Computers and Chemical Engineering 35 (2011) 2786–2798

Contents lists available at ScienceDirect

Computers and Chemical Engineering journal homepage: www.elsevier.com/locate/compchemeng

A sustainability root cause analysis methodology and its application Abhishek Jayswal a , Xiang Li a , Anand Zanwar a , Helen H. Lou a,∗ , Yinlun Huang b a b

Dan F. Smith Dept. of Chemical Engineering, P.O. Box 10053, Lamar University, Beaumont, TX 77710, United States Department of Chemical Engineering and Materials Science, Wayne State University, Detroit, MI 48202, United States

a r t i c l e

i n f o

Article history: Received 20 January 2011 Received in revised form 5 April 2011 Accepted 9 May 2011 Available online 14 May 2011 Keywords: Root cause analysis Sustainability analysis The Pareto analysis Fishbone diagram Biodiesel

a b s t r a c t In the design of chemical/energy production systems, a major challenge is to identify the bottleneck issues and improve its sustainability effectively. Due to the multi-dimensional feature of sustainability, how to account for the impacts of various design factors and the cause-and-effect relationships can be very difficult. This paper will present a sustainability root cause analysis method based on the combination of Pareto Analysis and Fishbone diagram. The sustainability of the process is assessed incorporating economic, environmental, societal and efficiency concerns. This methodology is able to help the designers focus the attention on the most important fundamental causes, discover opportunities for sustainability improvement and provide critical guidance to design for sustainability. The efficacy of this methodology will be demonstrated through a case study on a biodiesel production technology. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction Triple-bottom-line is the basic concept of sustainability, demanding a balance among economic, environmental, and social sustainability (Pintariˇc & Kravanja, 2006). The chemical industry, like other manufacturing industries, has been facing tremendous challenges due to economic globalization, environmental pressure, natural resource depletion, etc. The industry fully recognizes its commitment to product stewardship and sustainable development (Turton, Bailie, Whiting, & Shaeiwi, 2003). A sustainable design, as an evolving popular topic, should be based on the conception and realization of the full consequence of a particular economic activity on the environmental and societal dimensions. Implementing sustainability philosophy in early design stage requires much less effort and cost than to retrofit the process after its launch. However, in the design of chemical/energy production systems, a major challenge is to identify the bottleneck issues and improve its sustainability effectively. Due to the multi-dimensional feature of sustainability, how to account for the impacts of various design factors and the cause-and-effect relationships can be very difficult. This paper will present a sustainability root cause analysis methodology based on the combination of Pareto Analysis and Fishbone diagram. The Pareto analysis is used to identify the major causes while the complicated cause–effect relationships are illustrated by a Fishbone diagram. As shown in Fig. 1, the sustainability of the

∗ Corresponding author. Tel.: +1 409 880 8207; fax: +1 409 880 2197. E-mail addresses: [email protected], [email protected] (H.H. Lou). 0098-1354/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.compchemeng.2011.05.004

process is assessed incorporating economic, environmental, and societal concerns. This methodology is able to help the designers focus the attention on the most important fundamental causes, discover opportunities for sustainability improvement and provide critical guidance to design for sustainability. The efficacy of this methodology will be demonstrated through a case study on a biodiesel production technology. The conventional biodiesel production process is analyzed and the results are used to derive an improved process. 2. Sustainability root cause analysis The development of suitable sustainability assessment and design decision methodologies are of upmost importance. Due to the sophistication of processes, the fundamental causes leading to an inferior sustainability performance could be hiding deep down from the surface. This research aims at developing a novel yet practical methodology for conducting sustainability root cause analysis on chemical and energy production systems. The root-cause analysis framework is built on the combination of Pareto chart and the Fishbone diagram, in conjunction with a set of sustainability metrics for conducting comprehensive sustainability assessment on complex chemical and energy production systems along each dimension of sustainability. Due to the multi-dimensional nature of sustainability, a good root-cause analysis method depends on an appropriate sustainability assessment system. The procedures for sustainability assessment and root cause analysis in this study are shown as Fig. 2. When the designers are facing more than two designs, a series procedure is recommended (Li, Zanwar, Jayswal,

