Congress Resources: Papers, posters and presentations

< Return to abstract list

Centralised Versus Decentralised Planning Of Desalinated Water Supply System

Congress: 2015
Author(s): Maedeh P.Shahabi, Adam McHugh, Martin Anda, Goen Ho
Murdoch University1

Keyword(s): Sub-theme 10: Management of water resources,
AbstractMaedeh P. Shahabi a 1, Adam McHugh a, Martin Anda a, Goen Ho a

a Engineering & Information Technology, Murdoch University, 90 South Street, Murdoch, Western Australia, 6150

1 Corresponding author: E-mail address:

Postal address: Engineering & Information Technology, Murdoch University, 90 South Street, Murdoch, Western Australia, 6150


A recent trend in seawater desalination is the construction of larger-capacity plants, which has significantly contributed to freshwater supply for coastal cities around the globe. These large desalination plants enjoy economies of scale for the construction phase. However they may be relatively far from where water demand is located and may suffer from greater cost due to the need to construct a long water pipeline and associated energy for delivery of treated water to location of demand. This paper argues the need to identify the optimum geographical scale of a desalination system and thus resist the default mindset of "big is better" [1] in the water sector.


Comparative Levelised Cost (LC) is demonstrated for three geographical scale desalination supply system: A centralised scenario and two decentralised scenarios. For the centralised scenario we focus on a new 100 GL per annum SWRO desalination plant which has been planned to support future urban expansion in the northern corridor of Perth, Australia and to replace a loss of capacity in the groundwater system due to reduced rainfall. Decentralised scenarios are proposed as alternative options to supply water to the same area, where several medium size plants totally also produce 100 GL per annum. We created a spatial-temporal database of desalinated water demand from 2015-2035 (realistic hypothetical), vacant lands for accommodating plants and water storage facilities, and existing pipeline infrastructure in order to design scenarios with different geographical scales. Scenarios were compared with each other using LC. All spatial related tasks were performed using ArcGIS version 10. The three scenarios were defined, designed and compared for their economical performances. These were:

Centralised scenario: business as usual (BAU) planning of a single SWRO desalination plant with a capacity of 310,000 m3/day. Construction of 75 kilometres trunk main to connect the plant to Perth urban area was accounted for in this scenario. Water demand area for the plant consisted of 28 suburbs.

Decentralised scenario: Cluster scale 1, four SWRO desalination plants with sizes ranging between 65,000-100,000 m3/day with cumulative design capacity of 310,000 m3/day. The service zone of each plant consisted of four to ten suburbs.

Decentralised scenario: Cluster scale 2, eleven SWRO desalination plants with sizes ranging between 15,000-50,000 m3/day with total design capacity of 310,000 m3/day. The service zone of each plant consisted of one to four suburbs.

Results and discussion:

LC for scenarios lie between AU$2.92 to AU$3.63 per m3 water produced. The economically optimum scenario leads to approximately 20% lower total costs. The best economical performance belongs to the decentralised scenario: cluster scale 1, while the weakest economical performance belongs to centralised scenario. In all 3 scenarios highest cost contribution belongs to construction cost. Centralised scenario has the lowest plant capital cost due to economy of scale for plants size. However, the economy of scale for the plant construction cost is outweighed for the case study because a new water trunk main must be constructed in order to connect the plant to the water demand zones. The second important contributor influencing the total cost is the variable treatment plant O&M costs (including membrane, chemical, electricity and replacement parts), which is the same for all scenarios. Third important contributor influencing the total cost is the fixed O&M (including labour and insurance) costs, which is lower at centralised scenario with large plant size. This is mostly due to lower number of staff per production in the larger centralised plant. Water delivery pumping cost increases by 3% in centralised scenario due to the higher distribution distance in comparison with decentralised scenarios however it has low contribution to the system economic performance.


The current study posits that decentralisation of SWRO desalination plants can achieve lower cost. The key contribution of the present study is raising awareness of possible environmental and economic losses associated with the business as usual approach of centralised planning of desalination for supplying urban water demand. 1. Marques, R.C. and K. De Witte, Is big better? On scale and scope economies in the Portuguese water sector. Economic Modelling, 2011. 28(3): p. 1009-1016.

2011 IWRA - International Water Resources Association - - Admin