Elsevier

Applied Energy

Volume 269, 1 July 2020, 115071
Applied Energy

The economic value of a centralized approach to distributed resource investment and operation

https://doi.org/10.1016/j.apenergy.2020.115071Get rights and content

Highlights

  • First whole-system comparison of centralized vs decentralized electricity planning.

  • Coordinated planning could save between 7% and 37% of system costs over 15 years.

  • Decentralized decision-making leads to suboptimal adoption of rooftop solar.

  • Inefficiencies solved by setting rates close to the hourly locational marginal cost.

Abstract

Distributed energy resources have been almost exclusively deployed and operated under a decentralized decision-making process. In this paper, we assess the evolution of a power system with centrally planned utility-scale generation, transmission, distribution, and distributed resources. We adapt a capacity expansion model to represent both centralized and decentralized decision-making paradigms under various electricity rate structures. This paper shows that a centralized planning approach could save 7% to 37% of total system costs over a 15-year time horizon using a Western United States utility as a case study. We show that centralized decision-making deploys substantially more utility-scale solar and distributed storage compared to a decentralized decision-making paradigm. We demonstrate how a utility could largely overcome the complications of decentralized distributed resource decision-making by incentivizing regulators to develop electricity rates that more closely reflect time- and location-specific, long-run marginal costs. The results from this analysis yield insights that are useful for long-term utility planning and electric utility rate design.

Introduction

Decisions about the deployment and operation of distributed energy resources (DER) have almost exclusively been made “behind the meter” and based on the cost of equipment, financing and subsidies available, and electricity rates. Utility planners have generally not been involved in determining where and how much distributed generation should be installed. It follows that the decisions to invest in distributed resources have not stemmed from any coordinated planning effort, and the decisions to operate these resources generally do not respond to any coordinated dispatch process. In this paper, we investigate the economic value of coordinating the planning, procurement, and operation of DER.

To date, distribution utilities have grappled with DER adoption by using the feeder hosting capacity to cap the deployment of distributed generation in locations where system reliability may be adversely affected. Reliability concerns related to DER may include voltage and frequency control, reactive power management, less predictable power output, and bi-directional power flows that increase operational complexity [1]. Utilities occasionally need to upgrade or repower distribution systems and circuits to accommodate DER [2]. DER also benefit distribution systems by reducing losses, deferring system capacity expansion, alleviating feeder loading, and, in some cases, improving reliability for end-use customers [3]. Unfortunately, it is difficult to quantify these benefits because DER investment and dispatch are not coordinated with utility-level investment and dispatch. It should be noted that load serving entities (LSEs) that make bulk power system (BPS) investment and/or operation decisions have started to track these behind-the-meter resources more closely for their short- and long-term planning [4]. Furthermore, there are a small but growing number of utilities that are considering installing and operating owned behind-the-meter DER, which suggests this study could have immediate real-world applications [5].

The lack of control over any DER deployment and the traditional separation of distribution and BPS planning means that LSEs have limited opportunities to manage uncertainty associated with DER. Consequently, LSEs need to forecast adoption and operation of DER in an effort to characterize the net demand they will face in the future. It follows that decentralized additions of DER into the distribution network are most likely misaligned with the least-cost expansion of the BPS. This misalignment is exacerbated in cases where DER adoption forecasts are inaccurate.

DER adoption has increased substantially over the last 15 years driven by technology cost declines, favorable rate structures including net-metering and increased block pricing, and public awareness [6]. The residential photovoltaic (PV) market has grown 44% annually since 2005 and about 2.5% of households in the U.S. have installed a PV system [7]. This aggregate figure masks the fact that distributed PV is much more prevalent in certain states like Hawaii, California, and Arizona where 31%, 11%, and 9% of households have PV as of 2018, respectively. PV is the most prevalent residential distributed generation technology, but there could be substantial adoption of batteries as a distributed resource in the near future due to lower costs and increases in efficiency/capacity [8]. DER adoption is expected to continue to grow rapidly [9], which will only increase the inefficiencies associated with the lack of coordination between bulk power and distribution systems.

In this paper, we evaluate an important, yet understudied topic: the economic value of fully integrating DER into a long-term utility planning process. To explore this topic, we follow a typical utility planning process by (1) forecasting demand; (2) accounting for existing and future resources (utility-scale and DER); and (3) running a capacity expansion model to determine the least-cost pathway to meet future demand. We use a tool that evaluates least-cost expansion and operation of generation, transmission, distribution, and DER systems: the Grid Access and Planning (GAP) model [10]. The GAP model was originally programmed as a social planner with a centralized decision-making paradigm. We adapt GAP by splitting the customer (i.e. behind-the-meter) and utility decisions to simulate the decentralized decision-making approach that characterizes the industry. The centralized version of the tool provides the counterfactual to assess the differences with a decentralized decision-making framework associated with DER deployment and operation. Therefore, in this paper, the decentralized scenarios are the “business as usual” or status quo cases, whereas the centralized scenario is the simulated counterfactual against which the decentralized scenarios are compared.

