Interdependent Critical Infrastructure Model (ICIM): An agent-based model of power and water infrastructure

Abstract

The comfort, mobility, and economic well-being of a population depends on reliable and affordable electric power services, which in turn requires a sustainable water supply. It is therefore increasingly important to analyze the sustainability and resilience of mid- and long-term electric utility and water system capacity expansion plans. Due to the inherent interdependency between power and water critical infrastructure, these expansion plans should be analyzed with respect to potential challenges posed by climate change and other risks. Decision-makers therefore require tools that facilitate an integrated analysis that captures the interdependency of power and water to better inform future expansion plans. Here we develop an agent-based model of a typical regional power system that incorporates the features of specific plant types and their cooling systems that are dependent on adequate water supplies at appropriate temperatures to support full power operation. The effects of capacity expansion plans, power demand growth, climate change, and extreme events are analyzed through different scenarios designed to illustrate the utility of such a model and show where it can aid in mid- and long-term planning.

Introduction

Electric power and water provide vital services to society. As such they are designated by the U.S. government as critical infrastructure lifeline sectors. Nearly all the remaining U.S. critical infrastructure sectors depend on power and water to produce their goods and services [1]. Both electric power and water and wastewater systems draw upon natural water resources to enable the provision of their lifeline services. This common source—natural water resources—creates an interdependency between power and water critical infrastructure systems. When natural water resources become scarce due to drought, disasters, or mismanagement, electric power and water and wastewater systems may compete for use of that scarce resource. In a worst-case scenario, the scarcity of natural water resources could result in electric power and water shortages that adversely affect the health and economic wellbeing of a region.

The challenge of interdependency planning is deepened by the fact that different regions rely on different combinations of fuel for energy production, have different supply and demand profiles, have different amounts of water availability, and have different current and projected climates [2]. To reduce the possibility of future power and water shortages, regional and local power and water authorities should account for their mutual dependence on the same regional natural water resource supply. However, due to organizational barriers, differences in planning cycles, differences in geographic jurisdiction, and lack of joint planning tools, power and water authorities often do not engage in such joint planning.

All of this suggests a growing need for models and simulations that allow decision-makers to gain insights into the interactions and feedback structures of power and water on a local, regional, national, and global scale. Agent-based modeling (ABM) is a paradigm for analysis that is ideal for capturing complex effects and heterogeneity exhibited by agents in a system. In this paper we present the agent-based Interdependent Critical Infrastructure Model (ICIM), which is intended to facilitate joint long-term planning and thereby help authorities to avoid future power and water shortfalls.

This effort seeks to extend infrastructure models to a regional scale in a manner that can be employed by both national and regional long-term planners. By applying ICIM to a specific, constrained geographic region, we can identify potential vulnerabilities and failure points in power-water interdependencies. The characteristics of individual power sources are captured in the agent descriptions. Similarly, the dispatch algorithms used by the Electric Reliability Council of Texas (ERCOT) are approximated and encapsulated, as are the decision processes for determining water restrictions and usage outlined by the Lower Colorado River Authority (LCRA).

A key differentiator of ICIM is its focus on long-term planning. Existing, functional power-water planning models often focus on short-term, event based interdependencies (e.g., [3], [4], [5]). ICIM provides a capability that incorporates multiple time scales, starting at a granular 15-minute power demand level, and building up to a multi-year time frame that can support the analysis of the impact of extended periods of drought. By integrating the different decision-making time scales of power and water, ICIM can perform analysis and evaluation across several areas identified by the Department of Homeland Security as lifeline critical infrastructures including electric power, water, nuclear systems, dams, agriculture, and other critical elements and services in a regional economy (e.g., [6] and [7]).

In the sections that follow, we introduce ICIM as a tool template that can facilitate joint long-term planning for critical infrastructures. In Section 2 we review some of the federal and local policies that impact power and water management and discuss the modeling literature pertinent to the ICIM design. Section 3 presents the technical details of ICIM, including the construction of input variables. Section 4 describes the experiments designed to illustrate the utility of ICIM. The results of each experiment are presented in subsections immediately following each scenario description. Note that Thompson et al. [8] presented a preliminary abridged version of the model described in Section 3 and some of the results in Section 4. We conclude in Section 5 with a summary and discussion of future work.