A method and system for coordinated secondary frequency modulation control of a polygeneration energy storage system and conventional generating unit

Abstract

The application provides a kind of polymeric energy storage system and conventional unit collaborative secondary frequency modulation control method and system, based on model predictive control framework, the secondary frequency modulation state space model containing energy storage aggregation model and conventional unit is established and is discretized;Multi-objective optimization function containing frequency deviation penalty term, adjustment mileage cost term and energy storage energy state recovery term is constructed, and power and climbing constraint is set;In each control cycle, the optimal control instruction is solved rolling, and the aggregated energy storage power instruction and conventional unit control signal are executed respectively.The application can effectively suppress the invalid response of energy storage to high-frequency ripple of frequency, reduce the adjustment mileage, while actively maintaining the sustainable adjustment capability of energy storage, and is suitable for energy storage and conventional unit joint participation in frequency modulation auxiliary service market.

Description

Technical Field

[0001] This invention relates to the field of automatic power generation control technology, and in particular to a method and system for coordinated secondary frequency regulation control of a combined energy storage system and a conventional generating unit. Background Technology

[0002] The continuous expansion of new energy installed capacity has placed unprecedented pressure on power grid frequency regulation, making it difficult for traditional frequency regulation methods to maintain frequency stability independently. Against this backdrop, energy storage devices are demonstrating unique value, capable of precisely compensating for power shortages in a very short time, and are becoming a core force in ensuring power grid frequency stability. In practical engineering, energy storage systems are typically used in conjunction with existing conventional generator units for secondary frequency regulation to fully leverage their complementary advantages.

[0003] In existing technologies, methods for coordinating the control of energy storage and conventional generating units mainly face the following problems:

[0004] (1) There is a significant difference in response speed between energy storage systems and conventional generating units. Energy storage systems have millisecond-level or even microsecond-level response capabilities, while conventional generating units are limited by mechanical inertia and ramp rate, resulting in slower response speeds, reaching the second or minute level. If the Automatic Generation Control (AGC) signal is simultaneously sent to two types of resources, the energy storage system will frequently chase the high-frequency ripple component in the frequency signal. Even if the system's power generation and consumption balance has been achieved, energy storage will still generate a large amount of ineffective regulation mileage, leading to unnecessary charge-discharge cycle losses, accelerating battery aging, and increasing operating costs. In the prior art, a simple method is to increase the filter time constant to smooth the energy storage command, but this will limit the energy storage's rapid regulation potential when the actual frequency deviation occurs, weakening its due frequency regulation advantage.

[0005] (2) Existing coordinated control methods are not perfect in managing the energy state of storage. Some methods only perform forced recovery when the energy storage state of charge (SoC) exceeds a threshold. This passive recovery method may adversely affect AGC performance and even cause secondary frequency fluctuations. Although some methods introduce energy balance feedback loops, they fail to coordinate the recovery of energy storage state with frequency regulation and regulation mileage optimization, making it difficult to maintain the continuous regulation capability of energy storage while ensuring frequency regulation performance. When energy storage cannot discharge due to low energy state or cannot charge due to high energy state, its actual usable frequency regulation capacity will drop significantly, affecting the overall frequency regulation capability of the system.

[0006] Furthermore, most existing coordinated control methods are based on heuristic rules or simple first-order filter decomposition, lacking the ability to predict future system states (such as load disturbances and frequency change trends). This makes control decisions often lag behind system changes, making it impossible to achieve optimal coordination of multiple objectives while satisfying power constraints, ramp-up constraints, and energy storage state constraints.

[0007] Therefore, there is an urgent need to propose a secondary frequency regulation control method that can take into account frequency deviation, regulation mileage, and energy storage state optimization for the coordinated operation of energy storage systems (ESS) and conventional units. This method can leverage the rapid response advantage of energy storage, avoid ineffective regulation mileage, and actively maintain the sustainable regulation capability of energy storage. Summary of the Invention

[0008] In view of this, the purpose of the present invention is to provide a method and system for coordinated secondary frequency regulation control of aggregated energy storage system and conventional unit. This method is based on model predictive control framework, and constructs a multi-objective optimization function that includes frequency deviation, regulation mileage cost and energy storage state recovery. Under the premise of satisfying the physical constraints of the system, the optimal control command is solved in a rolling manner to realize the coordinated cooperation between aggregated ESS and conventional unit, thereby ensuring frequency regulation performance, reducing ineffective regulation mileage and actively restoring energy storage state.