Before proceeding, it should be noted that at present it might not be politically or technically feasible to centralize the decisions of deployment and operation of distributed resources. However, this analysis makes an important contribution to the literature by quantifying the economic impact of decentralized investment and operation as they already occur in practice—a topic that has not been evaluated before. More specifically, we measure the differences in technology types, sizes, locations, and operational schemes between centralized and decentralized investment and operation of the power system over its full scope, from distribution to transmission and generation

Based on this assessment, for example, electricity rates could be designed to achieve specific deployment or operation objectives that are a proxy for potentially more efficient decisions made by a centralized operator. Results from this paper have important policymaking implications as they highlight how long-term planning processes could be improved by integrating bulk power and distribution system planning [11]. This study answers the following specific research questions:

  • o

    What are the cost differences between decentralized versus centralized investment in and operation of DER?

  • o

    What are the technological and operational differences between a system managed with a decentralized decision-making structure versus a centralized one?

  • o

    What impacts do rates have on DER adoption and how does decentralized DER adoption compare to the centralized decision-making outcome under different rate structures?

  • o

    What changes could be implemented in the utility planning processes to address the challenges of decentralized decision-making?

This manuscript is organized as follows. Section 2 contains a short review of relevant literature. The GAP model, its adaptation to a decentralized framework, and its calibration are explained in Section 3. Section 4 includes a case study and Section 5 describes key results. Section 6 concludes with several findings that are relevant to utility IRP processes. All figures and tables are our own unless noted.

Section snippets

Relevant literature

Increased adoption of DER in the past decade is driving research into its impacts on distribution, transmission, and generation, and how these impacts are mediated by alternative decision-making paradigms. This paper addresses the second topic by comparing the effect that centralized and decentralized decision-making paradigms have on power system expansion with DER. This section then provides an overview of the literature on the impacts of DER adoption on power systems, with a focus on the

Methods

In this section, we introduce the GAP model employed to assess the differences between centralized and decentralized decision-making paradigms. The original GAP model implements a centralized decision-making paradigm for investment and operation of the power system. For this paper, the model is modified to emulate a decentralized decision-making paradigm by splitting the customer (behind the meter) decisions from the utility expansion decisions. These two model versions are then used to

Case study and input data creation

The GAP model has been employed previously to study conceptual systems. For this study, however, we deemed it useful to calibrate the model with input data that resemble an existing LSE’s generation, transmission, distribution, and customer-sited DER assets. We use an actual utility from the Western U.S.4

Results analysis and discussion

This results section is organized in the following manner. First, we report how centralized and decentralized decision-making outcomes differ in technology and location choices as well as operational decisions. Second, we detail the differences in costs between scenarios. Third, we show the impact of different rate structures on DER adoption for both decision-making paradigms. Finally, given cost-effective DER, we introduce the concept of price elasticity of net load as a metric that could

Conclusion

Electric utilities have traditionally relied on least-cost capacity expansion models with a single decision maker to support their resource planning processes and identify the most cost-effective generation and transmission investments to meet expected load. However, this approach is not well suited to incorporate behind-the-meter distributed resources, which in reality are installed and operated by customers attempting to minimize their own electricity bills or maximize profits rather than

CRediT authorship contribution statement

Juan Pablo Carvallo: Conceptualization, Methodology, Visualization, Supervision, Project administration, Writing - original draft, Writing - review & editing. Nan Zhang: Software, Formal analysis, Data curation, Writing - original draft, Writing - review & editing. Sean P. Murphy: Investigation, Writing - original draft, Writing - review & editing. Benjamin D. Leibowicz: Methodology, Writing - original draft, Writing - review & editing, Supervision, Project administration. Peter H. Larsen:

Acknowledgements

We are grateful for constructive comments provided by a group of technical staff at a Western U.S. utility, which requested anonymity. This Western U.S. utility served as an important case study for this analysis and allowed us to refine the model in ways that more accurately represented utility operation and planning decisions.

The authors would like to Joe Paladino (U.S. Department of Energy) for supporting this research and Larry Mansueti (formerly of the U.S. Department of Energy) for

References (59)

  • N.D. Laws et al.

    On the utility death spiral and the impact of utility rate structures on the adoption of residential solar photovoltaics and energy storage

    Appl Energy

    (2017)
  • J. Nelson et al.

    High-resolution modeling of the western North American power system demonstrates low-cost and low-carbon futures

    Energy Policy

    (2012)
  • A. Alberini et al.

    Response of residential electricity demand to price: The effect of measurement error

    Energy Econ

    (2011)
  • Y. Sun et al.

