Modeling Water Utility Investments and Improving Regulatory Policies using Economic Optimisation in England and Wales

Water utilities in England and Wales are regulated natural monopolies called 'water companies'. Water companies must obtain periodic regulatory approval for all investments (new supply infrastructure or demand management measures). Both water companies and their regulators use results from least economic cost capacity expansion optimisation models to develop or assess water supply investment plans. This presentation first describes the formulation of a flexible supply-demand planning capacity expansion model for water system planning. The model uses a mixed integer linear programming (MILP) formulation to choose the least-cost schedule of future supply schemes (reservoirs, desalination plants, etc.) and demand management (DM) measures (leakage reduction, water efficiency and metering options) and bulk transfers. Decisions include what schemes to implement, when to do so, how to size schemes and how much to use each scheme during each year of an n-year long planning horizon (typically 30 years). In addition to capital and operating (fixed and variable) costs, the estimated social and environmental costs of schemes are considered. Each proposed scheme is costed discretely at one or more capacities following regulatory guidelines. The model uses a node-link network structure: water demand nodes are connected to supply and demand management (DM) options (represented as nodes) or to other demand nodes (transfers). Yields from existing and proposed are estimated separately using detailed water resource system simulation models evaluated over the historical period. The model simultaneously considers multiple demand scenarios to ensure demands are met at required reliability levels; use levels of each scheme are evaluated for each demand scenario and weighted by scenario likelihood so that operating costs are accurately evaluated. Multiple interdependency relationships between schemes (pre-requisites, mutual exclusivity, start dates, etc.) can be accounted for by additional constraints. User-defined annual water saving profiles are used for DM schemes so that water conservation 'yields' can follow observed patterns. A two-stage optimization procedure is applied to deal with network infeasibilities which appear in large applications. We apply the model to a regional system of seven water companies in the South East of England, the driest part of the UK with its largest and fastest growing population. The model's spatial units are water supply zones, i.e. interconnected zones of equal supply reliability; each company contains between 3 and 8 of these. Economic benefits of greater sharing of resources among water companies (regional water transfers) are evaluated by considering bi-directional interconnections between all neighboring supply zones. Next we describe an extension of the model to investigate how current regulations incentivize companies to invest in an attempt to understand how better regulations can incentivize more water transfers. Finally, an attempt is made to change from the current model assumption of perfect cooperation between companies to one that represents the fact that each company is a private company seeking to maximize its own benefits. Limitations and advantages of the formulations are discussed and recommendations for capacity expansion modeling are made.