Life cycle sustainability assessment (LCSA) for selection of sewer pipe materials

Abstract

Sewer systems, over their life cycle, suffer deterioration due to aging, aggressive environmental factors, increased demand, inadequate design, third party intervention, and improper operation and maintenance activities. As a result, their state and overall long-term performance can be affected, which often requires costly and extensive maintenance, repair, and rehabilitation. Furthermore, these pressures can enhance the risk of failures (e.g., sewer leakage) which in turn can have serious impacts on the environment, public safety and health, economics, and the remaining service life of these assets. Effective asset management plans must be implemented to address long-term sustainability principles, i.e., economic growth, human health and safety, and environmental protection, simultaneously. The aim of this paper is to evaluate and compare four typical sewer pipe materials [i.e., concrete, polyvinyl chloride (PVC), vitrified clay, and ductile iron] and identify sustainable solutions. Two comprehensive life cycle sustainability assessment (LCSA) frameworks were applied. The first LCSA framework was based on the integration of emergy synthesis, life cycle assessment (LCA), and life cycle costing (LCC). In this framework, emergy synthesis has been applied to integrate the results from environmental analysis (i.e., LCA) and economic analysis (i.e., LCC) to an equivalent form of solar energy: a solar emergy joule. The second LCSA framework was based on a conventional, multi-criteria decision-making technique, i.e., the analytical hierarchy process, to integrate the results from environmental analysis (i.e., LCA) and economic analysis (i.e., LCC) and find the most sustainable solution over the sewer pipe life cycle. The results demonstrate that PVC pipe is the most sustainable option from both environmental and economic view points and can ensure a more sustainable sewer system.

Appendix

Emergy synthesis

Emergy synthesis is one of the environmental accounting methods that take into account the contribution of ecological products and services. Emergy can be defined as the available solar energy used up directly or indirectly to create a service or product and can be used to assess natural inflows and services within a system (Odum 1996). For instance, many joules of sunlight are required to make 1 J of fuel, several joules of fuel are needed to make a joule of electricity, many joules of electricity are required to support information processing in a university, and so forth. Because different kinds of energy are not equal in contribution, work is made comparable by expressing each in units of one form of energy previously required (Odum and Brown 2000). According to Odum (1996), in order to account for the existence of energies of different qualities, they must be considered in terms of one type of energy. The type of energy chosen as reference was solar energy, because it is basically the source of almost all flows in the biosphere.

The emergy of different products is calculated by multiplying mass (g) or energy quantities (J) by transformity, which is a transformation coefficient. Transformity is one example of a UEV and is defined as the emergy per unit energy. In the literature, emergy values and transformities are reported in scientific form (e.g., 3.42E+12 seJ/kg). For ease of use, emergy values can be reported using metric prefix of ‘tera’ (10¹²). Transformity is an intensive quantity and is measured in seJ/J (emergy per unit energy). It represents the inverse of an efficiency comparing two similar processes; a higher transformity means that more emergy is need to produce the same amount of output. Therefore, the transformity is a measure of hierarchical position in energy transformation chains (Zhang et al. 2006).The emergy of different product is calculated by multiplying mass (g) or energy quantities (J) by transformity, which is a transformation coefficient. The solar emergy U of a flow coming from a given process is

$$U = \sum\limits_{i} {{\text{UEV}}_{i} \times E_{i} } ,$$

(6)

where E _i is the actual energy content of the ith independent input flow to the process and UEV _i is the unit emergy value or solar transformity of the ith input flow (Pulselli et al. 2007).

It is common to measure solar transformity in seJ per joule of product (seJ/J) with a base that 1 emjoule is equivalent to 1 J of solar energy and transformity of solar energy is 1 seJ/J (Ulgiati et al. 1995). The solar transformity of the sunlight absorbed by the earth is 1.0 by definition. Solar transformities represent the position of any product or service in the hierarchical network of the earth’s biosphere (Odum 1996). For instance, if 6,000 solar emjoules are required to generate 30 J of natural gasoline, then the solar transformity of that gasoline is 200 seJ/J (6,000/30 seJ/J). Solar energy is the largest but most dispersed energy input to the earth. The higher the transformity of an item, the more available energy of another kind is required to make it (Brown and Ulgiati 2004). For convenience, it is very common to use transformity values derived from other studies. The use of emergy method is easy and its goal is to support designer’s decisions in the development/assessment of more sustainable products or process. Emergy method normalizes all the attributes of the system in a common metric unit, called solar emergy (Tilley and Swank 2003).

Multi-criteria decision-making (MCDM)

In the last three decades, MCDM research in different disciplines has expanded extensively. Hwang and Yoon (1981) reviewed and summarized MCDM methods and applications. In a MCDM process, a decision maker is required to choose among quantifiable or non-quantifiable and multiple criteria (Pohekar and Ramachandran 2004). Several MCDM methods have been applied in life cycle approach, like ELECTRE (Roy 1991), PROMETHEE (Brans et al. 1984) and GAIA (Brans and Mareschal 1994), AHP (Saaty 1980), and TOPSIS (Yoon and Hwang 1985).

