Process Safety and Environmental Protection 117 (2018) 100–110

Contents lists available at ScienceDirect

Process Safety and Environmental Protection

journal homepage: www.elsevier.com/locate/psep

Investigating the effect of inherent safety principles on system

reliability in process design
Nilesh Ade, Guanlan Liu, Ahmad F. Al-Douri, Mahmoud M. El-Halwagi, M. Sam Mannan ∗
Mary Kay O’Connor Process Safety Center, Artie McFerrin Department of Chemical Engineering, Texas A&M University, College Station, TX 77843-3122, USA

a r t i c l e i n f o a b s t r a c t

Article history: During the last decade, inherent safety has emerged as an area of interest in both academic and industrial
Received 8 September 2017 research. Various regulatory bodies have enforced the consideration of inherently safer design alterna-
Received in revised form 4 March 2018 tives. This enforcement, however, may not serve the purpose of reducing the risk associated with process
Accepted 18 April 2018
incidents due to the drawback of risk migration associated with inherent design philosophy. The phi-
Available online 23 April 2018
losophy of inherent safety has emerged from the need to reduce the consequence element of risk with
an objective to prevent high consequence–low likelihood events. Thus, this philosophy is plagued with
Keywords:
the drawback of risk migration where the lowering of consequence element can lead to an undesired
Inherent safety
Reliability
increased likelihood element, leading to an overall increase in risk associated with the system. The likeli-
Process design hood element of risk of the system under consideration depends on the system reliability. The developed
Risk methodology involves quantifying inherent safety based on the design stage under consideration using
Availability a quantification technique that utilizes process data available during the specific stage of design. This
Maintenance-downtime is followed by determining reliability and availability of the system using reliability databases or static
reliability modeling for various design alternatives considered during the specific design stage. Lastly, the
trend observed between quantified inherent safety and reliability/availability is used to determine the
required relationship between inherent safety and reliability. Thus, this developed methodology evalu-
ates the possibility of increased risk due to lowered system reliability caused by the implementation of
inherent design philosophy.
© 2018 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

1. Introduction pany personnel must be completed. Afterwards, the installation of

1.1. Hazard, risk and inherent safety
Due to the current economic downturn of oil and gas industry,
more research impetus has been dedicated towards the design of
processes that are superior in performance with respect to their
profitability. A process that is economically superior tends to be
generally safer as well, since a safer process has lesser downtime
and thus, more productivity. However, there is a tradeoff between
the economic and safety objectives in the design of a chemical
process or its equipment. As noted by Reniers et al. (2016), a
safety measure involves both costs and benefits, and a decision-
making process on whether to invest in a safety measure must
account for both aspects. There are many categories of the pos-
sible costs of safety measures. Initially, a proposed safety measure
must be selected and the applicable material and training for com-
∗ Corresponding author.
E-mail address: mannan@tamu.edu (M.S. Mannan).
the equipment involved may require the plant or part of it to be
shut down resulting in production losses. From an operating per-
spective, a safety measure may incur higher operational costs if
it requires higher utility consumption. Also, maintenance, includ-
ing possible production loss, and inspection costs for the proposed
measure must be accounted for. As for benefits of safety measures,
they can be viewed as avoided costs and are also divided into several
categories for consideration. For example, the damage to material
and property of the company and surrounding areas that could have
resulted from the absence of a proposed safety measure must be
evaluated. Another possible benefit is that a company can avoid
legal costs of lawyers and trial expenses stemming from those dam-
ages. Also, a company’s reputation can suffer from an accident at
its facility which can impact share price in the stock market. Fig. 1
shows a breakdown of the possible costs and benefits associated
with a safety investment.
The concept of inherent safety has increasingly received more
importance over the years due to its ability to design processes
with reduced hazards. However, a process that is inherently safer
with respect to process hazards may not be associated with low

