The global energy landscape is undergoing a profound transformation driven by the accelerating integration of variable renewable energy sources. This paradigm shift introduces significant challenges for grid operators, particularly in maintaining real-time balance between generation and consumption to ensure system frequency stability. Automatic Generation Control serves as the primary mechanism for this purpose, issuing dispatch signals to conventional generators. However, the inherent inertia and slower ramp rates of thermal power plants often limit their ability to meet the stringent performance metrics of modern AGC markets.
The deployment of cell energy storage system has emerged as a transformative solution to augment grid flexibility. These systems offer near-instantaneous power injection or absorption, perfectly complementing the slower thermal assets. The core challenge in such integrated operation lies in the finite energy capacity of the battery. An uncontrolled response strategy can quickly deplete or saturate the battery’s state of charge, rendering it unavailable for subsequent AGC commands and negating its economic and technical benefits. This necessitates an intelligent, coordinated control strategy where the thermal unit’s operation is dynamically adjusted based on the real-time SOC of the cell energy storage system.
Traditional hybrid control methods often treat the thermal unit and the battery as loosely coupled entities. The battery responds aggressively to AGC signals until its SOC hits predefined hard limits, at which point it disengages entirely, leaving the thermal unit to manage the remainder of the response alone. This approach leads to inefficient battery utilization, accelerated degradation due to deep cycling, and suboptimal overall AGC performance. The key to sustainable and high-performance operation is to maintain the SOC of the cell energy storage system within a preferred, central range, ensuring it is perpetually ready to respond. This paper investigates and formalizes a coordinated control paradigm where the thermal unit’s load setpoint is intelligently biased in real-time, informed by the SOC, to proactively regulate the battery’s energy level while seamlessly meeting AGC demands.
$$ SOC(t) = SOC(t_0) + \frac{1}{C_{rated}} \int_{t_0}^{t} \eta P_{bat}(\tau) d\tau $$
Where \( SOC(t) \) is the state of charge at time \( t \), \( C_{rated} \) is the rated energy capacity of the cell energy storage system, \( P_{bat}(t) \) is the battery power (positive for discharge), and \( \eta \) is the charging/discharging efficiency.
Fundamental Principles of AGC and Battery Dynamics
The integration of a cell energy storage system into the AGC framework of a thermal power plant creates a hybrid energy delivery system. The plant’s Remote Terminal Unit receives the AGC command \( P_{AGC} \) from the grid operator. In a coordinated setup, this command is not sent directly to the turbine governor alone. Instead, a high-level controller partitions the total required adjustment \( \Delta P = P_{AGC} – P_{gen} \) (where \( P_{gen} \) is the current generator output) between the thermal unit and the battery.
The battery’s response is governed by its available power and energy limits:
$$ P_{bat}^{min}(SOC) \leq P_{bat}(t) \leq P_{bat}^{max}(SOC) $$
$$ SOC_{min} \leq SOC(t) \leq SOC_{max} $$
The available power limits, \( P_{bat}^{min} \) and \( P_{bat}^{max} \), are often themselves functions of SOC to protect the battery. The primary objective of a basic controller is to minimize the Area Control Error by allocating the fast-responding component to the battery:
$$ P_{bat}^{ref} = K \cdot \Delta P $$
subject to the constraints above. However, this myopic strategy fails to manage the integral of power over time—the energy content or SOC.
Sustained upward or downward AGC trends can drive the SOC to its extremes. When \( SOC \approx SOC_{min} \), the cell energy storage system cannot discharge further to meet a positive \( \Delta P \). Conversely, when \( SOC \approx SOC_{max} \), it cannot charge to meet a negative \( \Delta P \). In these states, the battery is forced to “sit out” of the AGC response, degrading the combined system’s performance score and potential market revenue. Furthermore, frequent operation at low or high SOC extremes, and deep charge/discharge cycles, are primary drivers of lithium-ion battery degradation, directly impacting the project’s long-term economics.
Proposed SOC-Synchronized Coordinated Control Strategy
The proposed strategy introduces a dynamic feedback loop from the cell energy storage system’s SOC to the thermal unit’s load setpoint calculation. The core philosophy is to use the thermal unit’s excess capacity (when available) to subtly “pre-charge” or “pre-discharge” the battery, steering the SOC back towards a target zone (e.g., 45%-55%) during normal AGC response, thereby preventing it from reaching the disabling limits.
