking neutralization hydrogenation phenol recovery cumene process 1B cumox process 1B phenol production 1B phenol production 1D phenol production 1E

main reaction

heat of reaction index

pressure index

process yield index

temp index

IRI

ICI

OSI

2 2 3

1 1 0 0

1 1 1

0 1 1 2

4 5 5 2

59

1 1 0 0 0

0 0 0 0 0

0 1 1 1 1

1 2 1 1 1

6 7 6 6 6 6 6 6 9 6 6 6

side reaction

Table 9. Summary of ISI Calculations for Phenol Process Routes37 process name cumene oxidation process toluene oxidation process benzene oxidation process

no. of materials

no. of main reactions

ORI

OCI

OSI

WRI

WCI

TCI

HCI

HRI

25 10

7 2

16 13

43 12

59 25

7 8

9 8

68 19

9 6

5 13

7

1

8

6

14

8

6

14

6

8

the side reaction. Byproduct formation from side reactions can be limited by changing the conditions of the main reaction to improve selectivity. A change in the process conditions can be brought about in many ways including a change of the temperature, pressure, order of addition, solvents, and catalysts. The reaction network enables the easy identification of side reactions that deplete product, raw material, or intermediate. For cases where the product is depleted by a side reaction, alternatives such as a separate product as soon as possible and a decrease in the residence time of materials in the reaction system can be proposed. Recommendations to limit excess feed of raw materials can

be made when a side reaction depletes raw materials. Elimination of the side reaction can be proposed when the byproduct is formed through a chain of side reactions. A part of the knowledge base used for analysis of reaction networks is shown in Table 5. The reader is referred to work by Palaniappan37 for a complete list of heuristics. The analysis of the reaction pattern for material E in the above example reveals the formation of undesired byproduct E from product C. Therefore, inherently safer alternatives such as (1) remove the product C as soon as it is formed, (2) optimize the reaction conditions to reduce the formation of E, (3)

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6707

Figure 3. Reaction network for the cumene oxidation process.

Figure 4. Reaction network analysis for R-methylstyrene in the cumene oxidation process.

reduce the excess of A and B, and (4) use the process with higher yield can be identified. During this stage, index calculation and the hazard identification and minimization heuristics for prelimi-

nary route screening are also used. In addition, reaction graph analysis is used to identify the cause and effect of each byproduct and in index calculation; side reactions are also taken into account. Such an analysis

6708

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Table 10. Results of Inherent Safety Analysis of Phenol Processes context cumox process 1A phenol production 1B

safety issues

safety alternative

unit process oxidation is highly exothermic, and there is likelihood of fire and explosion hazards formation of byproduct R-methylstyrene through a side reaction in phenol production 1B resulting in a need for more separation units

phenol production 1E

formation of byproduct R-methylstyrene dimers through a side reaction of phenol production 1E

cumox process 1A

byproduct PDMC formation in cumox process 1A resulting in a need for more separation units

toluene oxidation A

liquid-phase reaction results in a large inventory of hazardous chemicals in the reaction system increased inventory of chemical in the process due to use of hazardous intermediate benzoic acid in toluene oxidation B use of extreme operating conditions

toluene oxidation B

benzene oxidation A

allows the chemists and design team to understand the route comprehensively and helps them to identify alternatives to develop an inherently safer process. 5. Inherent Safety Analysis of Phenol Production Process Routes The above-described methodology is illustrated in this section by comparing three different routes for the manufacture of phenol: cumene oxidation, toluene oxidation, and direct benzene oxidation.40-47 The process chemistry of the three different processes is shown in Tables 6 and 7. This section focuses on the analysis of the cumene oxidation route in detail and provides an overall comparison of the three process routes. The index calculated for each reaction step involved in the cumene oxidation process is shown in Table 8. Based on the index calculated for individual processes, routes are ranked and analyzed. The summary of the ISI of the three routes is shown in Table 9. The route index of the cumene oxidation process is high because of the higher number of main reactions. The TCI is also high because of the presence of a large number of chemicals involved and manufactured in this route. The temperature index for the benzene oxidation route is the highest, whereas the route index is the lowest when compared to the other two processes because it involves less hazardous chemicals, fewer reactions, and a higher yield. The hazards associated with each process route are identified through reaction network analysis. The reaction network for the cumene process chemistry generated automatically using the information about chemicals and reactions is shown in Figure 3. Once the reaction network is created, the cause and effect of each