https://doi.org/10.1016/j.psep.2018.04.011
0957-5820/© 2018 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
Fig. 1. Cost-benefits of safety measures.
risk as well. To understand this drawback of inherent safety, it is
essential to understand the difference between hazard and risk.
Hazard is defined as the property (physical or chemical) that has
the potential to cause harm to people, environment or property
whereas risk quantifies the level of human injury, environmental
damage, or economic loss in terms of both the incident probability
and the extent of the loss or injury (Crowl and Louvar, 1996). Inher-
ent design philosophy mainly focuses on reducing the consequence
element of risk (extent of loss) by reducing or minimizing the asso-
ciated process hazards (CCPS 2010), generally ignoring its effect
on the likelihood element. Thus, a process that is inherently safer
might have increased associated risk due to the phenomenon of risk
migration caused due to an increase in likelihood element. Inher-
ent safety is primarily based on the following 4 principles (Mannan,
2002):
• Minimization: reducing the volume of hazardous materials
• Substitution: replacing a chemical in the process with a relatively
safer alternative
• Attenuation: reducing the severity of operating conditions in the
process involving hazardous chemicals
• Simplification: implementing simpler process designs with lesser
number of equipment
1.2. Inherent safety
The Flixborough disaster in 1974 served as a cornerstone for
inherent design philosophy (Kletz, 1978). Inherent safety is defined
as the design approach to reduce or eliminate hazards with an
aim to reduce the consequence of incidents (CCPS, 2010). This
approach originates from the need to avoid low probability-high
consequence incidents. Inherent safety has found its way in several
regulations by various authoritative bodies including the UK health
and safety executive. The European Council Directive 96/82/EC of
1996 on the control of major-accident hazards involving danger-
ous substances, in its guidance document states, “Hazards should
be possibly avoided or reduced at source through the application
of inherently safe practices.” In the United States, the New Jersey
Department of Environmental Protection in its toxic catastrophe
prevention act and Contra Costa County, California in its industrial
safety ordinance have implemented the consideration of inherently
safer technologies for regulation of hazardous industrial facilities.
In February 2016, the US Environmental Protection Agency (EPA)
put forth a proposal to revise its risk management program to
include safer technologies and alternative analysis (STAA) in the
process hazard assessment for regulation of hazardous industrial
facilities. This proposal received criticism from various industrial
101
associations primarily due to the difficulty of implementing ISD in
existing facilities as compared to the design stage of grass-root facil-
ities. However, despite this industrial opposition, the amendment
was published by US-EPA in January 2017. These amendments,
however, have been delayed until February 2019 following a peti-
tion supported by various trade organizations including American
Petroleum Institute, the American Chemistry Council, the Amer-
ican Fuel & Petrochemical Manufacturers, the U.S. Chamber of
Commerce and the National Association of Manufacturers (Passut,
2017).
1.3. Equipment reliability
Reliability is the probability that a system or a component will
perform its desired function at the required time when used under
appropriate operating conditions whereas availability is the prob-
ability that a system or a component will perform its desired
function at the required time when maintained or operated in the
prescribed manner (Ebeling, 1997). Reliability serves as an impor-
tant parameter for estimating the downtime of chemical process
systems, However, quantifying the reliability of a system is insuf-
ficient in terms of estimating the maintenance downtime since
reliability and maintainability share a trade-off which overall con-
tributes towards the availability (and ultimately the maintenance
downtime) of the system. Various industries have manipulated
this relationship to improve the profitability of their processing
systems. Exxon Mobil in 1994 introduced the reliability and main-
tenance system program which reduced the maintenance cost by
$ 30 million (Exxon Mobil, 2001). Shell in its Pulau Bukom refining
facility in 1996, made design and operational modifications which
resulted in a 4-year run of a long residue catalytic cracking unit with
only 21 hours of downtime (Shell, 2001). British Petroleum saved
over $ 1.4 million in pump repairs by increasing the mean time
between failure of pumps in their facility in Lima, United States
(Griffith et al., 1998).
2. Motivation
An analysis of the occurring process safety incidents revealed
that most of these incidents occur during transient operations,
such as equipment maintenance, start-up, and shutdown (Duguid,
1998). Therefore, higher maintenance time may contribute to the
risk associated with the process. The maintenance downtime com-
prises of the corrective and preventive maintenance downtime.
The preventive maintenance downtime primarily depends on the
procedures and safety culture adopted by the company which is
generally dictated during the later stages of design of the facilities.
102 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
Fig. 2. Risk migration caused due to implementation of inherent safety principles.
The initial design stages of a chemical project are vital in deter-
mining the corrective maintenance downtime of the system since
major decisions pertaining to the number and type of different
equipment involved in the process are taken during these stages.
Also, it has been observed that implementing inherent safety prin-
ciples (ISP) becomes more and more difficult, as the design of
the chemical process plant progresses from initial stages to the
later stages (Kletz, 1991). Therefore, it becomes necessary to study
the effect of implementing ISPs on the corrective maintenance
downtime of the system during initial design stages, to analyze
the possibility of increasing the corrective maintenance downtime
since an increased corrective maintenance downtime can lead to
higher risk. It is, however, important to note that principle of sim-
plification would lead to the implementation of a lesser number
of equipment and thus, lesser corrective maintenance downtime.
However, the principles of minimization, substitution, and attenu-
ation can lead to the implementation of processes/equipment with
lower reliability and maintainability, thus increasing the corrective
maintenance downtime of the overall process. The phenomenon
of risk transfer caused due to an undesired increase in correc-
tive maintenance downtime by implementing the inherent safety
principles of minimization, substitution, and attenuation can be
described by the following figure (Fig. 2):
Apart from analyzing the relationship between inherent safety
and reliability from a systems perspective, it is important to eval-
uate the required relationship for a specific equipment such as a
reactor. This evaluation can be done at the later stages of design,
as important design data such as the material of construction is
available during these stages of design. This analysis will help in
understanding the impact of inherent safety principles and the
magnitude of this impact on the reliability of the equipment under
consideration.
3. Relevant work
A considerable depth of study has already been carried out in
designing process systems by considering inherent safety design
philosophy. To effectively understand and quantify the ease of
implementation and associated hazards of the available design
alternatives, various assessment methods, and safety indices have
been put forth. Prototype index for inherent safety (PIIS) was
one of the first indices developed with an objective to quantify
the inherent safety (or alternatively, the hazards) associated with
the chemical process routes (Edwards and Lawrence, 1993). Dow
Industries put forth the Dow fire and explosion index (Dow F&EI)
aimed at quantifying the hazards associated with a process design.
The Dow F&EI has received subsequent revision over the years and
is widely used in process industry. Inherent safety index (ISI) was
formulated with an aim to provide a simpler method for quantifica-
tion of inherent safety associated with a process design (Heikkila,
1999) and was developed to improve on the drawbacks of Dow F&EI
and PIIS. Similarly, the i-safe index was developed as a methodology
for selection of process routes based on the inherent safety of the
chemicals and the reactions involved in these routes (Palaniappan
et al., 2004). A significant deviation from the conventional method
of using non-dimensional indices for quantification of inherent
safety is the qualitative assessment for inherently safer design
(QAIS) method for application during preliminary design stage. The
described methods mainly focused on hazards associated with pro-
cess design alternatives, rather than focusing on the risk associated
with these alternatives. This drawback was countered by the for-
mulation of the concept of risk- based inherent safety (Rathnayaka
et al., 2014). A comparative analysis of the well-known safety
indices including Dow F&EI, ISI, i-safe and PIIS was subsequently
carried out to assess their agreement with each other as well as with
expert judgments in various process design stages (Kidam et al.,
2008). Significant depth of research has been carried out in reliabil-
ity, maintainability and availability studies of chemical processing
facilities as well. Reliability analysis of complex systems like chemi-
cal piping using simplistic Markov models (Gruhn et al., 1983) have
been put forth. A major contribution towards analyzing the relia-
bility of chemical systems was the development of methodology
wherein reliability and availability were obtained through block
diagrams (Henley and Gandhi, 1982). The concept of parallel and
standby redundancies was put forth, and dynamic programming
was applied to determine the optimum configuration of series sys-
tems (Rudd, 1962). A major step in reliability engineering was the
development of maintenance strategies based on the reliability of
chemical systems (Gaal and Kovacs, 1985). The widely-accepted
method of the fault tree for predicting chemical plant failure served
as a breakthrough in the field of reliability analysis (Freeman, 1983).
The developed concepts in reliability theory were effectively
applied in process design and optimization (Cos, 1973). A major
contribution in this regard was studying and optimizing the rela-
tionship between system reliabilities and unit reliabilities by
sensitivity analysis (Henley and Gandhi, 1982). Corrective mainte-
nance scheduling optimization based on reliability analysis served
as a cornerstone in developing the concept of reliability-based
 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110 103

Table 1
Commonly used safety indices and the corresponding inputs required.

