Designing ecologically-inspired robustness into a water distribution network

https://doi.org/10.1016/j.jclepro.2020.120057Get rights and content

Highlights

  • Mimicking ecological food webs creates sustainable water network.

  • Optimizing ecological robustness uniquely balances cost and resiliency.

  • Bio-inspired water network are cost effective under disturbances.

  • Ecological robustness also increases value cycling in distribution networks.

Abstract

Eco-Industrial Parks (EIPs), network of industries that collaborate by utilizing each other’s byproducts and wastes, are highly desirable for both the industries themselves, their environment, and governments due to their economic, environmental, and social advantages. Previous work has shown that EIPs are not as successful as they could be in terms of mimicking the behavior of biological ecosystems, highlighting that more work needs to be done for EIPs to truly mimic their biological-counterparts. The Kalundborg EIP, located in Kalundborg, Denmark, is a well documented example of an EIP with long-term success. Using the water network within the Kalundborg EIP as a case study, two bio-inspired networks are selected from an optimization based on the ecosystem metric robustness. The bio-inspired solutions are compared with a traditionally cost-minimized solution to understand what bio-inspired design can offer when a network is disturbed. Disturbances such as connection breakages and industry shutdowns are tested, showing that the bio-inspired designs require minimal recovery costs – in stark contrast to the traditional network solution. The results show that the bio-inspired designs reduce the network’s dependence on a scarce import (freshwater) and have higher overall network resilience in the event of disturbances. The three network solutions are discussed from a ecological perspective, explaining differences from the standpoint of ecosystem characteristics. The analysis highlights the benefits of using ecology to understand the nature of and improve the design of industrial networks.

Introduction

Water distribution networks (WDNs) are one of the critical infrastructure networks that society depends upon. When these networks experience disturbances, their complex nature can quickly escalate the impact, resulting in large-scale chain reactions and secondary failures (Shuang et al., 2017). This dependence and complexity makes these networks sensitive to perturbations caused by natural and man-made disturbances, supporting the need for new ways to improve the reliability and robustness of their designs (Shuang et al., 2017; Yazdani and Jeffrey, 2011).

Robustness and reliability lack wide-spread agreed upon definitions across disciplines. Reliability is defined here as the probability that user demand will be satisfied despite various failures occurring at any given time (Shuang et al., 2017). Disciplines such as supply chain network design however do make reasonably clear distinctions between the concepts of robustness and resilience (Behzadi et al., 2018). Robustness is defined here as the ability of a system to withstand disruption and failures (Behzadi et al., 2018), avoiding major performance disruptions, often via redundant, alternative pathways. Redundant paths usually consist of nodes and standby links and/or components that are not normally used at their full capacity (Behzadi et al., 2018). Topological redundancy in WDNs, redundancy provided by closed-loops of complex and meshed structure, enables quick recovery from failures by providing alternate flow paths between any given start and end point (Giudicianni et al., 2018). Historically, WDNs have been built using topological redundancy approach but without a systematic procedure related to the resultant impact on reliability (Di Nardo et al., 2017). Conventionally, hydraulic robustness methods like reliability/chance constrained formulation (Kapelan et al., 2005), variation in stochastic pressure (Jung et al., 2014) and structural robustness methods like average node-degree (Newman, 2011), average path length (Costa et al., 2007), spectral gap (Estrada, 2006), spectral radius (Bonacich, 1987) are used to design robust WDN. The most invulnerable-to-failure water distribution systems have been shown to be ‘regular’ graphs, or networks made up of an equal number of links incident to every component (Jacobs and Goulter, 1988, 1989; Yazdani et al., 2011). These redundant links in real life however are physical pipelines that are expensive to install and maintain, becoming an undesirable solution in most efficiency-focused industrial settings. Truly sustainable designs for WDNs and EIPs need a quantitatively supported understanding of the pros and cons of efficiency and redundancy goals, possibly finding a balance between the two that also considers cost. The urgency of this need is increased as critical infrastructure networks like WDNs are pre-requisites for health, prosperity and security (Dong et al., 2018).

Nature has proven to be an excellent source of design ideas for improving human products. Leonardo da Vinci’s flying machine was inspired by birds, while the modern Japanese bullet train was modeled after a Kingfisher’s beak (Hwang et al., 2015). Bio-inspiration on a systems–level has been less common but shows no less promise, with the characteristic structure and functioning of ecosystems being translated through the quantitative metrics of ecological network analysis (ENA). The metrics within ENA have been used to study carbon emissions in Beijing, China (Chen et al., 2015) and Vienna, Austria (Chen and Chen, 2012), the economic activity of six resource trade network around the world (Kharrazi et al., 2013), the reduction of wasted water in three Italian cities (Bodini et al., 2012), and the metabolism of urban wastewater in China (Zheng et al., 2019). Wetlands were added to a steel manufacturing facility in China to mimic the detritivores (low level species who break down dead organic matter) found in ecosystems, finding that the bio-inspired design was both cost effective and water efficient (Zhang et al., 2018). Some studies have shown that those human networks that more closely mimic the characteristics of ecological food webs also lower their costs and emissions (Layton et al., 2016a; Reap and Bras, 2014; Reap, 2009). These results suggest that using ENA metrics can help designers achieve a variety of traditional objectives via unconventional means. Ecologists have also quantitatively studied ENA metrics that characterize ecosystem robustness and cycling. These metrics are the focus of this paper, with the objective of understanding potential resilience benefits that may come from mimicking ecosystems. Ecologists have devoted significant time to studying the unique balance between efficiency and redundancy found in ecosystems (Ulanowicz, 2009). This paper models intentional attacks to pipelines on a traditionally designed WDN and two biologically-inspired WDNs to evaluate the resulting severity of cascading effects and the overall resilience.

