A holistic framework for environment conscious based product risk modeling and assessment using multi criteria decision making

Abstract

This article presents a holistic framework for environment conscious based product risk modeling and assessment. The attributes of product risk assessment are identified. The degree of interrelationships of the identified attributes is also established. A linked structure called product risk assessment digraph is developed to show various interrelationships among the identified attributes. For analysis of this linked structure, the concept of matrix is used. This facilitates the development of an environment conscious based product risk assessment index. Various product design alternatives are analyzed from risk perspective. The proposed methodology will facilitate product designers, manufacturing engineers, environmental analysts and risk experts in design and development of environment conscious based product. Two examples have been shown here in the present work to support the proposed methodology. Example one is to illustrate the proposed framework and second example is for validation purpose.

Introduction

The design of engineering systems is generally carried out using complex analytical and mathematical models which are integrated and operated under inescapable risk environment in a system development process (Haimes, 2008). This complexity is often expected to encounter multiple and conflicting objectives. The product designers must take into consideration the risk parameters at the conceptual design stage in a systematic manner. Lowrance (1976) defined the term risk as a measure of the probability and severity of adverse effects. An attempt was made to highlight the distinction between risk and safety. It was observed that measuring risk is an empirical, quantitative and a scientific activity. Aven, 2010, Aven, 2011, Aven, 2012, Aven, 2016, Aven and Vinnem, 2007 gave a historic analysis of risk, starting from definition and categorizing the risk and its various phases. Aven, 2011, Aven, 2016 described risk in general, into 9 categories, ranging from D1 to D9. The first category of risk i.e., D1 defined risk as expected value of the probability of an event occurrence and the utility of the consequences. The second category of risk i.e., D2 defined risk as the probability of an undesirable event, or the chance of a loss. In the similar manner, various definitions of risk were defined, D3 to D9.

In a product design and development process, risk analysis/mitigation has been a major concern for product designers during the past few years. According to Maarten Bonnema and Van Houten (2006), a product development process is divided into four phases. These are task clarification, conceptual design, embodiment design and detailed design. It is a fact that in a product design process 80% of the product cost is determined during the conceptual design phase (Cooper and Thompson, 2002, Li and Li, 2000, Okamura and Aoyama, 2009). The growing safety concern in design and development of products has gained much attention of designers and manufacturers during the past sometime. Moreover the strict regulations to improve product's performance, efficiency and, reliability are a greater challenge for product designers (Koh et al., 2007). Even after having the knowledge of process safety, availability of tools for analysis and safety management systems, accidents still take place in industries. It means, there is involvement of risk factor (s) in various products or sub products (Hollnagel, 1998, Weick and Sutcliffe, 2011). The issue of process safety which involves consideration of risk factor involves identification of all possible hazard scenarios and the related consequences using tools such as Hazard and Operability (HazOp), Layer of Protection Analysis (LoPA), Failure Mode and Effect Analysis (FMEA), etc. (Weick and Sutcliffe, 2011). It is also a fact that due to increased competitive scenario, complexity in product design and ever changing customer requirements, the product designers are under tremendous pressure to design and develop products with minimum or no risk. In a concurrent engineering (CE) approach involving design and development of complex products, the associated risk factors have impact on several strategic, financial and quality aspects (Ekanem et al., 2016, Mamtani and Green, 2006). The lack of exposure of the team comprising of designer, manufacturer and others results in high level of risk during the design and development of products at conceptual stage (Swain and Guttmann, 1983). Thus, the failures in a product cannot be completely eliminated, therefore while designing a safety – critical product, a thorough analysis of the potential consequences of its failures must be carried out by the designer. It also becomes the duty of the designer to focus on how to reduce the frequency of failures (Tseng et al., 2003).

A number of research investigations have been carried out in the area of risk analysis during the recent past. A relative reliability risk assessment study for original designs using the Analytical Hierarchy Process (AHP) technique during conceptual design phase was investigated by (Mamtani and Green, 2006). García et al. (2016) studied the Statistical equation modeling analysis for industrial projects. The study primarily focused on design of critical factors considering quality, cost, time and, success as the basis for analysis (García et al., 2016). Goswami and Tiwari (2015) proposed a methodology for product design evaluation. The methodology was aimed at product commercialization by considering qualitative consumer review-based product design characteristics. Zhao et al. (2016) proposed an economic input–output based hybrid life cycle assessment methodology which calculates the environmental impact of various types of trucks used for delivery purpose. While carrying out the LCA, three basic criteria for analysis were considered. These include: economic considerations, environment and, public health. Xiao et al. (2012) integrated the risk based supplier selection problem using fuzzy cognitive map (FCM) and fuzzy soft set approach. Three risk criteria for selection and evaluation of a best supplier on the basis of risk were selected for this purpose (Xiao et al., 2012). Kayis et al. (2007) studied the risk quantification for new product design and development in a Concurrent Engineering (CE) environment. This led to the development of an Intelligent Risk Mapping and Assessment System (IRMAS™) which facilitates the quantification of risks at all stages of the product life cycle in a systematic way. Khan (2001) proposed a model for dynamic risk assessment using bow-tie (BT) approach in which the occurrence probability of accident consequences changes. A failure probability model for primary events of BT was developed, with the help of physical reliability models.