Design stage Intial design stages Detailed design stage

Index/process Chemical Reaction route Operating Process flow Equipment design Plant Layout Process and
data required Properties proerties conditions diagram (PFD) instrumentation
diagram

Dow F&EI Yes Yes Yes No Yes Yes No

ISI Yes Yes Yes Yes No No No
i-safe Yes Yes Yes No No No No
PIIS Yes Yes Yes No No No No

maintenance studies (Krishnaswamy and Parker, 1984). Reliabil- available during these stages makes them suitable for their appli-
ity optimization was successfully integrated into process design cation in these stages for this study. Similarly, for detailed design
specifically during the conceptual stage of design (Goel et al., 2003). stages Dow F&EI has been utilized in this study since it can incorpo-
An important point to note is that despite this level of extensive rate equipment design data as well to quantify the levels of inherent
research in the fields of inherent safety and reliability, a study focus- safety in this stage.
ing on the relationship between these parameters has not been put With respect to the goal of studying the possibility of risk migra-
forth from the perspective of design highlighting the need for this tion caused due to the implementation of inherent safety principles,
study. it is vital to estimate the corrective maintenance downtime of the
system. The corrective maintenance downtime of the system is gov-
erned by the inherent availability of the system, where the inherent
4. Methodology
availability of the system itself depends on the inherent availabili-
ties of the equipment involved in the system. The number and type
i) For initial design stages (process selection and conceptual
of different equipment involved in the process can be obtained from
stage)
its process flow diagram (PFD).
To analyze the relationship between inherent safety and main-
The complete methodology used in this study can be described
tenance downtime with an objective to study the phenomenon of
by the following diagram:
risk migration in initial design stages, it is essential to quantify these
As mentioned previously, a process safety index based on i-safe
parameters. Inherent safety quantification depends on the stage of
and inherent safety indices is used in this study to quantify inherent
design under consideration. This is mainly because the depth of
safety of various design alternatives. The process safety index is
knowledge pertaining to the design of the chemical process system
calculated as follows:
increases along-with the progression of design stages. In this study,
Hazardous properties of all the chemicals (reactants and
the design stages are classified as per Towler and Sinnott (2013).
products) involved in the process such as reactivity, toxicity,
The stage dealing with the selection of process chemistry to man-
flammability, and explosiveness of chemicals are converted into
ufacture a required product is referred to as the process selection
non-dimensional indices as per inherent safety index methodol-
stage. The stage dealing with optimizing the process flow sheet for a
ogy (Heikkila, 1999). These indices are Nr (NFPA reactivity rating),
given chemistry is referred to as the conceptual stage and lastly, the
Nf (flammability index), Nt (toxicity index) and Ne (explosiveness
stage dealing with the detailed design of the equipment is referred
index).
to as the detailed design stage. The process selection stage and con-
Similarly, hazardous aspects of reactions such as the operating
ceptual stage have been categorized as initial design stages due to
conditions of temperature and pressure and heat of the reaction
their similarity in process data available in these stages.
are converted into non-dimensional indices as per inherent safety
A conventional method for quantifying inherent safety of a
index methodology (Heikkila, 1999). The inventory of chemicals
design (or the associated hazards) is using safety indices. A com-
required for the process is indirectly considered through the yield
parative study of the commonly used safety indices (Dow F&EI, PIIS,
of the chemical reactions involved. This yield is again converted
i-safe and ISI) revealed that ISI and i-safe indices had a strong agree-
into the non-dimensional index as per prototype index for inher-
ment with other indices and with judgement from process safety
ent safety methodology (Edwards and Lawrence, 1993). Thus, the
experts in the initial design stages (Kidam et al., 2008). Apart from
indices characterizing a reaction involved in the process are Rt
the agreement with other safety indices, it is essential to verify
(temperature sub-index), Rp (pressure sub-index), Rh (heat of reac-
that sufficient process data is available at a specific design stage
tion sub-index), Ry (yield sub-index).
to quantify the level of inherent safety through safety index. A
The process reaction and chemical index (PRCI) is a measure
study that reviewed the existing safety indices (Roy et al., 2016)
of the hazards relating to chemicals and reactions involved in the
described the information required for the commonly used safety
process and is computed as follows:
indices (Table 1):
1) Individual chemical index (ICI) = Nr + Nt + Nf + Ne (ICI is com-
Therefore, for initial design stages, due to the agreement of ISI
puted for all chemicals involved in the process)
and i-safe index with other indices and their ability to determine
the level of inherent safety by only utilizing the limited process data

Fig. 3. Methodology for analyzing the relationship between inherent safety and inherent availability.
104 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110


Table 2 12) PESI = Nj x IEQ,j for all j
Scoring of equipment.
Where j represents the different types of equipment and Nj rep-
Equipment Score (IEQ ) resents the number of a specific type of equipment involved in the
Equipment handling non-flammable, non-toxic materials 0 process.
Heat exchangers, pumps, towers & drums 1 Finally, the overall process safety index (OPSI) is calculated as,
Air-coolers, reactors & high hazard pumps 2 13) OPSI = PRCSI + PESI
Compressors & high hazard reactors 3 A higher value of OPSI indicates a more hazardous (i.e., less
Furnaces & fired heaters 4
inherently safer) process.
It should be also noted that the quantification of inherent safety
2) Individual reaction index IRI) = Rt + Rp + Rh + Ry (IRI is com-
through OPSI involves relative weighing of hazards and thus intro-
puted for all reactions involved in the process)
duces an inbuilt judgement that may not be consistent with other
3) Hazardous chemical index (HCI) = max (ICI)
safety indices, expert judgments on safety aspects and risk associ-
4) Hazardous reaction index (HRI) = max (IRI)
ated with these design alternatives.
5) Overall chemical index (OCI) = max (ICI)
Since OPSI is used for comparing the inherent safety of design
6) Overall reaction index (ORI) = IRI
alternatives by considering the hazards with respect to involved
7) Overall safety index (OSI) = (OCI + ORI)
reactions, chemicals and equipment in the process, it is important
8) Worst chemical index (WCI) = max (Nr ) + max (Nt ) + max
to understand the effect of different inherent safety principles on
(Nf ) + max (Ne )
OPSI, which is illustrated in the following diagram:
9) Worst reaction index (WRI) = max (Rt ) + max (Rp ) + max
As mentioned before, inherent availabilities are used in this
(Rh ) + max (Ry )  study to quantify the corrective maintenance downtime of pro-
10) Total chemical index (TCI) = ICI  cess systems. This is because inherent availability is defined as
11) Process reaction and chemical safety index (PRCSI) =
the steady state availability when considering only the corrective
(OSI + WCI + WRI + TCI)/4
maintenance downtime of the system. It is important to note that
Note: a) OSI, WCI, WRI and TCI are considered for the calcu-
inherent availability is an equipment design parameter. The inher-
lation of PRCSI to also evaluate the hazards associated with the
ent availabilities of the equipment themselves are based on their
worst-case scenario possible in a chemical process. b) Hazardous
mean time to failure (MTTF) and mean time to repair (MTTR) of the
chemical index (HCI) and overall chemical index (OCI) are numer-
equipment. The inherent availability of an equipment is given by:
ically identical.
Apart from considering the hazards with respect to chemicals MTTF
and reactions in a process, it is essential to consider the hazards Ainh =
MTTF + MTTR
with respect to the equipment involved in the process. Equipment
like furnaces can act as ignition sources leading to fire and explo- As observed from the expression, inherent availability is a
sion, similarly, failure in reactors handling toxic chemicals can lead parameter that captures the expertise level of the maintenance
to the release of hazardous chemicals to the environment. The scor- personnel (through MTTR) and the probability of the failure of
ing of equipment in this study is based on inherent safety index the equipment (through MTTF). If an equipment is highly reliable,
methodology (Heikkila, 1999) (Table 2). it would have high MTTF and thus higher inherent availability.
The process equipment safety index (PESI) is calculated as, The MTTF and MTTR values for general process equipment can be