Eco-industrial Parks (EIPs), a meso–level implementation of circular economy (Farooque et al., 2019), are networks of industries formed through symbiotic by-product exchanges. EIPs are reminiscent of the predator-prey food web representation of biological ecosystems (Ehrenfeld and Gertler, 1997; Layton et al., 2016a, 2016b). The exchanges that make up the structure of EIPs help producer industries avoid by-product disposal/cleaning costs and consumer industries avoid raw materials with lower cost alternatives (Lowe, 2001). These by-product exchanges have shown to result in emissions and water usage reductions as well and socially attracting both leading-edge corporations and smaller local ventures (Lowe, 2001). The mutually beneficial exchanges between industries are the drivers of the ‘closed-loop’ structural goal of EIPs, providing economic, environmental, and social advantages (Ehrenfeld and Gertler, 1997; Hein et al., 2015; Layton et al., 2016b).

A seminal example of a successful EIP can be found in Kalundborg, Denmark. The EIP known as “Kalundborg industrial symbiosis” (Ehrenfeld and Gertler, 1997) (shown in Fig. 1) exchanges waste, by-products, water, and energy based on self-created contractual dependencies (Jacobsen, 2006). Four major companies form the basic structure of the EIP: the 1300 MW (in 2002) Asnaes power plant, a biotech and pharmaceutical company - Novo group, an oil refinery - Statoil, and a plasterboard manufacturer – Gyproc (Ehrenfeld and Gertler, 1997; Jacobsen, 2006). The Kalundborg EIP initially formed in response to a large groundwater deficit in the region, caused by recent expansions of industries in the area (Jacobsen, 2006). The first EIP interaction was in 1961, when surface water from Lake Tisso replaced groundwater use for the industries (Ehrenfeld and Gertler, 1997; Jacobsen, 2006). This initial change became a springboard for other exchanges: Statoil started supplying wastewater and cooling water (1987 & 1991) to the power plant (Branson, 2016; Jacobsen, 2006) to use as boiler feeder water (Branson, 2016; Ehrenfeld and Gertler, 1997; Jacobsen, 2006). The power plant also started using sea water as cooling water to further reduce the stress on freshwater resources (Jacobsen, 2006). Since 1981, the town of Kalundborg has eliminated 3500 oil-fired residential furnaces by distributing the heat from the power plant through an underground pipe system (Ehrenfeld and Gertler, 1997). The Kalundborg EIP reported in 2002 an annual savings of 1.9 million m3 of groundwater, 1 million m3 of surface water, 20,000 tons of oil, 30,000 tons of coal, and a reduction in overall CO2 and SO2 emissions (130,000 and 3700 tons, respectively) (Ehrenfeld and Gertler, 1997; Jacobsen, 2006). The evolution of these symbiotic interactions making up Kalundborg was a complex process and happened over the course of more than 50 years, with each link developed as an independent business deal (Ehrenfeld and Gertler, 1997).

Despite the many documented benefits, a concern of companies looking to join EIPs is that the dependence on the by-products can reduce their resilience, make them vulnerable in a critical market or if a producing industry were to go out of business (Lowe, 2001). Quantifying the benefits of bio-inspired design for network resilience will enable industries to make more informed decisions joining an EIP. This will increase the appeal of these industrial communities and further their environmental, economic, and social benefits.

Designing these distribution networks is a complex task that requires a systems-of-systems perspective (Lowe, 2001). Previous studies (Branson, 2016; Ehrenfeld and Gertler, 1997; Jacobsen, 2006) of the Kalundborg EIP have sought to understand the decision-making parameters, benefits, and risks associated with the symbiotic exchanges to be able to mimic them in the design of ground-up EIPs. The structural similarity of the Kalundborg EIP to ecological food webs has been previously studied using basic, low fidelity ENA metrics (Layton et al., 2016b). The bio-inspired design of Kalundborg EIP in this paper focuses on mimicking the flow characteristics of ecological food webs that influence their robustness. Traditional investigations of water distribution network designs have focused on anything from a single plant to multiple industries, with the aim of minimizing things such as total water usage costs, total freshwater usage, capital investment costs, interconnections, and water treatment costs (Jezowski, 2010). Models following a variety of techniques have been used, including linear and nonlinear programming and heuristics methods. Previous studies on the water networks of EIPs have focused on understanding the effects of collaborative and non-collaborative techniques, achievement of overall cost reductions, and the minimization of freshwater consumption (Chew and Foo, 2009; Chew et al., 2009; Lovelady and El-Halwagi, 2009). This paper uses these traditional optimization goals to test a bio-inspired optimization method applied to a case study model based on the water distribution network of the Kalundborg EIP. The resultant bio-inspired approach is a novel method that uses ecological metrics mimic ecosystem characteristics in the design a WDN.