A. Jayswal et al. / Computers and Chemical Engineering 35 (2011) 2786–2798

2787

Nomenclature 0 Ech Ech Emix Eph Etot Exin Exout Exloss F H P Q S T

standard chemical exergy (J/mol) chemical exergy (W) mixing exergy (W) physical exergy (W) total exergy (W) inlet stream exergy (W) outlet stream exergy (W) exergy loss (W) mole flow rate (mol/s) enthalpy (J/mol) pressure (Pa) heat or cooling duty (W) entropy (J/(mol K)) temperature on thermodynamic scale

Superscripts 0 denotes standard state at normal temperature and pressure

Lou, & Huang, 2011). In this approach, the economic viability is checked first and non-profitable processes can be eliminated from further consideration. The next step is to evaluate the inbound safety, which represents the societal impacts. For chemical industries, social sustainability encompasses many aspects, including chemical safety, process equipment safety, possibility of working accidents, occupational disease, toxicity potential of the process, etc. However, the assessment of all these factors will be difficult during the early design stage of a chemical process, when only limited information is available. Moreover, the possibility of working accidents, occupational disease, toxicity potential of the process are all rooted from the inherent safety aspects of the chemical and the process equipment. In this work, we used a simpler method which focuses on the inherent chemical and process equipment safety to quantify the social dimension of a process. Then the various environmental impacts will be evaluated. This sequence is set because the designer needs to check the inbound safety issues first. If the inbound safety can not be satisfied, the potential environmental impacts must be undesirable. The efficiency is checked last because there are strict regulations on safety and environmental issues, but no such regulations on efficiency. This series procedure helps the designers to improve the efficiency of screening by eliminating non-viable processes systematically at the early design stage. The root cause analysis (Anderson & Fagerhaug, 2006) is originally a problem solving technique used in quality management (Wilson, Dell, & Anderson, 1993) and is a powerful tool to pinpoint the obstacles to improvement. It is usually used in a reactive mode to determine the causes of the problems which have already occurred. The root cause analysis can be performed using many tools such as 5 Whys (Ammerman, 1998), the Pareto chart (Surhone, Timpledon, & Marseken, 2010) and the Fishbone diagrams (Robitaille, 2004). 2.1. The Pareto analysis The Pareto analysis (Surhone et al., 2010), which is also known as 80–20 rule, is named after the Italian economist Vilfredo Pareto. The principle states that for many events, roughly 80% of the effects/problems come from 20% of causes. The Pareto analysis helps focusing the attention on the most important causes instead of wasting time and energy on minor ones. A combination of line and bar chart is prepared to identify the top 20% problems. The following steps need to be followed to prepare the chart:

Fig. 1. Sustainability assessment of chemical/energy production system.

1. A table is prepared showing all causes with their impacts in percentage. 2. The table is sorted out in descending order by percentage. 3. A third column is added to show cumulative percentage. 4. A line chart is prepared. The causes are on x-axis and their cumulative percentage values are on y-axis. 5. On the same chart a bar chart is added by causes on x-axis and percent impacts on y-axis. 6. A horizontal line is drawn at 80%. At the intersection of the 80% line with curve, a vertical line is added. 7. The point of intersection of the vertical line and x-axis separates the major causes to the left side and minor causes to the right side. This technique helps the users to identify the top causes that need to be addressed to resolve the 80% of the problems. Once the major causes are identified, tools like the Ishikawa diagram or Fish-bone diagram can be used to illustrate the root causes of the problems. Then efforts can be made to remove the major obstacles in order to develop a more sustainable process. This study is directed towards sustainability root cause analysis of chemical/energy production systems, considering the economic performance, environmental impacts and societal concerns as the major criteria for the analysis. Pareto charts are prepared for all the criteria to identify major causes and the Fish-bone diagram is prepared to illustrate the problems. 2.1.1. Comparing sensitivity analysis with sustainability root cause analysis Generally, sensitivity analysis is a technique used to determine how different values of an independent variable will impact a particular dependent variable under a given set of assumptions (Saltelli et al., 2008). In chemical engineering, sensitivity analysis is generally used to optimize or analyze a process flow sheet under different operating scenarios. In this work, the scope of process analysis is extended to include the triple aspects of sustainability, which is broader than the traditional economic-centered process analysis. On the other hand, traditional sensitivity analysis does not include Pareto analysis to prioritize the importance of the contributing factors, nor does it include Fishbone analysis to visualize the complicated relationships. With this sustainability root cause analysis methodology, the designers are equipped with a new tool in the exploration of new designs. In the mean time, root cause analysis and sensitivity analysis can be complementary. Once the critical issues are identified using the root cause analysis method, the designers can use sensitivity analysis to check how much the