Fig. 4. Effect of inherent safety principles on overall process safety index.

 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
obtained from the center for chemical process safety (CCPS) and
offshore reliability equipment data (OREDA) databases. It is impor-
tant to note that the implementation of linear failure rates available
in these databases to quantify reliability and availability is bound
to epistemic uncertainties and it might be beneficial in terms of
accuracy to implement relatively complex reliability models like
Weibull distribution in the suggested framework if comprehensive
reliability and failure data is available for the process equipment
under consideration.
In the initial design stages, redundant equipment are generally
not considered. Thus, the equipment in the process system can be
assumed to be in series with respect to their reliabilities and avail-
abilities. Therefore, the inherent availability of the system can be
expressed as:

i Also, since, the focus of this stage is the design of an individual
Ainh,sys = Ainh,i
n tial to analyses reliability alone rather than focusing on availability
Where,
Ainh,i is the inherent availability of ith equipment,
n is the total number of equipment in the process system,
& Ainh,sys is the inherent availability of the process system.
Thus, from the process flow diagrams (PFDs) of the process
designs, different types and number of equipment involved in
the process design can be known and the inherent availability of
the system can be estimated from the inherent availabilities of
equipment involved in the process. The corrective maintenance
downtime for the system is computed using the following expres-
sion:
corrective maintenance down time
Ainh,sys = 1 −
total process time
Finally, by computing OPSI and Ainh,sys for all process design
alternatives, the required relationship between these parameters
can be established.
This was followed by a predesign economic analysis (a study
estimate with a probable accuracy of ±30% for annualized profit) for
a defined capacity of 1 KTPA of Acetic acid (Peters and Timmerhaus,
1991). The economic analysis included the following steps:
Estimate the current fixed capital investment (FCI) based on a
known fixed capital investment at a certain period and chemical
engineering plant cost index by:
1) Current , FCI, by, 6-10th, rule =
current chemical engineering plant cost index
FCI, in, period ‘T’x chemical
engineering plant cost index at period T
 required capacity
0.6
x capacity for which FCI is known at period T
2) Annualize the calculated FCI by assuming appropriate num-
ber of expected service years of plant.
3) Estimate the annual operating time of plant from the calcu-
lated inherent availability (Ainh,sys ) of the system by:
Annual operating time of plant = Total annual time x inherent
availability of system (Ainh,sys )
4) Estimate the corrective maintenance labor cost by:
Corrective maintenance labor cost = Annual operating time x
corrective maintenance labor rate x [(1/Ainh,sys ) − 1] (Goel et al.,
2003)
5) Assuming the corrective maintenance material cost is roughly
the same as corrective maintenance labor cost the overall corrective
maintenance cost can be estimated as:
Overall maintenance cost = 2 x corrective maintenance labor
cost (Branan, 2005)
6) Estimate the cost of raw materials required for the desired
capacity by appropriate mass balance
7) Estimate the overall production cost (excluding overall main-
tenance cost) from the cost of raw materials by:
105
Overall production cost = 3.3 x cost of raw materials (Peters and
Timmerhaus, 1991)
8) Compute the annualized profit obtained for the given alterna-
tive by assuming an appropriate selling price of product and using
the overall production cost
ii) For detailed engineering stage
The described methodology in Fig. 3 can also be applied to
detailed engineering stage with certain modifications. The first step
in the methodology is to evaluate the design alternatives existing
for the design stage under consideration. The design alternatives in
detailed engineering stage for an equipment can be the same equip-
ment design under different operating conditions. The suitable
range of operating conditions for an equipment such as a reactor is
generally governed by thermodynamics and kinetics of the reaction
along with mass and heat transfer occurring within the equipment.
equipment and not the entire process system, it becomes essen-
as the reliability can be modelled as a function of the operating
conditions whereas, the maintainability of the equipment can be
considered to be independent of the operating conditions.
The modified methodology for detailed engineering stage can
be described as:
Since the detailed engineering stage comes after a process is
selected, Dow F&EI is deemed to be suitable for quantification of
inherent safety. The quantification of Dow F&EI is mainly based on
general process hazards such as exothermicity of the reaction, ease
of material handling and transfer, access to emergency equipment,
drainage and spill control measures at the facility, special process
hazards such as toxicity of materials involved, whether the process
is sub-atmospheric, closeness of operation to flammable range of
materials, possibility of leakage, possibility of dust explosion and
lastly on material hazards such as reactivity and flammability. The
Dow F&EI calculation in this study is based on Dow Fire and Explo-
sion index hazard classification guide, 7th edition.
The formulated methodology for detailed engineering stage is
described using an example of reactor design. The third step of the
described methodology requires the quantification of reliability of
the equipment (i.e., reactor) with respect to operating conditions,
the mode of failure for reactor under consideration is bursting due
to excess pressure and thus, the failure probability (mathematical
counterpart of reliability) can be modelled as a function of internal
pressure of the reactor. The relation between this failure probability
and operating pressure has been described by Arnold, 1972. In this
study, the induced stress in the reactor (due to internal/operating
pressure) and the allowable stress of the material of the reactor are
assumed to follow a normal distribution, and the overlapping of
these distributions governs the failure probability.
The problem of reactor design used in this study is partly based
on reactor design described by Suardin, 2005. The design problem
is described as follows:
Reaction: A → B
Feasible range of operating pressure (P): 8–14 atm
Required annual capacity: 645 million lbs of B
Reaction rate constant (k): 0.0025 s−1
Feasible conversion of reaction: 60% (defined on mass basis)
Molecular weight of A: 100 lb.mol/mol
Temperature of reaction: 800 ◦ F
Material of construction of reactor: A312 alloy steel (selected
based on operating conditions) with maximum allowable working
stress (M) of 14,750 psi from (Seider et al., 2004).
The governing equations for described methodology for reactor
design can be broadly divided into three sets of equations, equa-
tions pertaining to calculating reactor dimensions, equations for
reliability calculations and equations for Dow F&EI calculations.
a) Equations for reactor dimension calculation:
106 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110

