Heterogeneous frequency modulation unit distributed cooperative automatic power generation control method based on local linearization and sliding mode control

Abstract

The invention provides a heterogeneous frequency modulation unit distributed cooperative automatic power generation control method based on local linearization and sliding mode control. Aiming at the problem that the output power of an AGC (Automatic Gain Control) controller of a renewable unit is easy to be saturated after renewable energy is merged into a power grid, firstly, aiming at a complex nonlinear relation between the frequency adjustment amount of each frequency modulation unit and a regional control error, a super local linear model of each unit and a state observer based on sliding mode control are innovatively designed, and the design of the controller is simplified. On the basis, an AGC controller based on sliding mode control is further designed, the advantages of robustness and stability of sliding mode control are inherited, and the system frequency response performance is improved.

Description

Technical Field

[0001] The present invention belongs to the field of automatic power generation control, and in particular relates to a distributed cooperative automatic power generation control method for heterogeneous frequency modulation units based on local linearization and sliding mode control. Background Art

[0002] Generally speaking, Automatic Generation Control (AGC) services are provided by traditional thermal power units and hydropower units [1]. However, as the number of renewable energy units in new power systems increases significantly, the system's rotational inertia decreases, and the challenges faced by frequency regulation are becoming increasingly greater. In order to improve the system's frequency response performance, it is urgent to tap the frequency response potential of multiple types of power sources.

[0003] In recent years, domestic and foreign scholars have continuously proposed some new AGC strategies to enable wind turbines, photovoltaic generators and battery energy storage systems to participate in frequency response services. Reference [2] aims to solve the problem of imbalance between power generation and load faced by power systems containing a large amount of renewable energy, and integrates the prediction results of disturbances such as photovoltaics and wind power into the input of the model predictive controller for auxiliary control, thereby minimizing the deviation caused by uncertain disturbances. Reference [3] proposes a dual-time scale automatic generation control scheme based on stochastic model predictive control, which sets the AGC command of the fast response unit in each AGC cycle as a multi-step signal, while the traditional generator unit maintains a single-step signal, greatly exploring the frequency regulation capability of the fast response unit. Reference [4] proposes an AGC reinforcement learning algorithm for multi-regional interconnected power grids to optimize the control performance of the AGC controller in response to the increase in system disturbances caused by the continuous increase in the penetration rate of new energy and the decrease in frequency control capability caused by the insufficient capacity of traditional frequency response units. Reference [5] addresses the problem of frequency response performance degradation caused by the continuous increase in renewable energy penetration in the automatic generation control of power systems. It proposes an AGC allocation optimization strategy based on data-driven prediction of key frequency distribution parameters, and updates the AGC allocation coefficient in advance by performing particle swarm optimization on the prediction results. Reference [6] addresses the problem of randomness of renewable energy and proposes a cascade PI controller for renewable integrated power systems based on the Brown Bear optimization algorithm.

[0004] In addition, sliding mode control is gradually being applied to automatic power generation control. Compared with the PI-based AGC control strategy, sliding mode control improves system robustness and control performance. When there are random and uncertain interferences in the power grid, the power system can still maintain better frequency regulation performance. Reference [7] proposed a distributed sliding mode control strategy for optimal load frequency control in the power network, while achieving frequency regulation and economic dispatch. Reference [8] proposed an improved sliding mode variable and applied auxiliary control based on adaptive dynamic programming for online weight update, which improved the system's response capability when the load changes and system parameters are uncertain. Reference [9] proposed a sliding mode controller to reduce the system frequency deviation, and also used genetic algorithms to optimize the controller parameters.

[0005] The integration of renewable energy not only enriches the diversity of power generation resources, but also provides new possibilities for the flexible dispatch and efficient operation of the power system. However, the output power of renewable energy, especially wind power and photovoltaic power generation, is directly affected by natural environmental conditions and has significant intermittent and fluctuating characteristics. This characteristic makes it easy for wind and photovoltaic generators to reach the upper limit of output power in actual operation, limiting their ability to respond quickly to changes in grid frequency, and posing a challenge to the stability of the power system and the flexibility of frequency regulation.

[0006] In order to solve the above problems, people have been seeking an ideal technical solution.

