Design of Grid‑Forming Multi–Timescale Coordinated Control Strategies for Dynamic Virtual Power Plants
Yan Tong 1 , Qin Wang 1 , Sihao Chen 2 , Xue Hu 1 , Zhaoyuan Wu 3 · 2026
Read Design of Grid‑Forming Multi–Timescale Coordinated Control Strategies for Dynamic Virtual Power Plants by Yan Tong 1 , Qin Wang 1 , Sihao Chen 2 , Xue Hu 1 , Zhaoyuan Wu 3 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-04962.txt.
Introduction
Yan Tong and team offer a dynamic virtual power plant coordinating mixed resources across many timescales with grid-forming control. Grid-following converters use a phase-locked loop that can undermine balance in weak grids, while their design applies virtual synchronous generator control for inertia and damping. A participation factor method gauges each device's contribution through the loops linking rate to active power and voltage to reactive power. Fast resources deliver rapid reply and damping, middle resources smooth shifts, and slow resources handle steady state and low-rate regulation.
Catalog record
| CREATOR | Yan Tong 1 , Qin Wang 1 , Sihao Chen 2 , Xue Hu 1 , Zhaoyuan Wu 3 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27559 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.04962
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Design of Grid‑Forming Multi–Timescale Coordinated Control Strategies for Dynamic Virtual Power Plants Thanks: This study is supported by the State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources (Grant No.LAPS25006). Yan Tong 1 , Qin Wang 1 , Sihao Chen 2 , Xue Hu 1 , Zhaoyuan Wu 3 Abstract As the penetration level of distributed energy resources (DERs) continues to rise, traditional frequency and voltage support from synchronous machines declines. This weakens grid stability and increases the need for fast and adaptive control in a dynamic manner, especially in weak grids. However, most virtual power plants (VPPs) rely on static aggregation and plan-based resource allocation strategies. These methods overlook differences in device response times and limit flexibility for ancillary services. To address this issue, we propose a dynamic virtual power plant (DVPP) that coordinates heterogeneous resources across multiple time scales using grid-forming control. We first contrast grid-following and grid-forming converters: grid-following designs rely on a phase-locked loop which can undermine stability in weak grids, whereas our DVPP applies virtual synchronous generator control at the aggregate level to provide effective inertia and damping. Then, we introduce a dynamic participation factor framework that measures each device’s contribution through the frequency–active power and voltage–reactive power loops. Exploiting device heterogeneity, we adopt a banded allocation strategy: slow resources manage steady-state and low-frequency regulation; intermediate resources smooth transitions; and fast resources deliver rapid response and high-frequency damping. Comparative simulations demonstrate that this coordinated, timescale-aware approach enhances stability and ancillary service performance compared to conventional VPPs. I Introduction As the penetration level of distributed energy resources (DERs) continues to rise, ancillary services traditionally ensured by conventional thermal units are increasingly being shouldered by DERs. Because renewable DERs are variable and their available capacity changes over time, future power systems will require more robust frequency control and voltage regulation to maintain stability. In recent years, virtual power plants (VPPs) that aggregate the DERs have gained increasing attention as a means to deliver ancillary services and grid control [ 1 ] . The VPP represents a new paradigm for energy management and dispatch. It leverages information and communication technologies, automatic control, and intelligent algorithms to logically integrate distributed, multi-type DERs into a unified and controllable equivalent resource. Although these assets are geographically dispersed, they are integrated via the VPP platform into a single “entity” at the logical level, which can participate in electricity markets, receive dispatch instructions, and provide energy and ancillary services much like a conventional large power plant [ 2 ] . The core value of a VPP lies in unlocking the “aggregation effect” of distributed resources, which significantly enhances system flexibility, economic efficiency, and reliability. Conventional VPPs rely on relatively static aggregation, with resource allocation based on preset plans or historical data. A dynamic VPP (DVPP) extends this model by adding real-time intelligence and adaptive coordination, allowing operating strategies to adjust continuously to grid conditions, market signals, weather, and load changes across distributed generation and storage [ 3 ] . A DVPP coordinates heterogeneous DERs within their physical