The control logic is bifurcated based on the direction of the AGC command and the current SOC. The strategy employs a set of thresholds and dynamic biases, as summarized in the following algorithmic table.
| AGC Command Direction | SOC Condition | Thermal Unit Condition | Action & Bias Calculation |
|---|---|---|---|
| Increase (\(\Delta P > 0\)) | \( SOC \leq X \) (e.g., 10%) | \( P_{gen} < 0.9 \cdot P_{rated} \) | Battery discharge = 0. Add positive bias \( B_+ \) to \( P_{AGC} \). \( B_+ = f(P_{gen}, P_{AGC}, \text{Main Steam Pressure}) \). |
| \( SOC > X \) | Any | Standard coordinated discharge. No bias. | |
| Decrease (\(\Delta P < 0\)) | \( SOC \geq (1 – X) \) (e.g., 90%) | \( P_{gen} > Z \) (e.g., > Min Stable Load + Margin) | Battery charge = 0. Add negative bias \( B_- \) to \( P_{AGC} \). \( B_- = f(P_{gen}, P_{AGC}, \text{Main Steam Pressure}) \). |
| \( SOC < (1 – X) \) | Any | Standard coordinated charge. No bias. |
The bias functions \( f(\cdot) \) are designed to be adaptive. For a positive bias during an AGC increase when SOC is low:
$$ B_+ = \alpha \cdot (0.9 \cdot P_{rated} – P_{gen}) \cdot \left(1 – \frac{SOC}{X}\right) $$
Where \( \alpha \) is a tunable gain \( (0 < \alpha \leq 1) \). This ensures the bias is proportional to the available “headroom” in the thermal unit and the severity of the SOC deficit. A similar, symmetrically opposite logic applies for the negative bias \( B_- \).
The state machine for this control strategy can be described as follows. The system continuously monitors \( P_{AGC} \), \( P_{gen} \), and \( SOC \). When a new AGC command is received, its direction is assessed. Based on the direction and the SOC zone, one of the four primary states in the table is entered. If a bias is applied, the thermal unit follows the modified setpoint \( P_{AGC} \pm B \). During this period, the cell energy storage system responds to the original, unbiased AGC command based on its standard algorithm. This creates a deliberate, controlled offset between the generator output and the grid demand, which is compensated by the battery charging or discharging, thus moving its SOC. Once the SOC returns to the target band \( Y \) (e.g., 50% ±5%), the bias is ramped down to zero over a defined period, allowing the generator output to smoothly converge back to the original AGC command, concluding the correction cycle.
Implementation and System Integration Details
Implementing this strategy requires modifications at the plant’s distributed control system level. The primary controller responsible for calculating the final unit load demand must be enhanced with the SOC feedback logic. This controller receives inputs from the plant RTU (AGC command), the generator (actual output), the boiler-turbine system (main steam pressure, other constraints), and the cell energy storage system’s controller (real-time SOC and available power limits).
A critical integration point is the power allocation within the cell energy storage system itself. Most battery energy storage systems consist of multiple Power Conversion System units managing parallel battery strings. An advanced allocation strategy beyond simple proportional distribution is recommended to account for varying string impedances and slight SOC differences. A weighted allocation based on both available power and SOC deviation from the target can improve balance:
$$ P_{bat,i}^{ref} = \frac{ (P_{avail,i}) \cdot (1 + \gamma \cdot (0.5 – SOC_i)) }{ \sum_{j=1}^{N} (P_{avail,j}) \cdot (1 + \gamma \cdot (0.5 – SOC_j)) } \cdot P_{bat}^{total} $$
Where \( P_{bat,i}^{ref} \) is the power reference for the i-th PCS, \( P_{avail,i} \) is its available power, \( SOC_i \) is its string’s SOC, \( \gamma \) is a weighting factor, and \( P_{bat}^{total} \) is the total power requested from the cell energy storage system by the high-level controller.
Thermal and safety management of the cell energy storage system is paramount. The proposed strategy inherently promotes longevity and safety. By maintaining SOC in the mid-range, it reduces the prevalence of high-current charging at very low SOC and high-current discharging at very high SOC, both of which stress battery chemistry and generate excess heat. Furthermore, it minimizes the time spent at voltage extremes, which are correlated with accelerated aging and higher risks of thermal runaway. The controller can be further integrated with the Battery Management System to derate power commands if cell or module temperatures approach warning levels, ensuring operational safety is never compromised for performance.