W −W cylindrical shell of length L, to which are welded two ellipsoidal

1)Conversion of reaction with respect to mass(X) = A0 WA0
A

Where WA0 and WA are initial and final mass of A respectively. heads at opposite ends. The method employed is based on the
F 1
2)Volume of reactor(V ) = kCA0 [(1+ ∈ ) ln( 1−X )− ∈ X] weight of the shell and two 2:1 elliptical heads. The calculated free
A0 on board (FOB) purchase cost is for carbon-steel construction, and
Where FA0 is the initial flow rate of A, CA0 is the initial concen-
the vessel (Cv ) is assumed to be vertical and with a weight range of
tration of A and ∈ is the coefficient of volume expansion.
4200 < W < 1,000.000 lb, and is expressed as follows:
3)Volume of reactor(V ) = 4 D2 L
1) Cv = exp {6.775 + 0.18255*[ln(W)] + 0.02297 x [ln(W)]2 }
4)Length to diameter ratio of reactor(r) = L/D
(Seider et al., 2004)
Where D and L are the internal diameter and length of reactor
As previously stated, the vessel is constructed of A312 alloy steel,
respectively.
the vessel cost must be multiplied by a material factor, which was
5) Residence time of reactor ( ) = vV
0 found to be 1.7 from Seider et al. (2004). To calculate the vessel cost,
6) Design pressure of reactor(Pd ) = exp(0.60608 +
the weight of the shell (W) must be calculated as follows:
0.91615 ln P + 0.0015655(ln P)2 )
2) W = ␲(D + t)(L+0.8 D)t␳ (Seider et al., 2004)
Seider et al. (2004)
Apart from serving as a tool for designing of chemical projects,
Where, v0 is the initial flow rate of A.
the developed methodology can be used for analysis of past process
Pd × D
7) Thickness of reactor wall(t) = safety incidents. This study focuses on the incident of T2 Laborato-
2 × M − 1.2 × Pd
b) Equations for Dow F&EI calculation: ries explosion and fire. The application of described methodology
The main equations for Dow F&EI calculations are as follows: to this incident not only provides a novel view towards the failure
1) F3 = F1 .F2 mechanism but also serves as validation for the methodology.
2) F&EI = MF.F3 T2 Laboratories explosion and fire occurred on December 19,
Where, F1 , F2 , and F3 are the general process hazard factor, 2007, in Jacksonville, Florida (CSB, 2009). T2 Laboratories Inc.
special process hazard factor, and process unit hazard factor respec- manufactured specialty chemicals, mainly gasoline additives. This
tively and MF is the material factor. incident resulted in the death of 4 employees and injured 32
The Dow F&EI index is calculated for the reactor at different (including 4 employees). The 28 public members were working in
values of operating pressure (from 8 to 14 atm) and a sensitivity neighboring areas (CSB, 2009). The explosion occurred in a reac-
analysis is performed as per the procedure described by Suardin tor manufacturing methylcyclopentadienyl manganese tricarbonyl
(2005). (MCMT). MCMT is used as a supplement for leaded gasoline. The
The obtained equation relating Dow F&EI to operating pressure impact of the explosion was so immense that the debris of the
is as follows: exploded reactor was found at a distance of one mile and the shock
3) F&EI = 1.409P + 112.27 waves resulting from explosion damaged buildings at a radius of a
c) Equations for reliability calculation (Arnold, 1972): quarter mile from the facility.
P( D ) Chemical safety board carried out a detailed investigation of the
1) Mean hoop stress in the reactor(Sh ) = t
2
incident. The energy of explosion was estimated to be 1400 tons
p ( D )
2) Standard deviation of hoop stress(h ) = t 2 TNT equivalent (CSB, 2009). The explosion occurred due to a run-
Where, p is the standard deviation of operating pressure. away reaction leading to uncontrolled temperature and pressure
Sh +P elevation. CSB converged on the failure of the cooling system as the
3) Mean induced stress in reactor(Si ) = 2  primary cause leading to the explosion based on witness interviews
4)Standard deviation of induced stress(i ) = 12 p2 + h2
and elimination of other likely causes. CSB conducted laboratory
Failure probability of reactor(F) = = 0.0014(m4 ) − 0.0267(m3 ) + testing of the recipe used by T2 Laboratories and observed a run-
0.178(m2 ) − 0.5011(m) + 0.5055(5) away reaction occurring at a temperature of 390◦ F apart from the
Where, m is an integral parameter given by, desired reaction occurring at 350◦ F. The owners of T2 Laboratories
−Si + Sa were unaware of this runaway reaction leading to the improper
m=   design of cooling and relief systems. The reactor involved in the
i2 + m
2
explosion was constructed in 1962 and bought by T2 Laboratories
in 2001. After purchasing the reactor, certain modifications such
Where ␴m is the standard deviation of yield point. as replacing the nozzles were carried out leading to decrease in
6) Reliability of reactor (R) = 1 − F maximum allowable pressure of the reactor from 1200 psig to 600
Note: The polynomial equation for F is obtained by carrying out psig.
a regression analysis of F and m values as described by Arnold, The methodology described for detailed engineering stage
1972 and represents the overlapping probability of distribution of (Fig. 5) was applied to the incident of T2 Laboratories explosion and
induced and allowable stress. fire with slight modifications. Since the dimensions of the reactor
Following assumptions are made for the reactor design: are known, Dow F&EI and reliability of the reactor can be directly
The reactor length to diameter ratio (r) is 5. evaluated. An important aspect that can be understood is the vari-
Initial flow rate of A (v0 ) is 100 ft3 /s ation in Dow F&EI and reliability of the reactor as the pressure of
The reactor follows ideal isothermal plug flow characteristics. reaction continuously increases due to the occurrence of a runaway
The pressure relief valve is set at 15 atm. reaction.
The standard deviation of operating pressure (p ) is equal to the Following data has been used for application of the developed
mean. methodology to T2 laboratories explosion and fire:
The standard deviation of yield point (␴m ) is 15% of maximum Incident data (CSB, 2009):
allowable stress. Reactor Volume: 2000 gallon
In addition to the reactor dimension, reliability and Dow F&EI Reactor thickness: 3 inches
calculations, the cost of the reactor vessel was calculated using the Operating pressure: 50 psig
method described in Seider et al., 2004. Rate of pressure increase: 32000 psig/min
d) Equations for vessel cost calculation: Maximum allowable working pressure of reactor: 600 psig
The reactor vessel was assumed to be a pressure vessel, cylin-
drical in shape with an inside diameter (D) and consisting of a
 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
Fig. 5. Modified methodology for detailed engineering stage.
Raw materials: Molten sodium metal, methylcyclopentadiene
(MCPD) dimer and diethylene glycol dimethyl ether