[0009] To achieve the above objectives, the present invention adopts the following technical solution: a method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit, comprising the following steps:

[0010] Step 1: Establish a continuous state-space model for secondary frequency regulation that includes an energy storage aggregation model and a conventional generator set;

[0011] Step 2: Discretize the continuous state-space model to obtain a discrete state-space model;

[0012] Step 3: Based on the discrete state-space model, construct the prediction equations for model predictive control;

[0013] Step 4: Construct a multi-objective optimization function J, which includes a frequency deviation penalty term J. F Adjusting mileage cost item J RC and energy storage state of recovery term J ER ;

[0014] Step 5: Set constraints for the optimization problem, including aggregated energy storage power constraints, conventional unit power constraints, and conventional unit ramping constraints;

[0015] Step 6: In each control cycle, solve the constrained multi-objective optimization function J to obtain the optimal control signal sequence in the prediction time domain. Take the control signal of the first control step of the optimal control signal sequence, send the aggregated energy storage power command in the control signal of the first control step to the energy storage aggregation controller, and send the conventional unit control signal to the conventional generator set for execution.

[0016] In a preferred embodiment, step 1 defines a state variable vector:

[0017]

[0018] Where Δf is the frequency deviation, SoDE is the dispatchable energy state of aggregated energy storage, and Δp G Δp represents the change in output power of a conventional generating unit. t and Δp v For intermediate state variables of the unit; define the control signal vector u=[Δp E , Δu G ] T , where Δp E For the total power command of the aggregated ESS, Δu G For conventional generating units, the control signal is T, which represents transpose; the system disturbance is defined as d = Δp. mis This represents the power mismatch; the continuous state-space model expresses it as:

[0019]

[0020] Where A c B c D c It is a coefficient matrix.

[0021] In a preferred embodiment, step 2 employs the forward Euler method, using the control period Δt as the step size, to discretize the continuous model into:

[0022]

[0023] Where A d =I+ΔtA c B d =ΔtB c D d =ΔtD c ; I is the identity matrix.

[0024] In a preferred embodiment, the prediction time domain length is set to N in step 3. p The discrete state-space model is extended to a system of state transition equations in the prediction time domain, used to predict the future N. p The system status of the step.

[0025] In a preferred embodiment, the multi-objective optimization function J is constructed in step 4 as follows:

[0026]

[0027] J F J RC and J ER These are the frequency deviation penalty item, the conventional unit adjustment mileage cost item, and the energy storage state recovery item.

[0028] In a preferred embodiment, step 5 sets constraints for the optimization problem, such that the aggregated energy storage power Δp E (k) Conventional unit control signal Δu G (k) and the gradient [Δu] G (k)-Δu G [(k-1)] / Δt does not exceed its upper and lower limits, specifically:

[0029]

[0030]

[0031]

[0032] Where Δp ERTu and Δp ERTl These represent the upper and lower limits of aggregated energy storage capacity, Δu. Gu and Δu Gl These are the upper and lower limits of the control signal for conventional generating units, Δr Gu and Δr Gl These represent the upper and lower limits of the gradient rate.

[0033] In a preferred embodiment, in step 6, the constrained quadratic optimization problem described above is solved in each control cycle to obtain the optimal control signal sequence [u(0), u(1), …, u(N)] in the prediction time domain. p -1)], take the control signal u(0)=[Δp] for the first control step. E (0), Δu G (0)] T , will Δp E (0) Send to the energy storage aggregation controller, Δu G (0) Send to conventional generator set for execution;

[0034] The system runs continuously, repeating steps 1 to 6 above in each control cycle to achieve optimal coordinated control in the rolling time domain.

[0035] The present invention also provides a secondary frequency regulation control system for a combined energy storage system and a conventional generating unit, comprising a processor, a memory, and a bus. The memory stores machine-readable instructions executed by the processor. When the system is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the method for secondary frequency regulation control of a combined energy storage system and a conventional generating unit is as described above.

