Controller Synthesis of Wind Turbine Generator and Energy Storage System with Stochastic Wind Variations under Temporal Logic Specifications

Zhe Xu, Agung Julius, Ufuk Topcu, Joe H. ChowarXiv 2019 · arXiv

Introduction

As power grids transition toward renewable energy, managing the inherent volatility of wind power becomes a critical challenge. Traditional synchronous generators provide steady inertia, but wind turbine generators (WTGs) introduce stochastic fluctuations that can jeopardize grid stability. Grid operators enforce strict "grid codes"—performance standards that require frequency and voltage to return to nominal ranges within specific timeframes following a disturbance.

This paper presents a robust framework for controller synthesis that coordinates WTGs and energy storage systems (ESS) to satisfy these complex, time-sensitive requirements. By combining formal methods with stochastic control theory, the authors provide a way to guarantee grid stability with a calculable level of probability.

Methodology: Bridging Stochasticity and Determinism

The core of the proposed approach is the ability to design controllers for a simplified, deterministic "nominal" system that are guaranteed to work on the real, noisy stochastic system.

Metric Temporal Logic (MTL)

Grid codes are often expressed as time-bound constraints (e.g., "frequency must recover within 2 seconds"). The authors use Metric Temporal Logic (MTL) to formally define these requirements. Unlike simple setpoint tracking, MTL can express complex behaviors such as "always stay within safety limits and eventually reach a target region within a specific time window."

Stochastic Control Bisimulation

To handle wind noise, the authors introduce a "stochastic control bisimulation function." This mathematical tool quantifies the maximum expected "distance" between the trajectories of the real stochastic system and a simplified deterministic model.

Illustration of shrunken safety boundaries for robust synthesis.

Robust Synthesis and Feedback

The synthesis pipeline involves two main stages:

  1. Robust Feedforward Synthesis: The safety boundaries of the nominal system are "shrunken" by a margin derived from the bisimulation function. An optimization problem is then solved to find a trajectory for the nominal system that stays within these tighter bounds.
  2. Feedback Control Law: A feedback law is generated to ensure the real system tracks the nominal trajectory. This provides resilience against unexpected disturbances, such as sudden generation loss or extreme wind gusts.

Evaluation and Results

The framework was tested on standard power system benchmarks, including a four-bus system and the IEEE 9-bus system.

Schematic of the four-bus power system model featuring a thermal plant, wind farm, and ESS.

Key Findings

  • Probabilistic Guarantees: The authors proved that if the nominal trajectory satisfies the robustified MTL specification, the actual stochastic system will satisfy the original grid code with a probability of at least $1 - \epsilon$, where $\epsilon$ is a user-defined risk factor.
  • Resilience: In simulations involving a 150 MW generation loss combined with stochastic wind variations, the coordinated feedback controller successfully maintained grid frequency within the required MTL bounds, whereas feedforward-only approaches failed.
  • Scalability: The method effectively handled the IEEE 9-bus system, managing multiple transmission line constraints simultaneously, demonstrating its potential for larger-scale grid applications.

Conclusion

This research demonstrates that formal methods like MTL, when combined with robust stochastic control, offer a powerful language for defining and enforcing modern grid regulations. While the current approach uses linearization and can be somewhat conservative, it provides a rigorous foundation for integrating volatile renewable sources into the grid with high confidence.

Summary updated

Abstract

In this paper, we present a controller synthesis approach for wind turbine generators (WTG) and energy storage systems with metric temporal logic (MTL) specifications, with provable probabilistic guarantees in the stochastic environment of wind power generation. The MTL specifications are requirements for the grid frequency deviations, WTG rotor speed variations and the power flow constraints at different lines. We present the stochastic control bisimulation function, which bounds the divergence of the trajectories of a switched stochastic control system and the switched nominal control system in a probabilistic fashion.We first design a feedforward controller by solving an optimization problem for the nominal trajectory of the deterministic control system with robustness against initial state variations and stochastic uncertainties. Then we generate a feedback control law from the data of the simulated trajectories. We implement our control method on both a four-bus system and a nine-bus system, and test the effectiveness of the method with a generation loss disturbance. We also test the advantage of the feedback controller over the feedforward controller when unexpected disturbance occurs.
BibTeX
@misc{xu2019controller,
  author = {Zhe Xu and Agung Julius and Ufuk Topcu and Joe H. Chow},
  title = {{Controller Synthesis of Wind Turbine Generator and Energy Storage System with Stochastic Wind Variations under Temporal Logic Specifications}},
  year = {2019},
  eprint = {1911.11347},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1911.11347},
}
Figure