[0007] Reference 1: Li Zhongwen, Bai Ningning, Cheng Zhiping, et al. Research on distributed collaborative AGC method for heterogeneous frequency regulation units in new power systems[J]. Power System Technology, 2024, 48(06): 2327-2335.

[0008] Document 2: Ganger D, Zhang J, Vittal V. Forecast-Based Anticipatory FrequencyControl in Power Systems[J]. IEEE Transactions on Power Systems, 2018, 33(1):1004-1012.

[0009] Document 3: Shen Y, Wu W, Sun S.Stochastic Model Predictive Control BasedFast-Slow Coordination Automatic Generation Control[J]. IEEE Transactions on Power Systems, 2024, 39(3):5259-5271.

[0010] Reference 4: Xi L, Zhou L, Xu Y, et al. A Multi-Step Unified Reinforcement Learning Method for Automatic Generation Control in Multi-Area Interconnected Power Grid[J]. IEEE Transactions on Sustainable Energy, 2021, 12(2): 1406-1415.

[0011] Reference 5: Wang Z, Wang Y, Ding Z, et al. Optimal AGC allocation strategy based on data-driven forecast of frequency distribution key parameters[J]. Electric Power Sysrems Research, 2024, 226(109916).

[0012] Reference 6: Ojha S, Maddela C. Load frequency control of a two-area power system with renewable energy sources using brown bear optimization technique[J]. Electrical Engineering, 2024, 106(3): 3589-3613.

[0013] Reference 7: Trip S, Cucuzzella M, De Persis C, et al. Passivity-Based Design of Sliding Modes for Optimal Load Frequency Control[J]. IEEE Transactions on Control Systems Technology, 2019, 27(5): 1893-1906.

[0014] Document 8: Mu C, Tang Y, He H. Improved Sliding Mode Design for LoadFrequency Control of Power System Integrated an Adaptive Learning Strategy [J]. IEEE Transactions on Industrial Electronics, 2017, 64(8): 6742-6751.

[0015] Document 9: Al-Hamouz Z, Al-Duwaish H, Al-Musabi N. Optimal design of a slidingmode AGC controller: Application to a nonlinear interconnected model [J]. Electric Power Sysrems Research, 2011, 81(7): 1403-1409. Summary of the invention

[0016] The purpose of the present invention is to address the deficiencies of the prior art and to provide a distributed collaborative automatic power generation control method for heterogeneous frequency-regulating units based on local linearization and sliding mode control, so as to fully tap the frequency regulation potential of renewable frequency-regulating units and improve their dynamic response capability and frequency regulation efficiency in the power system.

[0017] In order to achieve the above object, the technical solution adopted by the present invention is:

[0018] Firstly, a distributed area control error (ACE) mining algorithm is designed. Each frequency response unit can estimate the ACE signal of each area in real time and accurately, and provide necessary state feedback for each heterogeneous frequency modulation AGC unit to design an independent AGC controller based on sliding mode control. Secondly, according to the complex nonlinear relationship between the frequency adjustment amount of each frequency modulation unit and the area control error, the super local linear model of each heterogeneous frequency modulation AGC unit and the state observer based on sliding mode control are innovatively designed to simplify the controller design. On this basis, an AGC controller based on sliding mode control is further designed, which inherits the robustness and stability advantages of sliding mode control and improves the system frequency response performance. In addition, a distributed proportional load allocation method is proposed. Through a reasonable power allocation strategy, the frequency modulation capacity can be released in time for the units with fast response capability to participate in the next stage of frequency response, thereby further improving its dynamic response speed and frequency regulation efficiency in the power system.

[0019] The present invention has the following beneficial effects and advantages:

[0020] 1. For new power systems containing a large amount of renewable energy, the AGC controller is locally linearized to solve the problem that the output power of the AGC controller is easily saturated, providing available capacity for the next round of frequency regulation, fully tapping the frequency regulation potential of renewable frequency regulation units, and improving their dynamic response capability and frequency regulation efficiency in the power system.

[0021] 2. Aiming at the complex nonlinear relationship between the frequency adjustment amount of each frequency-modulating unit and the regional control error, an ultra-local linear model of each unit and a state observer based on sliding mode control are innovatively designed. By linearizing the local characteristics of the system, the controller design is simplified.