[0036] Compared with existing technologies, this invention has the following advantages: Compared to existing coordinated control schemes that only focus on instantaneous power allocation or simple filter decomposition, the method provided by this invention, for the first time, incorporates frequency deviation, regulation mileage cost, and energy storage state recovery into a unified multi-objective optimization framework. Through model predictive control, the system model is used to predict future states, and the optimal control command is solved in a rolling manner while satisfying power constraints and ramp-up constraints. Compared with existing technologies, this invention has the following advantages: By using the regulation mileage cost term in the optimization function, the response of the energy storage system to high-frequency ripple components is actively suppressed, significantly reducing ineffective regulation mileage and extending the service life of the energy storage system while ensuring frequency regulation performance; the rolling optimization framework based on model predictive control can anticipate future disturbances and system state changes, achieving coordinated operation between aggregated ESS and conventional units, fully utilizing the rapid response capability of energy storage to compensate for the ramp-up lag of conventional units, while avoiding excessive regulation of energy storage; this method is compatible with existing AGC communication and control architectures, achieving multi-objective collaborative optimization of frequency regulation performance, economy, and energy storage sustainability without increasing hardware costs. Attached Figure Description

[0037] Figure 1 This invention relates to a secondary frequency regulation model for power systems;

[0038] Figure 2 This is a flowchart illustrating the collaborative secondary frequency modulation implementation proposed in this invention;

[0039] Figure 3 These are the system load curve and the power mismatch curve;

[0040] Figure 4 This refers to the overall operational adjustment performance of the method proposed in this invention within 24 hours. Detailed Implementation

[0041] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0042] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of this application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0043] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations according to this application; as used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise; furthermore, it should be understood that when the terms “comprising” and / or “including” are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0044] This invention provides a secondary frequency regulation control method for coordinating an energy storage system (ESS) with a conventional generating unit. Based on a model predictive control framework, it constructs a multi-objective optimization function that includes frequency deviation, adjustment mileage cost, and energy storage state recovery, and solves for the optimal control command while satisfying the physical constraints of the system, thereby achieving coordinated operation between the ESS and the conventional generating unit.

[0045] Appendix Figure 1 This invention relates to a secondary frequency regulation model for power systems. The system includes an energy storage aggregation controller, conventional generator sets, a system frequency response model, and the model predictive control-based coordinated secondary frequency regulation controller proposed in this invention.

[0046] In the k-th control step, the upper-level control loop will aggregate the power change Δp of the ESS. E (k) Provided to the secondary frequency modulation controller, this change is generated jointly by all aggregated ESSs participating in secondary frequency modulation. Δp E (k) The following restrictions must be met:

[0047]

[0048] Where Δp ERTu and Δp ERTl These represent the upper and lower limits of the change in ESS power.

[0049] Aggregate ESS dispatchable energy capacity D ag Defined as when N E Each ESS originates from its respective upper limit state of charge SoC ui Discharge to the lower limit SoC li The total power available for extraction. The dispatchable energy state (SoDE) of the aggregated ESS is defined as the ratio of remaining dispatchable energy to total dispatchable capacity. Δp E The relationship between (t) and SoDE(t) is as follows:

[0050]

[0051] According to the above formula, when Δp E When Δp > 0, the polymerized ESS discharges, and SoDE(t) decreases; conversely, when Δp > 0, the polymerized ESS discharges, and SoDE(t) decreases. E When SoDE(t) < 0, the aggregated ESS is charged, and SoDE(t) increases. When SoDE(t) = 0, the energy storage system is empty and cannot continue to discharge; when SoDE(t) = 1, the energy storage is at full capacity and cannot continue to charge.

[0052] Appendix Figure 2The diagram shown illustrates the collaborative secondary frequency modulation control block diagram provided by this invention. The controller receives feedback information such as system frequency deviation and aggregated energy storage status, and outputs an aggregated energy storage power command Δp. E and conventional unit control signal Δu G The specific implementation steps of the method of the present invention are as follows:

[0053] First, a continuous state-space model incorporating the energy storage aggregation model and the conventional generator set's secondary frequency regulation is established. Based on the system frequency response dynamics, the aggregated energy storage state dynamics, and the models of the conventional generator set's governor and prime mover, a state variable vector is defined:

[0054]

[0055] Where Δf is the frequency deviation. Δp t and Δp v Δp is an intermediate state variable of the unit. G The three quantities represent the changes in output power of conventional generating units, specifically as follows:

[0056]

[0057]

[0058]

[0059] Where R G T G T C T R and F H These represent the governor droop constant, governor time constant, steam chamber time constant, reheat time constant, and high-pressure cylinder power proportional coefficient, respectively. The control signal vector is defined as u=[Δp] E , Δu G ] T The system disturbance is d = Δp. mis This represents the power mismatch.