change process chemistry use an alternate catalyst that will improve the selectivity in the reaction change process chemistry to minimize nonhazardous byproduct R-methylstyrene formation in phenol production 1B from byproduct PDMC eliminate phenol production 1B or minimize the formation of byproducts eliminate phenol production 1E or minimize the formation of byproducts change process chemistry to minimize hazardous byproduct R-methylstyrene dimers formation in phenol production 1E from byproduct R-methylstyrene use an alternate catalyst that will improve the selectivity in the reaction change process chemistry to minimize byproduct PDMC formation in cumox process 1A from intermediate cumene use nonhazardous or less hazardous materials in excess look for a vapor-phase process chemistry coupling of reactions toluene oxidation B and toluene oxidation A into one reaction by changing the reaction conditions and catalyst look for safer process chemistry that involves milder reaction conditions

chemical other than reactants and products is analyzed to map the sequence of all material transformations that take place in this route. The reaction network for R-methylstyrene is shown in Figure 4. The cause and effect of R-methylstyrene in the process route is analyzed through the reaction pattern for that material. The hazards and inherently safer alternatives for elimination or reduction of byproduct R-methylstyrene are identified through reaction network analysis. Similarly, analysis is carried out for every material that is not a raw material or product, thereby identifying the effect of byproducts, intermediates, and processing aids used in the process route. The root cause for the formation of R-methylstyrene is identified to be the formation of byproduct phenylmethyldicarbinol (PDMC) formed as a result of the reaction of phenol production 1B and through the cracking process. Alternatives such as a change of the catalyst that improves the selectivity of the reaction, elimination, or reduction of the byproduct formed and a change of the process chemistry are therefore recommended. In addition to identifying the cause of the formation of R-methylstyrene, formation of byproducts such as dimer, o-cumylphenol, and pcumylphenol through side reactions has been identified as the effect of the R-methylstyrene. The root cause of the formation of PDMC is also identified to be the cumox process 1A, and the recommendations to change the process chemistry to minimize byproduct PDMC formation in the cumox process 1A from intermediate cumene and use of nonhazardous or less hazardous materials in excess are proposed. Hazards associated with individual reactions are also identified. For example, the possibility of fire and explosion hazards associated with the cumox process 1A is identified, and the alternative to look for safer process chemistry is suggested. The use

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002 6709

of hazardous chemicals such as benzene, phenol, and cumene in the process routes is also identified. The liquid phase and extreme operating conditions in the reaction are also identified, and alternatives to develop an inherently safer process are recommended. A part of the analysis results is shown in Table 10. Because of space constraints, the details of the other process routes and reaction network of each process route, analysis results, and index calculations are not presented here. The reader is referred to work by Palaniappan37 for a complete description of the process routes and the analysis results. 6. Conclusions The development of an inherently safer process and integration of safety analysis at each design stage are vital to the chemical process industry. In this paper, we have proposed a methodology for inherent safety analysis during the route selection stage. A modified inherent safety index for comparing different process routes has been developed, and its application was illustrated using routes for acetic acid manufacture. We have presented heuristics for inherent safety analysis of process routes and a graph-based approach to evaluate reaction networks. Three different routes for producing phenol were compared using the systematic methodology to illustrate that can be used to identify safety issues and generate inherently safer alternatives. The methodology helps to increase the awareness of the design team regarding consequences of early decisions on the final design, facilitates fast-track prototyping, and reduces the time and effort spent in safety analysis at later stages. An inherently safer process design and development methodology requires iterative application of chemical engineering and inherent safety principles at each decision point along the process life cycle. In the second paper of this series, we will describe the methodology for inherent safety analysis during the flowsheet development stage. The automation of the inherent safety methodology can complement the available design support tools along the process life cycle. In part 2, we will also describe the automation of inherent safety analysis as an expert system. Literature Cited (1) Kletz, T. A. Inherently safer plants. Plant/Oper. Prog. 1985, 4, 164-167. (2) Lien, K.; Perris, T. Future directions for CAPE research perceptions of industrial needs and opportunities. Comput. Chem. Eng. 1996, 20, S1551-S1557. (3) Koch, T. A.; Krause, K. R.; Mehdizadeh, M. Improved safety through distributed manufacturing of hazardous chemicals. Process Saf. Prog. 1997, 16 (1), 23-24. (4) Windhorst, J. C. A. Application of inherently safe design concepts, fitness for use and risk driven design process safety standards to an LPG project. In Loss prevention and safety promotion in the process industries; Mewis, J. J., Pasman, H. J., De Rademaeker, E. E., Eds.; Elsevier Science: New York, 1995; Vol. II, pp 543-554. (5) Mulholland, K. L.; Sylvester, R. W.; Dyer, J. A. Sustainability: Waste minimization, Green chemistry, and inherently safer processing. Environ. Prog. 2001, 19 (4), 260-268. (6) Larson, T.; Rapaport, D.; Swett, G. H. Making business sense: Add value to your EHS programs. Chem. Eng. Prog. 2000, Mar, 20-26. (7) Turney, R. Being enthusiastic about inherent SHE in a global engineering business. IBC U.K. conference on Inherent SHEsthe cost-effective to improved Safety, Health and Environment Performance, London, U.K., June 1997.