Chan et al. (2012) investigated the integration of failure analysis and risk analysis with quality assurance during the design phase of medical product development. The concept of quality assurance which takes into consideration risk analysis and failure analysis in product was used for this purpose. Zhu et al. (2017) proposed a risk based decision-making framework using interval numbers. The investigation involved an analysis of a risk based study considering static and dynamic situations, systematically applied to a supplier selection problem for key components of an aircraft. Bounit et al. (2016) carried out a research study involving assessment of the overall risk of machines using fuzzy system. Factors like integration of hygiene, safety, and environment system to assess the risk in a machine, were systematically analyzed to evaluate the overall risk. Ekanem et al. (2016) proposed a Qualitative Analysis Procedure called Phoenix, where it focused mainly on Human Reliability Analysis (HRA). The HRA methodology was used for carrying out Probabilistic Risk Assessment (PRA). There have been several developments in the application of first generation HRA techniques. These include Technique for Human Error Rate Prediction – THERP (Swain and Guttmann, 1983), Human Error Assessment and Reduction Technique – HEART (Williams, 1986) and, Success Likelihood Index Method Multi attribute Utility Decomposition – SLIM-MAUD (Embrey et al., 1984).

In addition to these, several other techniques and frameworks called second generation or advanced methods were developed. These include: Cognitive Reliability and Error Analysis Method – CREAM (Weick and Sutcliffe, 2011), Standardized Plant Analysis Risk Human Reliability Analysis – SPAR-H (Gertman et al., 2005) and, Information, Decision and Action in Crew context – IDAC (Chang and Mosleh, 2007). The apex body of risk assessment and management in United States have gained much importance and attention of researchers during the past few decades and thus described risk assessment as a decision making process, which makes use of empirical and quantitative tools for assessment (Lam et al., 2011). Ercan (2013) developed a fuzzy-data envelopment analysis (DEA) methodology for sustainability analysis of Intelligent Transportation Systems (ITS). The study involved the integration of triple bottom line (TBL) for sustainability analysis of ITS. An attempt was made to provide a holistic overview of socio-economic and environmental benefits by analyzing the ITS and sustainability performance comparison using the decision support system.

Radke et al. (2013) proposed a risk management based evaluation procedure of inventory allocations for make to order production. Schedule delays formed the basis for assessment of risk in such type of production environment. Radke et al. (2013) reported that material stock outs are one of the significant risks that affect the production activities. Ercan et al. (2015) proposed an optimization model based on multi-objective linear programming (MOLP), which facilitates determination of life cycle impact assessment (LCIA) for fleet of buses using alternative fuel options in USA. The LCIA study considered diesel, hybrid, electric battery, B20, CNG and, LNG buses for analysis. Bare (2006) carried out a research study on risk assessment and Life-Cycle Impact Assessment (LCIA) for human health cancerous and noncancerous emissions: integrated and complementary with consistency within the United States Environment Protection Agency (USEPA). A four stage risk assessment process considering Life Cycle Assessment (LCA) and risk assessment was the basis for analyzing such a risk assessment based study. These phases include: hazard identification, exposure assessment, dose response and, risk characterization (Bare, 2006).

It is clear from the aforementioned risk based research investigations that several efforts were made during the past few decades for product risk assessment (PRA) at design stage, particularly at conceptual design phase. The above research investigations also reveal that only some of the PRA attributes had been considered earlier by researchers. There is however, a need for a much critical analysis to ensure that more risk assessment attributes are analyzed at product design stage. Furthermore, some of the researchers have also considered PRA as a multi criteria decision making (MCDM) problem. The design criterion in risk assessment facilitates designers in analyzing the parameters. Therefore, the risk attributes such as Fictitious Need and Analysis (FNA), Problem Definition (PD), Design Solutions (DS), Physical Object and Social Acceptability (POSA), Economic and Financial Viability (EFV), Reliability and Safety (RS), Sustainability and Energy Conservation (SEC) and, Disposability and Recyclability (DR), are considered as criteria for evaluation/comparison of different design alternatives in a risk assessment process at conceptual design stage.

Based on the above literature review, an important uncovered aspect came to fore i.e., the systematic consideration of all risk assessment attributes, interrelationship among each attribute and their interdependence. Therefore, there is a need for critical understanding in this direction. For this purpose, one of the best available tools for handling such issues is the use of graph theory approach. In the present work, risk attributes, in terms of FNA, PD, DS, POSA, EFV, RS, SEC and, DR are identified. For structural analysis, the product risk model is developed, which is a digraph.

Every digraph has primarily:

• A set of vertices ……… V^g {V^g₁, V^g₂, ……. }

• A set of edges …………E^g = {e^g₁₁, e^g₁₂, ………} and;

• A mapping function (to connect an edge to a pair of vertices). Any increase in number of attributes may lead to a complex structure. Thus, the graph is modeled in the form of a matrix. This is useful in obtaining the sigma expression for risk assessment and also the index value. The concept of graph theory has been instrumental in addressing various problems of science and technology (Chen, 1997).