[0060] The continuous state-space model is represented as:

[0061]

[0062] Where A c B c D c The coefficient matrix is ​​represented as follows:

[0063]

[0064] ,

[0065] Where D is the load frequency droop coefficient and H is the system inertia constant.

[0066] Next, the continuous model is discretized. Using the forward Euler method with a control period Δt as the step size, a discrete state-space model is obtained:

[0067]

[0068] Where A d =I+Δt∙A c B d =Δt∙B c D d =Δt∙D c I is the identity matrix.

[0069] Then, the state transition equations for the prediction time domain are constructed. The prediction time domain length N is set. p The discrete state-space model is extended to:

[0070]

[0071] Where x(0) is the measured value of the system state in the current control cycle, x(1), x(2), ..., x(N) are the values ​​of the system state measured in the current control cycle. p ) is the predicted state variable vector. u(0), u(1), ..., u(N) p -1) and d(0), d(1), ..., d(N) p -1) represents the control signal vector and disturbance in the corresponding prediction step size, respectively.

[0072] Next, we construct the multi-objective optimization function. The optimization function J consists of three terms:

[0073] The first item is the frequency deviation penalty:

[0074]

[0075] Q F The first term is the frequency deviation penalty factor, where Δf is the frequency deviation. This penalty term is used to constrain the frequency deviation within the allowable range. The second term is the energy storage and conventional unit regulation mileage cost term, used to optimize the control frequency regulation mileage, expressed as:

[0076]

[0077] Q EMC To aggregate the ESS adjustment mileage penalty factor, Q GMC This is a penalty factor for the regulation mileage of conventional units. The introduction of the regulation mileage cost term can effectively suppress the response of regulation resources to high-frequency ripple components, reducing ineffective regulation mileage. The third term is the energy storage state of being recovery term.

[0078]

[0079] Q PN Q is the power equalization penalty factor. CR The energy state deviation penalty factor. ε(k) is a slack variable, constrained by inequalities:

[0080]

[0081] The above formula realizes the situation where the schedulable energy state SoDE(k) of aggregated energy storage exceeds the energy recovery threshold [E]. RTl E RTu Additional penalties are imposed at certain times, thereby actively restoring the energy storage state to the normal range.

[0082] The overall optimization function is the sum of the above three terms:

[0083]

[0084] Before solving the optimization problem, it is necessary to set the necessary constraints:

[0085] (1) Aggregate energy storage power constraint:

[0086] (2) Power constraints of conventional units:

[0087] (3) Ramp-up constraints for conventional units:

[0088] Where Δp ERTu and Δp ERTl These represent the upper and lower limits of aggregated energy storage capacity, Δu. Gu and Δu Gl These are the upper and lower limits of the control signal for conventional generating units, Δr Gu and Δr Gl These represent the upper and lower limits of the gradient rate.

[0089] In each control cycle, the controller acquires the current system state x(0) and obtains or predicts the future disturbance sequence d(0), d(1), ..., d(N). p -1), and then solve the above optimization problem to obtain the optimal control signal sequence [u(0), u(1), …, u(N)] in the prediction time domain. p -1)]. Take the control signal u(0)=[Δp] for the first control step. E (0), Δu G (0)] T , will Δp E (0) Send to the energy storage aggregation controller, Δu G(0) Send to conventional generator set for execution.

[0090] The system operates continuously, repeating the above steps in each control cycle to achieve optimal coordinated control in the rolling time domain. Through this mechanism, the energy storage system responds quickly when the frequency deviation is large to compensate for the ramp-up lag of conventional units, actively reduces output when the frequency deviation is small to avoid continuous changes in the energy storage state, and actively recovers when the energy state deviates from the normal range. Thus, while ensuring frequency regulation performance, it achieves multi-objective synergy of regulating mileage optimization and energy storage sustainability.

[0091] To verify the effectiveness of the method provided by this invention, Figure 3 (a) and (b) show the system load curve and Δp, respectively. mis curve. Figure 4 The overall operational control performance of the proposed frequency regulation method was demonstrated over a 24-hour period. Figure 4 (a) and Figure 4 (b) gives the combined energy storage system power Δp. E and generator power ∆p G Because the generator's power ramp-up rate is limited, its regulation frequency is lower than that of the energy storage system. Figure 4 (c) shows the frequency deviation ∆f under the proposed method, which is well limited to a small range. Figure 4 (d) illustrates the state-of-charge deviation (SoDE) of the aggregated energy storage system (ESS). Most of the SoDE is kept within the predetermined range, and although some SoDE exceeds the range, it quickly returns to the range.