5. Application of methodology, results and discussion
The described methodology for initial design stages is applied
to the case of Acetic acid manufacturing in the process selec-
tion stage. Various processes which utilize different chemicals,
reactions and equipment have been developed for Acetic acid
manufacture. The processes considered in this study are Ethylene
oxidation (U.S. Patent No. 7491843 B2, 2009), Acetaldehyde Oxi-
dation (Cheung et al., 2012), Low-Pressure Carbonylation (Cheung
et al., 2012), Ethane Oxidation (Soliman et al., 2012) and Butane
Oxidation (Cheung et al., 2012). The results obtained by applying
the described methodology on the mentioned case study are as
follows:
Analysing the graph obtained between OPSI and corrective
maintenance downtime for process selection stage (Fig. 6), it can be
observed that a process that is inherently safer in terms of involved
reaction, chemicals, and equipment may not necessarily have the
least corrective maintenance downtime. This is primarily because,
in process selection stage OPSI is governed by the inherent safety
principles of minimization, substitution, attenuation and simplifi-
cation (Fig. 4), whereas corrective maintenance downtime depends
only on the number and types of equipment present in the process
system and therefore, only depends on the principle of simplifica-
tion. Therefore, implementing a process that’s inherently safer in
terms of involved reaction, chemicals and equipment might lead
to an increase in the associated risk of the system by increasing its
corrective maintenance downtime.
In 1972, around 10% of the Acetic acid manufacture was carried
out by low-pressure carbonylation. However, due to the economic
superiority of this process, its dominance has increased over the
years and currently more than 90% of Acetic acid manufacture is
achieved using low-pressure carbonylation (Cheung et al., 2012).
Fig. 6. Corrective maintenance downtime (days) vs. OPSI for Acetic acid case study.
Note: The numbers in USD represent the annualized profit obtained from economic
analysis
107
Prior to the dominance of low-pressure carbonylation process, a
major proportion of the Acetic acid was produced by Acetaldehyde
Oxidation (Hintermann and Labonne, 2007). Thus, if regulations to
enforce the consideration of inherently safer designs are imposed,
upcoming facilities might need to reconsider the potential of
Acetaldehyde Oxidation (or any other alternative) for Acetic acid
manufacture over Low-Pressure Carbonylation due to its higher
associated inherent safety. This can lead to the implementation of
a process that has a higher corrective maintenance downtime and
thus ultimately higher associated risk.
Similarly, the described methodology is applied to the case
of Hydrodealkylation of Toluene to Benzene in the conceptual
stage of design. In this stage, different process design alternatives
that are similar in terms of reactions and chemicals but utiliz-
ing different number and type of equipment (i.e., different process
configuration or process flow diagrams) are evaluated in terms of
economic feasibility. For a selected chemical synthesis route, dif-
ferent configurations of process flow can be obtained based on the
implementation of recycling of mass and heat streams and the
feasibility analysis of these alternative configurations is typically
carried out in the conceptual stage. The processes considered are
described by Douglas, 1988; Bouton and Luyben, 2008; Konda et al.,
2006 and Mata et al., 2003. The results obtained by applying the
described methodology to the described case study are as follows:
From the graph obtained for the conceptual stage (Fig. 7), it can
be observed that the corrective maintenance downtime of pro-
cess systems increases with increase in OPSI. A detailed analysis
reveals that the number of equipment involved in the process sys-
tems increases as we move from a process with lower OPSI and
lower corrective maintenance downtime to a process with a higher
OPSI and a higher corrective maintenance downtime. Due to sim-
ilarity of process systems in terms of reactions and chemicals in
the conceptual stage, inherent safety principles of minimization,
substitution and attenuation become redundant, Therefore OPSI is
Fig. 7. Corrective maintenance downtime (days) vs. OPSI for Hydrodealkylation of
Toluene case study.
Note: Numbers above data points indicate the number of equipment and the annu-
alized profit for the design alternative with a capacity for 300 TPA capacity
108 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
essentially governed by the principle of simplification wherein a
process with lesser number of equipment will have a lower value
of OPSI and thus will be inherently safer. Similarly, corrective main-
tenance downtime of process systems increases with the number of
equipment involved in the system. This explains the trend observed
in the results for the conceptual stage (Fig. 7).
The described methodology for detailed engineering stage is
applied to the reactor design problem for the pressure range of
8–14 atm for the described reactor. The results obtained are as fol-
lows:
For detailed engineering stage, As the value of Dow F&EI of the
reactor system increases, the reliability of reactor decreases. This
trend (Fig. 8) exists because as pressure increases, the value of Dow
F&EI increases due to an increase in the value of special process haz-
ard factor (F2). This essentially implies that the process becomes
more hazardous at increasing value of pressures. Thus, ultimately,
the process becomes less inherently safer owing to the principle
of attenuation. Reliability subsequently decreases with increase in
pressure as this leads to an increased overlapping of induced and
allowable stress distributions. Thus, for detailed engineering design
stage, inherent safety and reliability decrease simultaneously with
the pressure of the reactor. An important point to note is that at Dow
F&EI value of 128 (Fig. 8), which is deemed to be intermediately
hazardous as per the guideline (Dow, 1994), the failure probabil-
ity of reactor is as high as 0.25. Thus, in the event of failure of all
layers of protection designed to prevent bursting of the reactor
(such as basic control systems, safety instrumented systems such
as pressure relief valve and others), there is a 0.25 probability of
the reactor bursting. This severity of the situation is not reflected
by the Dow F&EI value. Lastly, the cost of the vessel increases with
increase in the operating pressure, as higher operating pressure
requires a thicker reactor leading to an increased weight of material
of construction for the reactor (Fig. 8).
Similarly, the developed methodology for detailed engineering
stage is applied to T2 laboratories explosion and fire incident and
following results are obtained: Fig. 9.