2788

A. Jayswal et al. / Computers and Chemical Engineering 35 (2011) 2786–2798

Fig. 2. Sustainability root cause analysis flow sheet.

change on a root cause factor can affect the final sustainability performance.

and the societal concerns are used as the major criteria and are discussed as follows. 2.3. Economic matrix

2.2. The Fishbone diagram The Fishbone diagram (Kelleher, 1995), which is also known as Ishikawa diagram, was proposed by Kaoru Ishikawa in 1960s, who founded quality management processes in Kawasaki shipyards. It is considered one of the seven basic tools of quality control. The Fishbone diagrams are used for the representation of the major problems in a process. In this study, the results of the Pareto analysis are used to prepare the Fishbone diagram to represent the major causes which drive the process away from being sustainable. The Fishbone diagram helps to visualize and convey the important relationships between the seemingly disconnected elements. For the purpose of preparing the Fishbone diagram and the Pareto analysis, the economic analysis, the environmental impacts

About half of the papers found in a study by Pintariˇc and Kravanja were using cost and profit and only 10% were using net present value (NPV) as the economic objective function (Pintariˇc & Kravanja, 2006). The internal rate of return (IRR) with NPV is considered as the most useful indicator for economic performance by Othman, Repke, Wozny, and Huang, 2010 and Li et al. (2011), since IRR and NPC includes initial investment, annual profit, annual depreciation, salvage value and interest on investment. When using NPV for profit calculation, a positive value means that the project is feasible and negative values indicate otherwise. Therefore, when comparing between alternatives, one with the largest positive NPV value will be the best choice. IRR is designed to reflect the highest, after-tax interest or discount rate at which the project can just

A. Jayswal et al. / Computers and Chemical Engineering 35 (2011) 2786–2798 Table 1 Impact categories in WAR algorithm (Pintariˇc & Kravanja, 2006). General impact category

Human toxicity Ecological toxicity

Global atmospheric impacts Regional atmospheric impacts

Impact category

Measure of impact category

Ingestion Inhalation/dermal Aquatic toxicity

LD50 OSHA PEL Fathead Minnow LC50 LD50 GWP ODP AP PCOP

Terrestrial toxicity Global warming potential Ozone depletion potential Acidification potential Photochemical oxidation potential

break even (Turton et al., 2003). Apparently, a project that yields IRR with higher value than the cost of capital is considered to be profitable. In the following case study, only the annualized capital cost and operating cost are used for the evaluation of economic performance because the product is always the same: the biodiesel (Seider, Seader, & Lewin, 2004). The capital cost is calculated using Aspen Icarus software while the operating cost is calculated based on the mass and energy flow information obtained by Aspen Plus simulation. The cost of chemicals and utilities is based on the market prices shown in Table 4. Results of the root cause analysis are used to develop a new heterogeneous catalyst process and the economic performance of both processes is compared using NPV and IRR. 2.4. Environmental matrix Life cycle analysis (LCA) is a powerful method for assessing the environmental impact of products, processes and activities from “cradle to grave” and the EPA waste reduction algorithm is a typical LCA matrix to quantify the various environmental impacts of a process. It must be pointed out that the existing indicator sets does not include a very important type of efficiency analysis – exergy analysis. The authors argue that exergy analysis is particularly useful for the assessment of energy/fuel production systems. Therefore, the environmental metrics in this study consists of two subcategories: various environmental impacts and the efficiency analysis. 2.5. Environmental impacts The LCA method uses the material flow and energy flow information to assess a process from “cradle-to-grave”, i.e., from raw material acquisition through production, use, and final disposal. In 1999 Young and Cabezas (US EPA) introduced a specific and precise methodology namely WAR algorithm to assess the environmental impact of chemical process (Young & Cabezas, 1999). The environmental impacts at the manufacturing stage can be analyzed using WAR algorithm for the overall life cycle of the chemical production process. The WAR algorithm uses the potential environmental impacts of chemicals to provide a relative indication of environmental friendliness or unfriendliness of the chemical process. The PEI indexes consider the mass and energy balance to evaluate environmental impacts of the process. A database of relative environmental impact scores has been created and embedded in the WAR software. As shown in Table 1, the eight different categories are divided into four general environmental impact categories (Young & Cabezas, 1999). The scores given by the WAR software helps the users to compare different process alternatives based on the potential environment impact of the process. The lower PEI index represents