[0092] Compared to existing solutions that only focus on instantaneous power allocation or simple filter decomposition, the method provided in this invention incorporates frequency deviation, regulation mileage cost, and energy storage state of being recovery into a unified multi-objective optimization framework. Through a rolling optimization mechanism of model predictive control, the coordinated control of aggregated energy storage systems (ESS) and conventional generating units is elevated from passive response to proactive optimal decision-making without increasing hardware costs, significantly improving the economy of the frequency regulation system and the continuous regulation capability of energy storage. This method is particularly suitable for large-scale application scenarios where energy storage and conventional generating units jointly participate in the frequency regulation ancillary services market.

Claims

1. A method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit, characterized in that, Includes the following steps: Step 1: Establish a continuous state-space model for secondary frequency regulation that includes an energy storage aggregation model and a conventional generator set; Step 2: Discretize the continuous state-space model to obtain a discrete state-space model; Step 3: Based on the discrete state-space model, construct the prediction equations for model predictive control; Step 4: Construct a multi-objective optimization function J, which includes a frequency deviation penalty term J. F Adjusting mileage cost item J RC and energy storage state of recovery term J ER ; Step 5: Set constraints for the optimization problem, including aggregated energy storage power constraints, conventional unit power constraints, and conventional unit ramping constraints; Step 6: In each control cycle, solve the constrained multi-objective optimization function J to obtain the optimal control signal sequence in the prediction time domain. Take the control signal of the first control step of the optimal control signal sequence, send the aggregated energy storage power command in the control signal of the first control step to the energy storage aggregation controller, and send the conventional unit control signal to the conventional generator set for execution.

2. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 1, a state variable vector is defined: Where Δf is the frequency deviation, SoDE is the dispatchable energy state of aggregated energy storage, and Δp G Δp represents the change in output power of a conventional generating unit. t and Δp v For intermediate state variables of the unit; define the control signal vector u=[Δp E , Δu G ] T , where Δp E For the total power command of the aggregated ESS, Δu G For conventional generating units, the control signal is T, which represents transpose; the system disturbance is defined as d = Δp. mis This represents the power mismatch; the continuous state-space model expresses it as: Where A c B c D c It is a coefficient matrix.

3. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 2, the forward Euler method is used, with the control period Δt as the step size, to discretize the continuous model into: Where A d =I+ΔtA c B d =ΔtB c D d =ΔtD c ; I is the identity matrix.

4. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 3, the prediction time domain length is set to N. p The discrete state-space model is extended to a system of state transition equations in the prediction time domain, used to predict the future N. p The system status of the step.

5. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 4, the multi-objective optimization function J is constructed as follows: J F J RC and J ER These are the frequency deviation penalty item, the conventional unit adjustment mileage cost item, and the energy storage state recovery item.

6. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 5, constraints are set for the optimization problem, allowing the aggregated energy storage power Δp to be... E (k) Conventional unit control signal Δu G (k) and the gradient [Δu] G (k)-Δu G [(k-1)] / Δt does not exceed its upper and lower limits, specifically: Where Δp ERTu and Δp ERTl These represent the upper and lower limits of aggregated energy storage capacity, Δu. Gu and Δu Gl These are the upper and lower limits of the control signal for conventional generating units, Δr Gu and Δr Gl These represent the upper and lower limits of the gradient rate.

7. The method for coordinated secondary frequency regulation control of a converged energy storage system and a conventional generating unit according to claim 1, characterized in that, In step 6, the constrained quadratic optimization problem described above is solved in each control cycle to obtain the optimal control signal sequence [u(0), u(1), …, u(N)] in the prediction time domain. p -1)], take the control signal u(0)=[Δp] for the first control step. E (0), Δu G (0)] T , will Δp E (0) Send to the energy storage aggregation controller, Δu G (0) Send to conventional generator set for execution; The system runs continuously, repeating steps 1 to 6 above in each control cycle to achieve optimal coordinated control in the rolling time domain.

8. A secondary frequency regulation control system for coordinated energy storage system and conventional unit, comprising a processor, a memory, and a bus, wherein the memory stores machine-readable instructions executed by the processor; characterized in that, When the system is running, the processor and the memory communicate via a bus, and the machine-readable instructions are executed by the processor as described in any one of claims 1 to 7, which is a method for coordinated secondary frequency regulation control of aggregated energy storage system and conventional unit.

Images