In recent years, the rapid integration of renewable energy sources, particularly solar power, has transformed the global energy landscape. As a key component, solar inverters play a critical role in converting direct current from photovoltaic panels into alternating current for grid integration. However, the high penetration of solar energy introduces challenges to grid stability, including frequency fluctuations and voltage instability. Traditional power control systems in solar power stations, such as Automatic Generation Control (AGC) and Automatic Voltage Control (AVC), often operate on non-real-time systems with communication delays, resulting in slow response times—typically in the order of seconds. This inadequacy hinders the ability of solar inverters to provide rapid support during grid disturbances. To address this, we have developed a millisecond-level rapid power control system that leverages the fast power exchange capabilities of solar inverters. Our system enables solar power stations to participate in primary frequency regulation and dynamic reactive power response, enhancing grid stability. This article details the architecture, control algorithms, technological implementations, and field test results of our proposed system, demonstrating its effectiveness in real-world applications.
The existing power control framework in solar power stations relies heavily on AGC and AVC subsystems, which communicate with solar inverters and reactive power compensation devices like Static Var Compensators (SVC) or Static Var Generators (SVG). As illustrated in the system diagram, these components are interconnected through monitoring networks. However, the use of non-real-time operating systems and standard communication protocols introduces significant latency. For instance, AGC/AVC systems process instructions with delays that can extend to several seconds, making them unsuitable for emergency grid support. In contrast, solar inverters inherently possess the capability for rapid power adjustment, often within milliseconds, due to their power electronic-based design. By optimizing this potential, our system bypasses traditional bottlenecks, enabling solar inverters to respond swiftly to grid demands. This shift is crucial as power systems evolve toward higher renewable penetration, where inertia decreases and the need for fast frequency and voltage support increases.
Our rapid power control system architecture comprises three main components: the solar power station, a rapid power control device, and a high-speed communication network. The rapid power control device is installed at the grid connection point of the solar power station, where it directly samples voltage and current signals to monitor frequency and voltage variations in real time. This device employs advanced algorithms to compute the required active and reactive power adjustments based on grid conditions. Subsequently, it issues control commands to the solar inverters via a high-reliability fiber-optic ring network. This network utilizes the Generic Object Oriented Substation Event (GOOSE) protocol for multicast communication, ensuring low-latency and reliable transmission of commands. By grouping and controlling multiple solar inverters simultaneously, the system achieves coordinated power regulation, transforming the solar power station into an active participant in grid stability.
The control process begins with continuous monitoring of the grid connection point. The rapid power control device calculates the frequency and voltage deviations from setpoints, such as the nominal frequency of 50 Hz. For active power control, the system implements a droop control strategy to enable primary frequency regulation. The droop characteristic defines the relationship between frequency deviations and active power output. For example, if the frequency rises above a deadband, the solar inverters reduce their active power output proportionally. Conversely, during frequency dips, they increase output, provided reserve capacity is available. This approach allows solar inverters to mimic the inertia-like response of traditional generators. The active power adjustment, denoted as ΔP, is computed based on the droop parameters, and it is distributed among the solar inverters using an equal margin distribution strategy to ensure balanced operation and minimize iterations.