2789

more environmentally friendly process. However, the design is always constrained by by-products, environmental impacts of the waste and the energy consumption. The WAR algorithm and any other indicator of environmental impact don not consider the efficiency of the process. In this study, the authors suggest use of efficiency analysis to compliment WAR algorithm and complete analysis of the environmental impacts. 2.6. Indicators on efficiency performance As we all know, the first law of thermodynamics states that energy can neither be created nor destroyed; it just changes its forms. The law does not provide enough information about the potential work producible by a form of energy or that lost during energy transformation. Note that different types of energy display different qualities. In thermodynamics, the exergy of a system is the maximum useful work possible during a process that brings the system into equilibrium with a heat reservoir. Exergy analysis will allow accounting for irreversibility in a process and provides a more detailed tracking mechanism for energy and chemical generation and consumption. Defined as the maximum amount of work available from a stream, exergy analysis differentiates the qualities of energies and chemicals (Szargut, 2005; Szargut, Morris, & Steward, 1988). A chemical/energy production system has three types of streams: the material streams, the power streams, and the heat flows. The exergy of a material stream should be a sum of chemical exergy, physical exergy and the mixing exergy. The exergy of a power stream (electricity) is simple to calculate, as 1 kW of electricity represents 1 kW of exergy. The exergy of a heat stream with a constant temperature at the reference pressure can be obtained based on the concept of Carnot efficiency. The calculation method is detailed in the seminal work by Szargut (2005) and Szargut et al. 0 ) can be (1988). The standard chemical exergy of a compound (Ech calculated from the standard formation enthalpy and the Gibbs energy (or obtainable in standard tables). The chemical exergy of a multi-component stream (Ech ) can be evaluated by calculating the 0 ) of component i summation of the standard chemical exergy (Ech,i multiplied by its molar flow rate (Fi ) as follow: Ech =



0 (Fi × Ech,i )

The physical exergy (Eph ) quantifies the amount of work arising when a stream (containing unmixed components) changes from a process condition (T, P) to the reference state (T0 , P0 ), i.e.,



Eph = [

−

(Fi × Hi ) −







(Fi0 × Hi0 )] − T0 × [

(Fi × Si )

(Fi0 × Si0 )]

In addition, the mixing exergy (Emix ) is the amount of work when a stream (consisting of pure components) is mixed at the process condition (T, P). It can be calculated as: Emix = [F × H −



(Fi × Hi )] − T0 × [F × S −



(Fi × Si )]

Therefore, the total exergy of a material stream at the process conditions can be derived as: Etot = Ech + Eph + Emix The exergy of a power stream (electricity) can be readily calculated, as 1 kW of electricity represents 1 kW of exergy. The exergy of a heat flow with a constant temperature at the reference

2790

A. Jayswal et al. / Computers and Chemical Engineering 35 (2011) 2786–2798

pressure can be obtained based on the concept of Carnot efficiency, i.e., Ex =



1−

T0 T



×Q

where, Q is the total amount of heat or cooling duty. When dealing with a heat stream with a linear temperature change from Tl to Th without phase transition, the following expression applies:



Ex =



1 − T0 × ln

Th /Tl Th − Tl



×Q

It is known that exergy loss (Exloss ) due to irreversibility is the exergy difference between the input (Exin ) and the output (Exout ). The exergy efficiency () of a system is a ratio of Exout to Exin . Exout = Exin According to this efficiency, those process steps with a relatively large exergy loss can be pinpointed, and system improvement opportunities can be readily identified. In this project, energy efficiency and exergy efficiency will be used as an integral part of the overall environmental sustainability assessment. 2.7. The societal impacts Social indicators involve many soft criterions, which are based on human intuitive and it is normally influenced by the decision maker’s knowledge and experience. It is difficult to quantify these criteria in numerical equations but it can be scaled based on specific and distinctive measurement. Herder and Weijnen (1998) performed a study to explicitly define soft quality criteria in process design decision making. Since some criteria, such as acceptable for environment, efficient use of raw material and total life cycle aspects have been already included in the environmental impact assessment and efficiency analysis, the authors view social assessment consideration as mainly about safety analysis in the chemical process. Safety is the second nature of chemical industry. To prevent casually and injury is of paramount importance. In addition, there are always legal issues and concerns on companies’ image for which safety should be considered at the design stage. To evaluate safety at the design stage will help minimize potential undesirable consequences. Safety analysis is a systematic examination of the structure and functions of a process system, aiming at identifying potential accident conditions, evaluating the risk and identifying measures to mitigate or eliminate risk. There are many safety analysis methods available. Dow Fire and Explosion Hazard Index and Mond Index are two of the widely used methods in process industries (Mond, 1985; Dow, 1987). These indices are mainly related to fire and explosion rating of a plant and are best suited at later design stage when equipment, chemicals and process conditions are known. Another accurate method for safety analysis is HAZOP (Hazard and Operability Analysis). Normally HAZOP studies are conducted using P&ID (Piping and Instrumentation Diagrams) to find out possible process disturbances and their consequences. Since a very minute detail about plant operating conditions and control is required for HAZOP study, it is not suitable to be used in early design stage. Prototype Index of Inherent Safety (PIIS) developed by Edwards and Lawrence (1993) is also a good safety analysis method (Edwards & Lawrence, 1993). PIIS is mainly used to analyze raw materials used and sequence of the reaction steps. This method focuses on reaction, and it is not suitable for safety analysis of the whole process plant. Details of process plant operations and control are normally not fully available at the early design stage. Nevertheless safety of all the chemicals and

Table 2 Structure of Inherent Safety Index (Heikkila, 1999).

Chemical inherent safety index, ICI Heat of main reaction Heat of side reaction, max Chemical interaction Flammability Explosiveness Toxic exposure Corrosiveness Process inherent safety index, IPI Inventory Process temperature Process pressure Equipment safety Isbl Osbl Safe process structure

