Abstract
Performances of AGC and generators' valve controllers, connected through AGC allocation strategy, both influence active power balance in power systems. In this paper, authors consider that when distributing the whole regulation demand, AGC should give consideration to the bottom-level generator valve controller' performance. Following this idea, an AGC allocation parameter optimization method is proposed considering valve controllers' characteristics. Area Control Error (ACE) is selected as optimization criterion and optimal distribution coefficients are calculated under associated constraints, thus improving control performance of power system frequency and tie-line transmission power. The simulation of two-area system with four generators shows the effectiveness of the proposed method.
Introduction
In modern power grids, Automatic Generation Control (AGC) serves as the primary mechanism for maintaining system frequency and managing tie-line power exchanges. Traditionally, AGC operates in a hierarchical fashion: a high-level controller determines the total regulation demand, which is then distributed among individual generators. However, standard allocation methods often rely on simplistic rules, such as equal proportions or available capacity, ignoring the dynamic performance differences between various generator valve controllers.
The authors identify that generators equipped with different control technologies—ranging from traditional PID controllers to advanced nonlinear "inverse system" controllers—respond differently to regulation signals. Failing to account for these "reflexes" can lead to suboptimal grid stability. This paper proposes a performance-aware AGC allocation strategy that optimizes distribution coefficients based on the specific characteristics of the underlying hardware.
Methodology
The core of the proposed method is a feedback loop where the AGC Area Controller adjusts its allocation strategy based on the observed performance of the generator units.
The methodology involves three main components:
- Controller Characterization: The study analyzes two types of valve controllers: a linear PID-based model and a nonlinear model based on inverse system theory. The nonlinear controller is shown to better manage the inherent complexities of steam turbines.
- Optimization Model: An objective function is defined to minimize the integrated Area Control Error (ACE). ACE is a critical metric combining frequency deviation and tie-line power error. The optimization is subject to physical constraints, including generator output limits and NERC Control Performance Standards (CPS1).
- Algorithmic Solver: To determine the optimal distribution coefficients, the authors utilize a Particle Swarm Optimization (PSO) algorithm with improved inertia weights. This allows the system to intelligently weight generators with superior response characteristics.
Simulation and Results
The effectiveness of the method was validated using a four-generator, two-area interconnected power system model simulated in MATLAB/SimPowerSystems. The simulation parameters were derived from established literature to ensure a realistic benchmark.
When subjected to a 400 MW load disturbance, the PSO-optimized strategy significantly outperformed the traditional equal-proportional allocation. The algorithm assigned a higher regulation burden (65%) to the generator with the advanced nonlinear controller, while the PID-controlled unit handled the remaining 35%.
Key performance indicators showed marked improvement:
- The overall performance index (error measure) decreased from 5.999 to 4.632.
- Tie-line power integration deviation was reduced from 131.3 MW to 108.9 MW.
- As shown in the ACE plot, the proposed strategy results in lower peak deviations and faster recovery times following a disturbance.
Conclusion
This research demonstrates that AGC allocation should not be "blind" to the underlying control technology of the generators. By leveraging the superior dynamics of nonlinear valve controllers through optimized dispatch, power systems can achieve greater frequency stability and more reliable inter-area power exchange. Future work will focus on scaling this optimization for larger grids and integrating the highly variable response characteristics of renewable energy sources.
Summary updated