    Revisiting the residential electricity demand in the United States: a dynamic partial adjustment modelling approach

    Social Sci J

    (2017)
  • Horowitz KAW, Ding F, Mather BA, Palmintier BS. The Cost of Distribution System Upgrades to Accommodate Increasing...
  • A.M.L. da Silva et al.

    Distributed energy resources impact on distribution system reliability under load transfer restrictions

    IEEE Trans Smart Grid

    (2012)
  • A.D. Mills et al.

    Planning for a distributed disruption: innovative practices for incorporating distributed solar into utility

    Planning

    (2016)
  • G. Barbose et al.

    Utility-owned residential rooftop solar: an analysis of shareholder and ratepayer impacts

    (2019)
  • G. Barbose et al.

    Tracking US photovoltaic system prices 1998–2012: a rapidly changing market

    Prog Photovoltaics Res Appl

    (2015)
  • D. Feldman et al.

    Q4 2017/Q1 2018 Solar Industry Update

    (2018)
  • Lazard. Lazard’s Levelized Cost of Storage Analysis—Version 4.0....
  • MarketWatch. Rooftop Solar PV Market 2019 grow at +15% and projected to grow double in upcoming years by in-depth...
  • J. Carvallo et al.

    Distributed resources shift paradigms on power system design, planning, and operation: an application of the GAP Model

    Proc IEEE

    (2019)
  • N.M. Frick et al.

    A Framework for Integrated Analysis of Distributed Energy Resources: Guide for States

    (2018)
  • B. Tamimi et al.

    System stability impact of large-scale and distributed solar photovoltaic generation: the case of Ontario, Canada

    IEEE J. Magazine IEEE Trans Sustain. Energy

    (2013)
  • H. Haghighat et al.

    A bilevel approach to operational decision making of a distribution company in competitive environments

    (2012)
  • A. Canova et al.

    Electrical Impact of Photovoltaic Plant in Distributed Network

    IEEE Industry Appl Annual Meeting

    (2007)
  • N. Kakimoto et al.

    Voltage Control of Photovoltaic Generator in Combination With Series Reactor

    IEEE Trans Sustain Energy

    (2011)
  • M. Thomson et al.

    Impact of widespread photovoltaics generation on distribution systems

    IET Renew Power Gener

    (2007)
  • Cited by (18)

    • Opportunities and challenges of mainstreaming distributed energy resources towards the transition to more efficient and resilient energy markets

      2022, Renewable and Sustainable Energy Reviews
      Citation Excerpt :

      The technological progress of distributed generation units, the global focus on reducing CO2 emissions and the creation of opportunities for local investors are some important reasons for including DG in generation expansion planning [56]. Centralised generation investments can be planned in a complementary way with distributed generation investments, especially considering that wholesale market prices are generally influenced by peaks in demand [57]. While DG investments could reduce the overall revenues of the entire electricity sector, this could compromise the recovery of the sector's total costs, which include the cost of producing, transporting and distributing energy.

    • Quantifying the emissions impact of repurposed electric vehicle battery packs in residential settings

      2022, Journal of Energy Storage
      Citation Excerpt :

      Similarly, grid operators can sponsor peak-shaving tariffs to reduce overall peak loads, decrease yearly consumption, and provide greater financial incentives to homeowners with BTM-BSS [52,53]. On the other hand, increased adoption of distributed PV arrays at residences tends to displace utility-scale PV investment [54], which is often the system-optimal investment. To ensure that these BSSs and PV arrays do not impede utility investments in renewables, a centralized-coordinated dispatch strategy or decentralized strategy with a system-optimal goal may be of interest for utility operators, such as storing excess rooftop solar and abating the use of peaker power plants.

    • Coordinated management of centralized and distributed generation in an integrated energy system using a multi-agent approach

      2022, Applied Energy
      Citation Excerpt :

      The model identifies and describes key parameters that determine burgeoning trends in system development, including the following: investment and operating costs, system flexibility requirements, demand response capabilities, demand for heating of buildings, capital expenditures for the network spend, and network losses. In [9] the authors consider the development of an energy system with centralized generation, transmission, distribution, and distributed energy resources. They adapt the capacity expansion model to represent both centralized and decentralized decision-making paradigms under different structures of electricity tariffs.

    • The key role of aggregators in the energy transition under the latest European regulatory framework

      2022, International Journal of Electrical Power and Energy Systems
      Citation Excerpt :

      The case study conducted by Arefifar et al. shows that collective energy management at system level procures higher benefits and lower system losses, compared to performing the optimisation for the individual MGs [99]. The investigation by Carvallo et al. yields similar results, remarking that centralised decisions can lead to significant cost reductions [100]. Molzahn et al. emphasise the potential advantages of distributed algorithms over centralised ones, such as the improved cyber security, system robustness, and potential computational superiority regarding maximum problem size and solution speed [23].

    View all citing articles on Scopus
    View full text