Symbol

Score

IRM IRS IINT IFL IEX ITOX ICOR

0–4 0–4 0–4 0–4 0–4 0–4 0–2

II IT IP IEQ

0–5 0–4 0–4

IST

0–4 0–3 0–5

equipment should be considered. In choosing any safety analysis method, requirement of minimum data and coverage of all the aspects of safety should be the main criteria. Heikkila (1999) proposed a very good safety analysis method named Inherent Safety Index, which requires less information compared to other methods while it covers many aspects of safety (Heikkila, 1999). As suggested by Heikkila in their report, this method is suitable for comparison between various alternative designs of a process. In inherent safe design, the main principles are to avoid the use of hazardous material and aim for a simpler process. The possibility of affecting the inherent safety of a process decreases as the design proceeds and more and more engineering and financial decisions have been made. Implementing the principles of inherent safety during the conceptual design phase will help the designers root out inferior designs at the earliest stage. As shown in Table 2 (Heikkila, 1999), calculation of Inherent Safety Index is divided into two sub-indices: Chemical Inherent Safety Index and Process Inherent Safety Index. These sub-indices are further divided into other sub-indices. Chemical Inherent Safety Index covers parameters related to hazards presented by the chemicals in the plant and process Inherent Safety Index deals with the hazards due to equipment and inventory in the plant. Scores are given for sub-indices based on the parameters of the individual components. Calculated by adding all the sub-indices together, the resulting total Inherent Safety Index can be used to compare the safety of different designs. A lower score indicates safer design. Nevertheless, the original Inherent Safety Index method has its own limitation since it does not count the complexity of the process or the quantity of the chemicals used in the process. So, an Enhanced Inherent Safety Index (EISI) method was proposed by the research team to enhance its functionality (Li et al., 2011). In the EISI method, again two sub-indices, i.e., the Chemical Inherent Safety Index and the Process Inherent Safety Index are given similar to ones in the Inherent Safety Index. In Chemical Inherent Safety Index, the scores are calculated by multiplying the severity (i.e., flammability) of the chemicals with the flow rates of the chemicals instead of considering only maximum value. All the scores for individual chemicals are added together to obtain the total Chemical Inherent Safety Index. In the Process Inherent Safety Index, scores are given for individual equipment and multiplied by the number of pieces of equipment. The scores of all pieces of equipment are added together to get the total Process Safety Index. 2.8. Normalization of the indexes Normalization is the process of organizing data in a database. It is a systematic way of ensuring that a database structure is suitable

A. Jayswal et al. / Computers and Chemical Engineering 35 (2011) 2786–2798

for general-purpose querying and free of certain undesirable characteristics. As different indicators have different values, which may spread over several orders of magnitude, it is necessary to normalize the values to a unified scale to compare them easily. Most often, weights to each criterion are chosen subjectively for each indicator. Normalized vector value indicates the percentage importance of a criterion in the overall decision of its parent level. In the following case study, all the values will be normalized from 0 to 1. 3. Illustrative case study: biodiesel processes Biodiesel is a renewable fuel suitable for both centralized and distributed production at a wide range of scales. Currently, the most common method to produce biodiesel is by transesterification reaction of vegetable oils (e.g., canola, palm, jatropha, palm kernel, sunflower, and waste vegetable oils), the main component of which is triacylglycerol (TAG). When TAG reacts with alcohol, fatty acid methyl esters (FAME, or biodiesel) and glycerol will be generated. The catalyst can be either enzymes, acids, or bases (Zabeti, Daud, & Kheireddine, 2009). The transesterification reaction is illustrated as follows (Singh, 2008): H2 C-OCOR´ ROCOR´ H2C-OH Catalyst HC-OCOR´´ + 3ROH ROCOR´´ + HC-OH H2 C-OCOR´´´ Triglyceride

Alcohol

ROCOR´´´

H 2C-OH

Mixture of alkyl esters

Glycerol

The conventional homogeneous process, which produces biodiesel using alkali based catalyst. Alkali-catalyzed transesterification (also known as alcoholysis) uses an alkali such as NaOH or KOH as catalyst to convert TAG into biodiesel (Hass & McAloon, 2006). The preferred methanol to oil molar ratio is 6:1. At 65 ◦ C, a 93–98% conversion of TAG is achieved within 1 h. In this case study, the process simulator Aspen Plus is used. The outputs of process simulators provide the inventory data needed to perform process assessment and selection. A flow sheet of the conventional process is provided in Fig. 3 (Aspen, 2008). In a conventional process, vegetable oil is mixed with liquid caustic catalyst and methanol in a plug flow reactor namely “REACTOR”. The un-reacted methanol is then separated from product mixture in a distillation column “MEOHCOL”. The product biodiesel is water washed in the separator “WASHCOL” and the final product is obtained after separating un-reacted vegetable oil in a distillation column “ESTCOL”. The aqueous phase from WASHCOL separator is neutralized by H3 PO4 in a reactor “NEURT” and an additional distillation column “GLYCRCOL” is required to separate by-product glycerol from water and methanol. In comparison to both the enzyme- and acid-catalyzed transesterification reactions, the high yield in a relatively short reaction time and the relative low cost makes the alkali-catalyzed method the dominating production method in current industrial practice. The transesterification reaction requires a low water (