[0022] 3. An AGC controller based on sliding mode control is designed to replace the traditional PI controller. It inherits the robustness and stability advantages of sliding mode control, effectively reduces the regional control error caused by the frequency regulation unit, and improves the frequency regulation capability of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is the dynamic response model of the AGC unit.

[0024] Figure 2 The present invention has an automatic power generation control block diagram of a heterogeneous frequency modulation AGC unit.

[0025] Figure 3 It is a performance comparison of different AGC methods;

[0026] (a) Regional control error; (b) Frequency deviation; (c) Load disturbance; (d) CPS1 value.

[0027] Figure 4 are the output and per unit value of each unit under different methods;

[0028] (a) Output changes of each unit under the distributed AGC method proposed in the present invention;

[0029] (b) the per-unit output value of each unit under the distributed AGC method proposed by the present invention;

[0030] (c) Output changes of each unit under DPI method 1;

[0031] (d) Unit output value of each unit under DPI method 1;

[0032] (e) Output changes of each unit under DPI method 2;

[0033] (f) Unit output value of each unit under DPI method 2. DETAILED DESCRIPTION

[0034] The present invention is further described in detail below in conjunction with the accompanying drawings.

[0035] Example 1

[0036] like Figure 1 As shown, this embodiment provides a method for controlling distributed coordinated automatic power generation of heterogeneous frequency modulation units based on local linearization and sliding mode control, which is used in a distributed power system composed of multiple heterogeneous frequency modulation AGC units, including the following steps:

[0037] (1) A single-region power system AGC dynamic response model is used to describe each heterogeneous frequency regulation AGC unit;

[0038] (2) The sum of known constants and unknown disturbances F of each heterogeneous frequency-regulated AGC unit in a single-region power system 1,i and regional control error ε i , respectively perform super local linear model state observations and obtain the corresponding estimated values and i represents the i-th heterogeneous frequency modulation AGC unit;

[0039] Among them, the regional control error ε i The ACE signals of each heterogeneous frequency modulation AGC unit are obtained by distributed estimation of the ACE discovery algorithm;

[0040] (3) Output power and estimated value of each heterogeneous frequency modulation AGC unit and estimated values Adopting AGC algorithm based on sliding mode control to control the regional control error of each heterogeneous frequency modulation AGC unit to zero;

[0041] (4) In the final stage of frequency regulation, each heterogeneous frequency regulation AGC unit is assigned the frequency regulation task in proportion to the available capacity allocated by the set distributed proportional load allocation algorithm.

[0042] AGC dynamic response model for single-area power system

[0043] The steps for establishing the AGC dynamic response model of the single-area power system are as follows:

[0044] 1-1) The first-order transfer function is used to represent the dynamic response model of the governor-turbine controller of each heterogeneous frequency modulation AGC unit:

[0045]

[0046] Where, the subscript i is the unit number, T t,i is the time constant;

[0047] 1-2) Define the regional control error ε(t):

[0048] ε(t)=β·Δf(t)

[0049] Among them, Δf is the system frequency deviation, β is the frequency deviation coefficient;

[0050] 1-3) Primary control output power of each heterogeneous frequency modulation AGC unit Calculated as:

[0051]

[0052] Among them, R i The droop coefficient of the i-th AGC frequency regulation unit;

[0053] 1-4) For the regional control error, the state equation is:

[0054]

[0055] Among them, Δp t,i represents the output power of the i-th unit, Δp d represents the power disturbance, T ps and K ps Represent the system time constant and gain constant respectively;

[0056] The relationship between the regional control error and the output power of each frequency modulation unit is:

[0057]

[0058] Will With Δp d Considered as a disturbance.