Mathematically, the active power regulation for frequency disturbances can be expressed as follows. For frequency up-regulation (e.g., when frequency exceeds the upper deadband), the target active power for the i-th solar inverter, \( p_i \), is given by:
$$ p_i = \begin{cases} p_{0_i} – k(p_{0_i} – p_{L_i}) & \text{if } k < 1 \\ p_{L_i} & \text{if } k \geq 1 \end{cases} $$
where \( p_{0_i} \) is the current active power of the i-th solar inverter, \( p_{L_i} \) is the lower active power limit, and \( k \) is the proportional coefficient calculated as:
$$ k = \frac{\Delta P}{\sum_{i=1}^{N} (p_{0_i} – p_{L_i})} $$
Here, \( N \) represents the number of operational solar inverters, and \( \Delta P \) is the total active power adjustment derived from the frequency droop characteristics. Similarly, for frequency down-regulation, the target active power is:
$$ p_i = \begin{cases} \eta p_{H_i} & \text{if } k \geq 1 \\ p_{0_i} + k(p_{H_i} – p_{0_i}) & \text{if } k < 1 \end{cases} $$
with \( p_{H_i} \) being the upper active power limit of the solar inverter, and \( \eta \) a calibration parameter (defaulting to 1). The coefficient \( k \) is:
$$ k = \frac{\Delta P}{\sum_{i=1}^{N} (p_{H_i} – p_{0_i})} $$
This ensures that solar inverters adjust their output efficiently, even under varying solar irradiance conditions. To facilitate active power increase during frequency dips, the solar power station may operate with reserved capacity or integrate energy storage systems, enhancing flexibility.
For reactive power control, the system focuses on voltage support by dynamically adjusting the reactive power output of solar inverters. The rapid power control device continuously estimates the grid impedance at the connection point using an intelligent multi-state sequence discrimination algorithm. This allows real-time computation of the required reactive power to maintain voltage stability. The target reactive power for the i-th solar inverter, \( q’_i \), is determined based on the total reactive power demand for capacitive (Q_c) or inductive (Q_l) support:
$$ q’_i = \begin{cases} q_{\max_i} & \text{if } \xi_c \geq 1 \\ q_i + \xi_c (q_{\max_i} – q_i) & \text{if } \xi_c < 1 \\ q_i + \xi_l (-q_{\max_i} – q_i) & \text{if } \xi_l < 1 \\ -q_{\max_i} & \text{if } \xi_l \geq 1 \end{cases} $$
where \( q_i \) is the current reactive power output, \( q_{\max_i} \) is the maximum reactive power the solar inverter can inject, and \( -q_{\max_i} \) is the maximum it can absorb. The coefficients \( \xi_c \) and \( \xi_l \) are calculated as:
$$ \xi_c = \frac{Q_c}{\sum_{i=1}^{N} (q_{\max_i} – q_i)} \quad \text{and} \quad \xi_l = \frac{Q_l}{\sum_{i=1}^{N} (-q_{\max_i} – q_i)} $$
This equal proportional distribution algorithm ensures that reactive power is balanced across all solar inverters, preventing overloading of individual units and optimizing thermal management. Initially, solar inverters operate at unity power factor (i.e., \( q_i = 0 \)), but they swiftly switch to reactive power mode during voltage sags or swells, providing dynamic support within milliseconds.
The technological innovations in our system enable these rapid responses. First, high-precision frequency measurement is achieved through a combination of hardware and software techniques. The rapid power control device employs low-pass filtering and microsecond-level high-speed sampling to capture grid signals accurately. In software, algorithms such as zero-crossing curve fitting and hard Discrete Fourier Transform (DFT) are used to compute frequency with an accuracy better than 0.001 Hz under ideal conditions and 0.003 Hz in the presence of harmonics. This ensures reliable detection of frequency deviations for primary frequency regulation. Second, system impedance measurement is performed autonomously using the intelligent multi-state sequence discrimination algorithm, which periodically updates the impedance values without manual intervention. This facilitates quick voltage perception and reactive power control. For instance, during voltage transients, the device calculates the required reactive power within tens of milliseconds, enabling solar inverters to respond almost instantaneously.
Communication reliability is another critical aspect. The fiber-optic ring network connects the rapid power control device to the solar inverters, providing redundancy and high bandwidth. Each node in the ring supports bidirectional data exchange and intelligently filters duplicate packets, ensuring that only the earliest valid commands are executed. In case of a single link failure, the network maintains operation through alternative paths. The use of GOOSE protocol further enhances reliability by incorporating retransmission mechanisms, which guarantee command delivery even in noisy environments. This robust communication infrastructure allows for millisecond-level command propagation, which is essential for coordinating multiple solar inverters in large-scale power stations.
