Joint Expansion Planning of Power and Water Distribution Networks

Sai Krishna Kanth Hari, Ahmed S. Zamzam, Byron Tasseff, Russell Bent, Clayton BarrowsarXiv 2023 · DOI · arXiv

Introduction

Modern urban infrastructure relies on the deep interdependence of power and water distribution networks. Traditionally, these systems are planned and operated in silos. However, water pumps are significant electricity consumers, and power generation often requires substantial water for cooling. This paper presents a framework for the joint expansion planning of these networks, modeling them as a single, interdependent system to optimize resource allocation and improve resilience.

Methodology: The Joint MINLP Framework

The researchers formulate the expansion problem as a Mixed Integer Nonlinear Program (MINLP). This approach captures both discrete investment decisions (e.g., whether to install a new pipe or battery) and the nonlinear physical laws governing each network.

Network Modeling

The power network is modeled using high-fidelity AC power flow equations to account for voltage magnitudes and thermal limits. On the water side, the Hazen-Williams equation is employed to model nonlinear head loss in pipes. The critical "link" between the two systems is the water pump, whose power consumption is modeled as a function of the water flow rate, effectively treating it as a controllable load on the power grid.

Schematic of the interdependent power and water distribution test network used for model validation.

The study utilizes a synthetic test network (Figure 1) comprising a 3-bus power distribution system and a 3-node water network. Expansion options include PV units, battery energy storage, water pumps, and storage tanks.

Key Insights: Cross-Network Synergies

The joint planning approach reveals significant efficiencies that are missed in independent planning.

Water Tanks as Virtual Batteries

A primary finding is that water tanks can serve as "virtual batteries" for the power grid. By installing a tank, the system gains the flexibility to pump water during off-peak hours and store it for later use. This shifts the electrical load of the pump away from peak periods, freeing up capacity for other power-only loads without requiring expensive electrical storage.

Budget Sensitivity and Substitution

The optimal investment strategy is highly sensitive to the available budget. Joint planning allows for "cross-network substitution," where a component in one network helps satisfy demand in the other. For instance, at certain budget levels, adding a water tank is the most cost-effective way to increase the satisfaction of peak power demand.

Peak hour power load satisfaction vs. budget, demonstrating how different infrastructure components contribute to system capacity.

As shown in Figure 2, at a budget of 0.4, the addition of a water tank (a water-side component) significantly increases the peak hour power-only load satisfied from 1.0 MW to 1.15 MW. This illustrates the power of leveraging interdependencies.

Results and Conclusion

Numerical experiments demonstrate that joint planning reduces infrastructure redundancy. In many scenarios, a single investment (like a tank or a battery) can satisfy the elevated demands of both networks simultaneously. Specifically, the "virtual battery" effect of water tanks allowed for a 15% increase in satisfied peak power demand in the test case.

While the current MINLP formulation faces scalability challenges due to its computational complexity (NP-hard), this research provides a proof-of-concept for more integrated urban planning. Future work will focus on developing specialized algorithms to solve these problems for large-scale city networks and exploring broader interdependencies, such as water-for-cooling in thermal power plants. By breaking down the silos between utility sectors, cities can achieve greater resilience and cost-efficiency in their infrastructure expansion.

Summary updated

BibTeX
@misc{hari2023joint,
  author = {Sai Krishna Kanth Hari and Ahmed S. Zamzam and Byron Tasseff and Russell Bent and Clayton Barrows},
  title = {{Joint Expansion Planning of Power and Water Distribution Networks}},
  year = {2023},
  eprint = {2312.10224},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2312.10224},
}
Figure