Case Study and Performance Analysis
The efficacy of the proposed SOC-synchronized control strategy was validated through deployment at a power generation facility equipped with two 300 MW thermal units and a 10 MW / 5.6 MWh cell energy storage system. Prior to implementation, the plant’s AGC performance in the regional frequency regulation market was suboptimal, with a composite performance score \( K \) averaging around 1.0 (against a maximum possible of 2.0 in that market). The performance score is typically calculated as a weighted sum of three key metrics: Regulation Rate \( K_1 \), Response Time \( K_2 \), and Regulation Accuracy \( K_3 \):
$$ K = 0.25 \times (2K_1 + K_2 + K_3) $$
Following the implementation, a marked and sustained improvement was observed. The strategy ensured the cell energy storage system was almost always available to contribute its fast response, while the thermal unit’s biased actions kept the battery’s SOC hovering in the effective response zone. The table below presents a 12-hour snapshot of performance metrics post-implementation.
| Time Interval | Composite Score \( K \) | Rate Score \( K_1 \) | Response Time Score \( K_2 \) | Accuracy Score \( K_3 \) |
|---|---|---|---|---|
| 00:00 – 01:00 | 1.73 | 2.52 | 0.96 | 0.90 |
| 01:00 – 02:00 | 1.77 | 2.62 | 0.95 | 0.89 |
| 02:00 – 03:00 | 1.80 | 2.71 | 0.95 | 0.83 |
| 03:00 – 04:00 | 1.74 | 2.56 | 0.94 | 0.90 |
| 04:00 – 05:00 | 1.88 | 2.89 | 0.96 | 0.79 |
| 05:00 – 06:00 | 1.88 | 2.84 | 0.96 | 0.87 |
| 06:00 – 07:00 | 1.89 | 2.96 | 0.96 | 0.68 |
| 07:00 – 08:00 | 1.75 | 2.65 | 0.96 | 0.72 |
| 08:00 – 09:00 | 1.89 | 2.87 | 0.94 | 0.89 |
| 09:00 – 10:00 | 1.79 | 2.68 | 0.95 | 0.83 |
| 10:00 – 11:00 | 1.81 | 2.71 | 0.96 | 0.87 |
| 11:00 – 12:00 | 1.83 | 2.75 | 0.96 | 0.87 |
| Average | 1.81 | 2.71 | 0.95 | 0.84 |
The data demonstrates a remarkable improvement. The average composite score \( K \) rose to 1.81, and the regulation rate score \( K_1 \) achieved an average of 2.71, closely approaching the market maximum of 3.0. This directly translates to superior economic outcomes. Under the performance-based market rules, the enhanced scores resulted in a significant increase in daily revenue from frequency regulation services. Beyond economics, operational data indicated a substantial reduction in battery-related thermal alarms and stress events, confirming the positive impact of mid-range SOC management on the health and safety of the cell energy storage system. The thermal unit itself operated more smoothly, as the battery absorbed the most abrupt power demands, reducing wear and tear on the boiler and turbine systems.
Conclusion and Future Perspectives
The integration of cell energy storage system with conventional thermal power generation offers a potent solution for enhancing grid frequency regulation services. However, the full potential of this hybrid system can only be unlocked through deep coordination that intelligently manages the battery’s finite energy reserve. The control strategy presented in this work, which dynamically adjusts the thermal unit’s load setpoint based on the real-time SOC of the cell energy storage system, provides a robust and effective methodology. By proactively maintaining the SOC within a central operating band, the strategy ensures the battery’s continual availability for AGC response, maximizes the combined system’s performance metrics in regulation markets, and concurrently extends battery life by mitigating deep discharge/charge cycles and operation at extreme SOC levels.
The successful field implementation validates the practical feasibility and significant benefits of the approach. Future research directions could focus on several enhancements. First, the bias calculation function \( f(\cdot) \) could be optimized using model predictive control frameworks that forecast short-term AGC signal trends, allowing for even more anticipatory SOC management. Second, the strategy could be extended to incorporate the economic trade-offs between regulation revenue, fuel costs for the thermal unit during biasing actions, and the quantified degradation cost of the cell energy storage system, formulating a real-time optimization problem. Finally, as grids evolve, similar coordination principles could be applied to hybrid systems involving energy storage paired with other slow-ramping assets or even aggregated distributed energy resources, solidifying the role of intelligent, state-aware control in the future of stable and economical power systems.