The MCDM methods are capable of performing the solution procedure regardless of the functional relationship for the objectives and constraints, and secondly, the number of attributes and alternatives applicable to the model is computationally limitless. However, the MCDM methods have two weak points. First, the MCDM methods are lacking in the delivery of the absolute optimum; however, they are capable of deciding over the best options among selected alternatives. Second, if the weights of the criteria are not properly assigned, it may fail to reveal “true” decisions. AHP that was used in this study has been successfully implemented in various engineering applications, especially for infrastructure management.

AHP is a systemic method commonly used for decision-making (Sadiq et al. 2004; Saaty 1997). AHP can solve complex decision-making problems involving few alternatives with numerous criteria. The process of comparing the relative importance or preference of a parameter (objectives or criteria) with respect to other parameter is based on pair-wise comparisons. One of the major advantages of AHP is using pair-wise comparisons to determine weights and derive priority index in comparison to other weighting methods where weights are assigned arbitrarily. AHP can use subjective assessment of relative weights (importance, likelihood, or preference) to a set of priority ratio scale and overall scores (Sadiq et al. 2003).

Usually a hierarchical model is developed to degenerate complex problems into simpler and manageable elements which create different hierarchical layers or levels. The first level of each hierarchy is a goal or an objective, whereas at the last level there is an evaluation of alternatives. The intermediate layers contain criteria and sub-criteria (Tesfamariam and Sadiq 2006).

Saaty (1980) proposed pair-wise comparisons at each level in the hierarchy using a reciprocal matrix. The pair-wise judgment matrix thus developed, indicates dominance or relative importance of one element over another (Saaty 1980). The result of the pair-wise comparison on n criteria is summarized in an n × n matrix as follows:

$$A = \left[ {\begin{array}{cccc} {a_{11} } & {a_{12} } & \ldots & {a_{1n} } \\ {a_{21} } & {a{}_{22}} & \ldots & {a_{2n} } \\ \vdots & \vdots & \ldots & \vdots \\ {a_{n1} } & {a_{n2} } & \ldots & {a_{nn} } \\ \end{array} } \right],\quad a_{ii} = 1,\quad a_{ji} = 1/a_{ij} ,\quad a_{ij} \ne 0.$$

(7)

Each element a _mn in the upper triangular matrix expresses the importance intensity of a criterion (or property) m with respect to another criterion n. The weights of the criteria in each level of the hierarchy are determined by taking the geometric mean of each column of the final judgment matrix and then normalizing the derived matrix. Finally, the weights at the lowest level will be obtained by multiplying the weights of the corresponding criteria in higher levels from the highest level to that level. In a case of n criteria, a set of weights in each level of hierarchy could be written as follows:

$$W = \left( {w_{1} ,\;w_{2} , \ldots ,w_{n} } \right)\quad {\text{where}}\quad \sum\limits_{1}^{n} {w_{n} } = 1.$$

(8)

There are several mathematical techniques that can be used to calculate the vector of priorities (weights) from matrix, such as eigenvector, geometric mean, and arithmetic mean. Preliminary investigation has been shown that there is no significant difference based on the selection of a specific technique. Normalization based on geometric means of the rows has been recommended because it provides an easy approach to obtain approximate priorities (weights; Saaty 1990). In this method, the normalization is required for each column of the matrix and then averaging over each row. One of the common issues in generating pair-wise comparison matrix is non-consistency; that is ∀ i, j: a _ij  ≠ w _i /w _j . To ensure consistency in the pair-wise comparisons and associated weight estimation, a consistency value is recommended. In pair-wise comparison matrices, the eigenvalue λ and eigenvector W (priority vector) value may help solving eigenvalue Eq. (3).

$$(A - \lambda )W = 0.$$

(9)

In Eq. (3), W is the priority vector which is associated with the matrix of comparisons and n is the dimension of this matrix. Saaty (1980) recommended a maximum eigenvalue λ _max > n for inconsistent matrices. If consistency index (CI) is sufficiently small, the estimate of the weight w is acceptable. The CI is defined as following:

$${\text{CI}} = \left( {\lambda_{\hbox{max} } - n} \right)/(n - 1),$$

(10)

where CI is consistency index that indicates whether a decision maker assigns consistent values (comparison) in a set of evaluation (Tesfamariam and Sadiq 2006). The final inconsistency in pair-wise comparison is computed using consistency ratio (CR).

$${\text{CR}} = {\text{CI}}/{\text{RI}},$$

(11)

where RI is the random index, determined by averaging CI of a randomly generated reciprocal matrix (Saaty 1980).

It is noted that making a comparison between different criteria is a challenging task. There is no widely agreed method to determine the relative importance of different impacts. Decision-making based on AHP technique can cause confusion and does not deal effectively with redundancy of selected criteria. Normalization of different attributes may fail to find the true solution of the alternatives. For this reason, more advanced method needs to be developed that can address the dilemma of non-commensurate units in MCDM problems.