Focusing on the T2 Laboratories explosion and fire, one of the
contributing causes towards this incident was the decrease in max-
imum allowable working pressure (MAWP) of the reactor from
1200 psig to 600 psig following replacement of certain components
of reactor such as nozzle. Following are the results obtained after
10 centiseconds from the onset of runaway reaction for MAWPs of
1200 psig and 600 psig from the developed methodology.
Application of developed methodology to the case study of T2
Laboratories explosion and fire reveals that although Dow F&EI can
capture the increasing hazard associated with rising pressure from
runaway reaction in the reactor, it fails to increase at the same rate
as the probability of failure of the reactor due to excess pressure. It
is also important to note that, the initial value of Dow F&EI is con-
Fig. 8. Reliability vs Dow F&EI for representative case study of reactor.
Fig. 9. Reliability vs Dow F&EI & Failure probability with increasing pressure due to
runaway reaction for T2 Laboratories case study.
Table 3
Reactor reliability at different MAWPs for T2 Laboratories explosion and fire.
MAWP (psig) Dow F&EI Reactor reliability
1200 246.67 0.9999
600 246.67 0.5588
siderably high since molten sodium (highly reactive in nature and
thus, having a high material factor) is involved in the reaction. Also,
the methodology can predict the decrease in reliability of reactor
(Table 3) due to a decrease in its MAWP which is not reflected by
the Dow F&EI value (since F&EI considers the relief set pressure
and not MAWP of the reactor). This result is in alignment with the
analysis from the CSB report.
From the results obtained for the 3 stages it can be inferred that
the relationship between inherent safety varies with the design
stage under consideration. This relationship depends on the impact
of the 4 inherent safety principles on the parameters of inherent
safety and reliability. In the process selection stage, Since the design
alternatives vary in both chemistry and the equipment involved,
there is possibility to implement all the 4 inherent safety prin-
ciples i.e. substitution (by using less hazardous raw materials),
minimization (by using reactions with higher yield and thus lesser
quantities of raw materials), attenuation (using processes with less
severe operating conditions) and simplification (using processes
with lesser number of equipment). Although it is important to
understand that implementing processes with different raw mate-
rials will lead to the implementation of different process flows i.e.
different types and number of equipment ultimately affecting the
reliability of the system. In conceptual stage, the design alternatives
vary only in terms of process flows and are similar in terms of the
process chemistry. Therefore, the impact of the chemistry depen-
dent inherent safety principles of substitution, minimization, and
attenuation becomes insignificant. As a result, both the parameters
of inherent safety and reliability depend on the principle of simpli-
fication where a process with a lesser number of equipment can be
deemed to be inherently safer and reliable. Lastly, in the detailed
engineering stage, that deals with the design of specific equip-
ment, the impact of the inherent safety principle of simplification
becomes insignificant as the design alternatives vary with respect
to the design of specific equipment and not with the design of pro-
cess flows i.e the number and type of equipment are fixed. The effect
of principles of substitution and minimization becomes insignifi-
 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110
cant as the design alternatives do not differ in terms of chemistry.
However, the effect of the principle of attenuation becomes sig-
nificant as the inherent safety and reliability associated with the
equipment might degrade due to increase in the severity of oper-
ating conditions.
6. Conclusions
In this work, the relationship between inherently safer design
and equipment reliability has been investigated. The aim of this
study was to determine the possibility of risk escalation caused
due to lowering of system reliability during the implementation
of inherent safety principles applied with an objective to lower
the consequence element of risk. This lowering of system relia-
bility can lead to an undesired increase in likelihood element of
risk ultimately leading to an increase in the associated risk. The
study included the development of a safety index (OPSI) for initial
design stages that could effectively reflect the impact of inherent
safety principles (including simplification) on the associated inher-
ent safety of a design alternative. During the analysis of process
selection stage using this index, it was observed that a process
alternative that is inherently safer in terms of the reactions, chem-
icals, and equipment involved, may be associated with a higher
corrective maintenance downtime and therefore may in turn be
riskier since most of the process incidents occur during mainte-
nance, start-up, and shutdown. When a similar analysis was carried
out for the conceptual stage, the number of equipment involved in
a process alternative emerged as a critical factor with regards to
inherent safety and reliability of the alternative. It was observed
that a process with a relatively lesser number of equipment might
be inherently safer (owing to the principle of simplicity) and more
inherently available. Finally, when the required relationship was
studied for a specific equipment such as the reactor during detailed
engineering stage, the severity of operating conditions was deter-
mined to play a crucial role in the associated inherent safety and
reliability of the reactor. Operating at the lesser severity of oper-
ating pressure, the reactor was deemed to be relatively inherently
safer and more reliable, and these parameters were observed to
deteriorate at different rates with an increase in the severity of
operating pressure.
An important conclusion that can be drawn from this study is
that the relationship between inherent safety and reliability is com-
plicated and varies during different stages of design. The impact
of inherent safety principles on the overall inherent safety of the
design alternative is governed by the design stage under consid-
eration, and the cumulative impact of these principles decreases
along with the progression of design resulting in this complicated
relationship. Application of the developed methodology to the case
study of T2 Laboratories explosion and fire revealed that the quan-
tification of inherent safety in terms of non-dimensional indices
(OPSI or Dow F&EI) is plagued with an inbuilt judgement (relative
weighing) of hazards that may not be reflective of the overall risk
associated with a design alternative as observed through reliability