(8) Douglas, J. M. Conceptual design of chemical processes; McGraw-Hill: Singapore, 1988. (9) Kletz, T. Process plants: a handbook for inherently safer design; Taylor & Francis: Philadelphia, PA, 1998; pp 1-19 and 152-180. (10) Mansfield, D.; Cassidy, K. Inherently safer approaches to plant design. Ind. Chem. Eng. Symp. Ser. 1994, 134, 285-299. (11) Edwards, D. W.; Lawrence, D. Assessing the inherent safety of chemical process routes: Is there a relation between plant costs and inherent safety? Process Saf. Environ. Prot. 1993, 71, Nov, 252-258. (12) Heikkila, A. M. Inherent safety in process plant design. An index-based approach; Technical Research Center of Finland, VTT Publications: Espoo, Finland, 1999. (13) Turney, R. D. Designing plants for 1990 and beyond. International Conference on Safety and Loss Prevention in the Chemical and Oil Processing Industries, Singapore, 1989; Paper H2. (14) Hawksley, J. L. Process Safety Management: A U.K. Approach. Plant/Oper. Prog. 1988, 7 (4), 265-269. (15) Lutz, W. K. Take chemistry and physics into consideration in all phases of chemical plant design. Process Saf. Prog. 1995, 14 (3), Jul, 153-160. (16) Lutz, W. K. Advancing Inherent Safety in Methodology. Process Saf. Prog. 1997, 16, 86-88. (17) French, R. W.; Williams, D. D.; Wixom, E. D. Inherent Safety, Health, Environmental (SHE) Reviews. Process Saf. Prog. 1996, 15 (1), 48-51. (18) Preston, M. L.; Hawksley, J. L. Inherent SHEs20 years of evolution. Ind. Chem. Eng. Symp. Ser. 1997, 141, 11-23. (19) Mansfield, D.; Turney, R.; Rogers, R.; Verwoerd, M.; Bots, P. How to integrate Inherent SHE in process development and plant design. Ind. Chem. Eng. Symp. Ser. 1995, 139, 467-484. (20) Rogers, R. L. An overview of the ‘INSET’ ToolKit. IBC U.K. conference on Inherent SHEsThe cost-effective to improved Safety, Health and Environment Performance, London, U.K., June 1997. (21) Schabel, J. INSET Toolkit Stages I and IIs Route Selection and Optimisation. IBC U.K. conference on Inherent SHEsThe cost-effective to improved Safety, Health and Environment Performance, London, U.K., June 1997. (22) Mansfield, D. INSET Toolkit Stages III and IVsProcess Front End and Detailed Design. IBC U.K. conference on Inherent SHEsThe cost-effective to improved Safety, Health and Environment Performance, London, U.K., June 1997. (23) Malme´n, Y. Applying the INSET ToolkitsCase Studies from the Fine Chemicals Industry. IBC U.K. conference on Inherent SHEsThe cost-effective to improved Safety, Health and Environment Performance, London, U.K., June 1997. (24) Ellis, G. Applying the INSET ToolkitsA Polyurethanes Case Study. IBC U.K. conference on Inherent SHEsThe costeffective to improved Safety, Health and Environment Performance, London, U.K., June 1997. (25) Rushton, A. G.; Edwards, D. W.; Lawrence, D. Inherent safety and computer aided process design. Process Saf. Environ. Prot. 1994, 72, May, 83-87. (26) Journet, C. Computer aids for the safe design and operation of chemical plants. Ph.D. Thesis, University of Bath, Bath, U.K., 1993. (27) CCPS. Making EHS an integral part of design; AIChE: New York, 2001; pp 1-11. (28) Ashford, N. A.; Zwetsloot, G. Encouraging Inherently safer production in European firms: a report from the field. J. Hazard. Mater. 2000, 78, 123-144. (29) Englund, S. M. Design and operate Plants for Inherent safetys1. Chem. Eng. Prog. 1991, Mar, 85-90. (30) Englund, S. M. Design and operate Plants for Inherent safetys2. Chem. Eng. Prog. 1991, May, 79-91. (31) Englund, S. M. Inherently Safer Plants: Practical Applications. Process Saf. Prog. 1995, 14 (1), Jan, 63-70. (32) Hendershot, D. C. Alternatives for reducing the risks of hazardous material storage facilities. Environ. Prog. 1988, 7 (3), Aug, 180-184. (33) CCPS. Guidelines for design solutions for process equipment failures; American Institute of Chemical Engineers: New York, 1998; pp 37-122. (34) Turney, R. D.; Mansfield, D. P.; Malmen, Y.; Rogers, R. L.; Verwoered, M.; Suokas, E.; Plaiser, A. The INSIDE Project on