To validate our system, we conducted extensive field tests at a pilot solar power station. The tests focused on primary frequency regulation and dynamic reactive power response under various operating conditions. For primary frequency regulation, we defined parameters such as a frequency deadband of ±0.06 Hz, a droop rate of 3%, and power adjustment limits of 10% of the rated power for both increase and decrease. We tested four scenarios: limited and unlimited power operation at low (20-30% of rated power) and high (50-90% of rated power) output ranges. In frequency step disturbance tests, the system demonstrated response times as low as 0.08 seconds and control deviations under 0.34%. The following table summarizes key results from these tests:
| Test Type | Step Target (Hz) | Response Lag (s) | Response Time (s) | Settling Time (s) | Initial Power (MW) | Final Power (MW) | Control Error (%) |
|---|---|---|---|---|---|---|---|
| T1 (Up-step, Low Power Limited) | 50.21 | 0.07 | 0.11 | 0.83 | 5.49 | 3.54 | 0.02 |
| T2 (Down-step, Low Power Limited) | 49.79 | 0.09 | 0.12 | 0.89 | 5.50 | 7.45 | 0.02 |
| T3 (Up-step, Low Power Unlimited) | 50.21 | 0.05 | 0.08 | 1.01 | 5.60 | 3.64 | 0.04 |
| T4 (Up-step, High Power Limited) | 50.21 | 0.10 | 0.15 | 0.40 | 14.02 | 12.06 | 0.04 |
| T5 (Down-step, High Power Limited) | 49.79 | 0.08 | 0.12 | 0.69 | 12.02 | 14.04 | 0.34 |
| T6 (Up-step, High Power Unlimited) | 50.21 | 0.05 | 0.11 | 0.35 | 15.51 | 13.55 | 0.04 |
Additionally, we simulated real grid frequency disturbances to assess performance under dynamic conditions. In these tests, the active power output of the solar inverters closely tracked the theoretical values, with minimum output response合格率 exceeding 73.71% and integral power合格率 over 96.53%. This confirms the system’s ability to provide consistent frequency support across varying operational states.
For dynamic reactive power response, we tested the system under voltage sag and swell conditions. During a voltage swell of 513 V, the solar inverters injected capacitive reactive power within 27.12 ms, stabilizing the voltage with an overshoot below 0.5% of the nominal voltage. Similarly, during a voltage sag of 508 V, they absorbed reactive power within 23.15 ms, achieving steady-state accuracy within 0.2% of the nominal voltage. The reactive power control capability was further validated with a maximum adjustment error of 1.91% and a response time of 24.36 ms. These results demonstrate that solar inverters can effectively participate in voltage regulation, reducing reliance on traditional compensation devices like SVG or SVC.
Long-term operation data from the pilot site further substantiates the system’s reliability. Over a single day, the grid connection point voltage adhered closely to the target curve, with deviations remaining within 0.2% of the nominal voltage and no instances of voltage limits being exceeded. Extended monitoring over a week showed consistent performance, meeting all grid compliance requirements for voltage control. The integration of solar inverters into rapid power control not only enhances grid stability but also improves the economic efficiency of solar power stations by minimizing the need for additional hardware.
In conclusion, our rapid power control system represents a significant advancement in the utilization of solar inverters for grid support. By leveraging fast communication, precise measurement techniques, and advanced control algorithms, we have enabled solar power stations to provide millisecond-level responses to frequency and voltage disturbances. The field tests confirm that the system achieves primary frequency regulation with response times under 0.15 seconds and dynamic reactive power response within 30 ms, surpassing traditional approaches. This capability is crucial for future power systems with high renewable penetration, where fast and reliable power control is essential. As solar energy continues to grow, the role of solar inverters in grid stability will become increasingly important, and our system provides a scalable and effective solution to meet these challenges.