analysis. Thus, implementation of inherent safety design philoso-
phy should be supported by a thorough risk and reliability analysis
to make better safety related decisions in process systems.
Acknowledgements
This research was sponsored by the Mary Kay O’Connor Process
Safety Center at Texas A&M University.
References
American institute of chemical engineers (AIChE), 1994. Dow’s Fire and Explosion
Hazard Classification Guide, 7th ed (New York).
109
Arnold, H., 1972. Pressure Vessel Reliability as a Function of Allowable Stress, Thesis.
Oak Ridge national laboratory, Tennessee.
Bouton, G., Luyben, W., 2008. Optimum economic design and control of a gas perme-
ation Membrane Coupled with the Hydrodealkylation (HDA) Process. Ind. Eng.
Chem. Res. 47 (4), 1221–1237.
Branan, C., 2005. Rules of Thumb for Chemical Engineers: a Manual of Quick, Accurate
Solutions to Everyday Process Engineering Problems. Elsevier, Amsterdam.
CCPS, 2010. Final Report: Definition for Inherently Safer Technology in Pro-
duction, Transportation, Storage, and Use. Prepared by Center for Chemical
Process Safety. The American Institute of Chemical Engineers, for U.S.
Department of Homeland Security, Science & Technology, Chemical
Security Analysis Center, Aberdeen Proving Ground, MD (July 19, 2011
from http://www.aiche.org/uploadedFiles/CCPS/Resources/KnowledgeBase/
IST%20Final%20Definition%20Report.pdf. (Accessed September 23 2011).
Chemical Safety Board (CSB), 2009. Investigation Report. T2 Laboratories, inc. run-
away reaction, Jacksonville, Florida.
Cheung, H., Tanke, R., Torrence, G., 2012. Acetic acid. In: Elvers, B. (Ed.), Ullmann’s
Encyclopedia Industrial Chemistry. , 5th ed. Weinheim, VCH, pp. 209–237.
Cos, N.D., 1973. Application of reliability theory to process design. Chemical Engi-
neering Progress Loss Prevention, vol. 7., pp. 12.
Crowl, D.A., Louvar, J.F., 1996. Chemical Process Safety: Fundamentals with Appli-
cations. Prentice-Hall, Englewood Cliffs, N.J.
Douglas, J., 1988. Conceptual Design of Chemical Processes. McGraw-Hill, New-York.
Duguid, I., 1998. Analysis of past incidents in the oil, chemical and petrochemical
Industries. Loss Prevent. Bull. Inst. Chem. Eng. 142, 3–6.
Ebeling, C., 1997. Introduction to Reliability and Maintainability Engineering, 1st ed.
McGraw-Hill, Long Grove, Illinois.
Edwards, D., Lawrence, D., 1993. Assessing the inherent safety of chemical pro-
cess routes: is there a relation between plant costs and inherent safety? Trans.
IChemE B 71, 252–258.
Exxon Mobil, 2001. Financial and Operating Review.
Freeman, R.A., 1983. Problems with risk analysis in the chemical industry.
Plant/Oper. Progress 2, 185–190.
Gaal, Z., Kovacs, Z., 1985. The reliability of chemical technological systems 111 Uti-
lization of the results of the study. Hung. J. Ind. Chem. 13, 397–404.
Goel, H., Grievink, J., Herder, P., Weijnen, M., 2003. Optimal reliability design of
process systems at the conceptual stage of design. Ann. Reliab. Maintainability
Symp., 40–45.
Griffith, J., Kuenzli, D., Monus, P., 1998. Proactive Manufacturing Acceler-
ating Step Change Breakthroughs in Performance [online], BP, available
at, http://www.marshallinstitute.com/inc/eng/resources/Areas/articles/Body/
articles/98 npra.pdf.
Gruhn, G., Neumann, W., Seidel, R., 1983. Reliability analysis of complex chemical
engineering systems. Hung. J. Ind. Chem. 11, 275–282.
Heikkila, A., 1999. Inherent Safety in Process Plant Design, an Index-based Approach,
Thesis, Technical Research Center of Finland, vol. 384. VIT publications, Espoo
Finland, pp. 129.
Henley, E.J., Gandhi, S.L., 1982. Process reliability analysis. Am. Inst. Chem. Eng. J. 21,
190–194.
Hintermann, L., Labonne, A., 2007. Catalytic hydration of alkynes and its application
in synthesis. Synthesis 8, 1121–1150.
Jobson, S., Watson, D.J., 2009. U.S. Patent No. 7491843 B2. Washington, DC: U.S.
Patent and Trademark Office.
Kidam, K., Hassim, M., Hurme, M., 2008. Enhancement of inherent safety in chemical
industry. Chem. Eng. Trans. 13, 287–294.
Kletz, T.A., 1978. What you don’t have, can’t leak. Chem. Ind., 287–292.
Kletz, T.A., 1991. Plant Design for Safety: a User-friendly Approach. Hemisphere
Publishing Corporation, New York.
Konda, N., Rangaiah, G., Lim, D., 2006. Optimal process design and eff;;ective
plantwide control of industrial processes by a simulation-based heuristic
approach. Ind. Eng. Chem. Res. 45 (17), 5955–5970.
Krishnaswamy, R., Parker, N.H., 1984. Corrective maintenance and performance
optimization. Chem. Eng., 93–98.
Mannan, S., 2002. Challenges in Implementing Inherent Safety Principles in New &
Existing Chemical Processes, Mary Kay O’Connor Process Safety Center.
Mata, T., Smith, R., Young, D., Costa, C., 2003. Evaluating the environmental friendli-
ness, economics and energy efficiency of chemical processes: heat integration.
Clean Technol. Environ. Policy 5 (3), 302–309.
Palaniappan, C., Srinivasan, R., Tan, R., 2004. Selection of inherently safer process
routes: a case study. Chem. Eng. Process. 43, 647–653.
Passut, C., 2017. At Industry’s Request, EPA Stays Risk Management Program
Rule for Three Months from, March 14. March 17, 2017, from http://www.
naturalgasintel.com/articles/109758-at-industrys-request-epa-stays-risk-
management-program-rule-for-three-months (Accessed March 17 2017).
Peters, M.S., Timmerhaus, K.D., 1991. 4th ed. In: Plant Design and Economics for
Chemical Engineers. McGraw-Hill, New York.
Rathnayaka, S., Khan, F., Amyotte, P.R., 2014. Risk-based process plant design con-
sidering inherent safety. Saf. Sci. 70, 438–464.
Reniers, G., Talarico, L., Platrinieri, N., 2016. In: Paltrinieri, Nicola, Khan, Faisal (Eds.),
Cost-Benefit Analysis of Safety Measures. Dynamic Risk Analysis in the Chemical
and Petroleum Industry. Butterworth-Heinemann, Amsterdam, pp. 195–205.
Roy, N., Eljack, F., Gutierrez, A., Zhang, B., Thiruvenkataswamy, P., Halwagi, M., Man-
nan, M.S., 2016. A review of safety indices for process design. Curr. Opin. Chem.
Eng. 14, 42–48.
Rudd, D.F., 1962. Reliability theory in chemical system design. Ind. Eng. Chem. Fund.
1, 138–143.
110 N. Ade et al. / Process Safety and Environmental Protection 117 (2018) 100–110

Seider, W.D., Seader, J.D., Lewin, D.R., 2004. Product and Process Design Principles:
Synthesis, Analysis, and Evaluation. Wiley, New York.
Shell, 2001. Impact Magazine Issue 2.
Soliman, M., Al-Zeghayer, Y., Al-Awadi, A.S., Al-Mayman, S., 2012. Economics of
acetic acid production by partial oxidation of ethane. APCBEE Proc. 3, 200–208.
Suardin, J., 2005. The Integration of Dow’s Fire and Explosion Index into Process
Design and Optimization to Achieve an Inherently Safer Design, Thesis. A&M
University, Texas.
Towler, G., Sinnott, R., 2013. Chemical Engineering Design. Butterworth-
Heinemann, Boston, pp. 3–32, http://dx.doi.org/10.1016/B978-0-08-096659-5.
00001-8, ISBN 9780080966595.