6710

Ind. Eng. Chem. Res., Vol. 41, No. 26, 2002

Inherent SHE in process development and designsThe toolkit and its application. Ind. Chem. Eng. Symp. Ser. 1995, 141, 203-216. (35) DeVito, S. C.; Garrett, R. L. Designing safer chemicals: Green chemistry for pollution prevention. ACS Symp. Ser. 1996, 640, 2-21. (36) Tyler, B. J. Using the Mond index to measure inherent hazards. Plant/Oper. Prog. 1985, 4 (3), Jul, 172-175. (37) Palaniappan, C. Expert system for design of inherently safer chemical processes. M.Eng. Thesis, National University of Singapore, Singapore, 2001. (38) Jacobs, R.; Jansweijer, W. A knowledge-based method for the automatic derivation of reactor strategies. Comput. Chem. Eng. 2000, 24, 1803-1813. (39) Jacobs, R.; Jansweijer, W. A knowledge-based system for reactor selection. Comput. Chem. Eng. 2000, 24, 1781-1801. (40) Parkinson, G. A one step phenol process moves to commercialization. Chem. Eng. 1998, 105 (4), Apr, 27. (41) Mcketta, J. J.; Cunningham, W. A. Encyclopedia of chemical processing and design; Marcel Dekker: New York, 1990; Vol. 35, pp 376-385.

(42) Kharitonov, A. S.; Panov, G. I.; Sheveleva, G. A.; Pirutko, L. V.; Voskresenskaya, T. P.; Sobolev, V. I. Catalysts for production of phenol and its derivatives. Patent Publication No. WO 9527560, 1995. (43) Wittcoff, H. A.; Reuben, B. G. Industrial organic chemicals; Wiley: New York, 1996; pp 237-241. (44) Weissermel, K.; Arpe, H. Industrial organic chemistry; VCH: New York, 1997; pp 544-554. (45) Wells, M. G. Handbook of petrochemicals and processes; Gower: Aldershot, England, 1991; pp 283-287. (46) Fouhy, K. One Step phenol process offers higher yield. Chem. Eng. 1997, 104 (2), Feb, 15. (47) Elvers, B.; Hawkins, S.; Schulz, G. Ullmann’s Encyclopedia of industrial chemistry; VCH: New York, 1991; Vol. A19, pp 302306.

Received for review March 8, 2002 Accepted October 10, 2002 IE020175C