and operational constraints to deliver fast, system-level frequency and voltage control [ 4 ] . Existing DVPP control designs predominantly rely on aggregations of grid-following (GFL) DERs: frequency and voltage at the point of common coupling (PCC) are measured via a phase-locked loop (PLL) and used to modulate device outputs. Due to this reliance on measured frequency, GFL-based DVPPs generally require a stiff grid to operate stably; under high DER penetration or weak-grid conditions, their responsiveness and PLL tracking performance may degrade substantially and even trigger instability [ 5 ] . Aligned with the growing use of grid-forming (GFM) resources, this paper proposes a DVPP architecture built on GFM units to deliver robust dynamic ancillary services. To make the DVPP function as a single power plant from the grid’s perspective, it must coordinate wind, solar, storage, and controllable loads—each with different response times—across both time and location. This aggregation is not a simple sum of outputs; it is a layered, timescale-aware, and constraint-driven coordination scheme. The response characteristics of major resource types are summarized in Table I . At the aggregation layer, the DVPP should present its overall, externally visible dynamic capability to the market and system operator. Allocation and performance assessment are then carried out at this aggregate level, rather than through device-by-device commands. TABLE I : Typical Response Characteristics of DVPP Assets Asset Type Dynamic Characteristics Response Speed Adjustable Time Horizon Primary Control Objective Wind farm Strongly correlated with wind speed, pronounced stochasticity Minutes to hours Mid-term (5 minutes to hours) Power smoothing, production forecasting Photovoltaic plant Correlated with solar irradiance, pronounced diurnal cycles Seconds to minutes Short to mid-term (seconds to hours) Maximize energy yield, mitigate power fluctuations Energy storage system Capable of rapid charge/discharge, highly agile Milliseconds to seconds Very short-term (milliseconds to minutes) Frequency regulation, peak shaving, emergency reserve Controllable load / demand response Dependent on end-user behavior, inherent delays Minutes to hours Medium to long term Peak-valley balancing, load shifting Microgrid / virtual power plant Coordinated control enabled, flexible operational strategies Seconds to minutes Short to mid-term Power coordination, backup provision • Note: Renewable generation (wind and solar) evolves comparatively slowly, storage responds the fastest, whereas loads and microgrids bridge the intermediate temporal layers. The main contributions of this paper are summarized as follows: 1) We propose a DVPP framework that combines GFM capability with dynamic aggregation. By introducing virtual synchronous generator (VSG) control at the aggregation layer, the DVPP provides quantifiable effective inertia and damping, thereby improving stability margins under weak-grid and high-penetration conditions. 2) We develop a coordination strategy centred on dynamic participation factors that explicitly accounts for time-varying resource capacities and disparate response timescales. The strategy enables bandwidth-aware task allocation from milliseconds to hours and aligns the aggregated external behaviour with system-level requirements, offering a deployable control paradigm for ancillary services. 3) To facilitate reproducibility, we provide the algorithms required for the experimental design as well as the datasets used in the experiments. In summary, this paper develops a DVPP architecture supported by GFM control. Section II presents the control design and implementation pathway for GFM inverters within the DVPP. Section III proposes a dynamic participation factor framework for the coupled \omega-P power and v-Q power channels. Guided by device-level physical constraints and timescale characteristics, the desired aggregate behavior is decomposed across wind/PV, energy storage, and controllable loads, and multi–timescale coordination is achieved via low-/high-/band-pass characteristics. Section IV conducts three comparative studies on a modified 3-machine, 9-bus system to validate the effectiveness of the proposed method in terms of frequency coherency, Rate Of Change Of Frequency (ROCOF) suppression, enhancement of effective inertia and damping, and feasible decoupling and limit handling for active/reactive power. II GFM Control Design for DVPP II-A GFL and GFM Power electronic inverters for DERs are commonly divided into two categories [ 6 ] , as illustrated in Fig. 1 : GFL and GFM. A GFL inverter employs a PLL to track the phase angle of the existing grid voltage and subsequently regulates the AC-side current. In general, a GFL inverter consists of three control loops, namely the outer-loop power controller, the inner-loop current controller, and the PLL: • Outer-loop power