Section snippets

Ecological network analysis

Food webs are aggregated representations of ecosystems that focus on their prey to predator interactions (Layton et al., 2016b). Ecological network analysis (ENA) is an analysis technique used by ecologists to quantify the characteristics surrounding the structure and behavior of food webs using a wide variety of metrics. The metrics fall into two categories, those that require only structural information or those that use both structural and flow magnitude information. Ecologists use

Traditional properties of interest

Table 3 lists the final costs (US$/yr), total freshwater water consumption (m3/yr), and an ENA analysis of the solutions to the three (traditional, BI1, and BI2) optimization scenarios. All three solutions use the local groundwater and surface water to fulfill the needs of any industry that is not fulfilled by other sources.

The traditional solution uses a total of 3,692,000 m3/yr of freshwater and has a total cost of $12,902,800. Fig. 7a and the flow matrix [T] in Table A.1 in the appendix

Solution analysis

WDNs are particularly sensitive to perturbations caused by natural and man-made disasters (Shuang et al., 2017). Damage to key components of the WDN can have cascading failures and lead to network breakages (Shuang et al., 2017). All the solutions were evaluated for their robustness by simulating network attacks, modeled as pipeline breakages and industry shutdown scenarios. These disturbance simulations are set up to answer if an interconnected network can remain functional, maintaining

Conclusion

The ecological metric robustness presents a novel and potentially superior method for improving the design of industrial distribution networks, especially to protect these networks against outages and disturbances. The using a basic water network model, based on the real Kalundborg EIP, as a case study highlights the improvements afforded by the bio-inspired design approach. The validation of these results, checked using multiple variations on linkage and actor shutdowns, support this

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (62)

  • B.D. Fath

    Quantifying economic and ecological sustainability

    Ocean Coast Manag.

    (2015)
  • B.D. Fath et al.

    Cyclic energy pathways in ecological food webs

    Ecol. Model.

    (2007)
  • J.T. Finn

    Measures of ecosystem structure and function derived from analysis of flows

    J. Theor. Biol.

    (1976)
  • A. Kharrazi et al.

    Quantifying the sustainability of economic resource networks: an ecological information-based approach

    Ecol. Econ.

    (2013)
  • R.E. Ulanowicz

    The dual nature of ecosystem dynamics

    Ecol. Model.

    (2009)
  • A. Yazdani et al.

    Resilience enhancing expansion strategies for water distribution systems: a network theory approach

    Environ. Model. Softw

    (2011)
  • B. Zheng et al.

    Metabolism of urban wastewater: ecological network analysis for guangdong province, China

    J. Clean. Prod.

    (2019)
  • Water system and disasters

  • L. Astrid et al.

    Industrial ecosystems and food webs: an expansion and update of existing data for eco-industrial parks and understanding the ecological food webs they wish to mimic

    J. Ind. Ecol.

    (2016)
  • J. Bascompte et al.

    Plant-animal mutualistic networks: the architecture of biodiversity

    Annu. Rev. Ecol. Evol. Syst.

    (2007)
  • P. Bonacich

    Power and centrality: a family of measures

    Am. J. Sociol.

    (1987)
  • S.R. Borrett et al.

    enar: an r package for ecosystem network analysis

    Methods Ecol. Evol.

    (2014)
  • F. Briand

    Environmental control of food web structure

    Ecology

    (1983)
  • S. Chen et al.

    Network environ perspective for urban metabolism and carbon emissions: a case study of vienna, Austria

    Environ. Sci. Technol.

    (2012)
  • S. Chen et al.

    Nonzero-sum relationships in mitigating urban carbon emissions: a dynamic network simulation

    Environ. Sci. Technol.

    (2015)
  • L.d.F. Costa et al.

    Characterization of complex networks: a survey of measurements

    Adv. Phys.

    (2007)
  • Costwater

    Running costs of wastewater treatment plants

  • T. Dave et al.

    Bio-inspired Design for Resilient Water Distribution Networks

    (2019)
  • T. Dave et al.

    Bio-inspired design for resilient water distribution networks

  • J. Ehrenfeld et al.

    Industrial ecology in practice: the evolution of interdependence at kalundborg

    J. Ind. Ecol.

    (1997)
  • E. Estrada

    Network robustness to targeted attacks. the interplay of expansibility and degree distribution

    Eur. Phys. J. B Condens. Matter

    (2006)
  • Cited by (0)

    View full text