[0059] State observation of hyperlocal linear model

[0060] The method for observing the state of a super local linear model is:

[0061] 2-1) Design of AGC super local linear model for single-region power system:

[0062]

[0063] Among them, η1 and ψ1 are constant coefficients, F1 represents the sum of known constants and unknown disturbances in a single-region power system, and ξ1 is expressed as the differential of F1;

[0064] 2-2) For the AGC super local linear model, its state observer is designed as:

[0065]

[0066] in, and Represent the estimated values ​​of ε and F1, μ 1,i represents the sliding mode control function, λ1 represents the gain factor;

[0067] 2-3) The observation error of the state observer is:

[0068]

[0069] 2-4) The sliding surface of the state observer is:

[0070]

[0071] Among them, the sliding mode parameter α 1,i >0;

[0072] 2-5) The sliding mode control function of the state observer is:

[0073]

[0074] The sliding mode parameter β 1,i , β 2,i >0,

[0075] 2-6) The estimation formula is:

[0076]

[0077] AGC based on sliding mode control

[0078] The AGC algorithm based on sliding mode control is:

[0079] 3-1) Output power of sliding mode controller The output power Δp of the heterogeneous frequency modulation AGC unit t,i The relationship is:

[0080]

[0081] in, The second item of the secondary frequency regulation output power of the AGC frequency regulation unit;

[0082] 3-2) The sliding surface of the sliding mode controller is:

[0083]

[0084] Among them, the sliding mode parameter α 2,i >0;

[0085] 3-3) Considering Δw as a disturbance, the AGC control law based on sliding mode control is designed as:

[0086]

[0087] Among them, the sliding mode parameter β 3,i ,

[0088] Distributed ACE Discovery and Proportional Load Distribution

[0089] 4-1) The distributed ACE mining algorithm is expressed as:

[0090]

[0091] Among them, ε(t) is the actual ACE value of the system, and are the ACE estimates of the i-th and j-th heterogeneous frequency modulation units respectively, and the gain coefficient ρ1>0. If the i-th heterogeneous frequency modulation AGC unit is connected to the control center, then c i =1, otherwise c i =0.

[0092] 4-2) Distributed proportional load distribution algorithm is:

[0093] The second term of the secondary frequency modulation output power of the i-th heterogeneous frequency modulation AGC unit for:

[0094]

[0095] Among them, the gain coefficient ρ4>0;

[0096] w i,j is an element of the symmetric matrix W, defined as:

[0097]

[0098] In the formula, the subscripts i and j represent the sequence number of the heterogeneous frequency modulation AGC unit, n i 、n j represents the number of adjacent heterogeneous frequency modulation AGC units, γ>0; Δp pu,i (t) is the per-unit value of the actual output power of the ith heterogeneous frequency modulation AGC unit, defined as:

[0099]

[0100] Where Δpt,i(t) is the actual output power of the i-th heterogeneous frequency modulation AGC unit, Δp max,i and Δp min,i They are respectively the upper and lower limits of the regulation capability of the heterogeneous frequency modulation AGC unit.

[0101] Case Analysis

[0102] In order to verify the method of Example 1, in Simulink Figure 2 The power system shown in the figure contains five heterogeneous frequency-regulated AGC units and is built and simulated.

[0103] The detailed parameters of the five frequency modulation units are shown in Table 1. Figure 2 The system model parameters shown are set to K ps =0.01Hz / MW, T ps =20s, β=120MW / Hz. The control cycle of AGC is generally between 1 and 16s. The example analysis sets it to 2s, so that AGC updates the true value of the regional control error every 2s. The time interval of agent communication is set to 10ms.

[0104] Table 1 Parameters of various heterogeneous frequency modulation AGC units

[0105]

[0106] The relevant parameters of the distributed ACE mining algorithm are set to ρ1 = 20. For the designed sliding mode state observer and the designed sliding mode controller, their relevant parameters are shown in Table 2. The parameters of the distributed proportional load distribution algorithm are designed to be ρ4 = 0.2.

[0107] Table 2 Parameters of the proposed controller and observer

[0108]

[0109] Performance comparison of different AGC methods

[0110] In this case analysis, the proposed method is compared with the following two AGC methods: (1) a distributed PI controller is used instead of the proposed AGC controller without considering proportional power allocation, which is called DPI method 1; (2) based on DPI method 1, distributed proportional power allocation is considered, which is called DPI method 2.

[0111] Figure 3 This is a performance comparison chart of the AGC method proposed in the present invention and the other two methods, and the statistical results are shown in Table 3. Figure 4 The output and per-unit change curves of each unit under the distributed AGC method, DPI method 1 and DPI method 2 proposed in the present invention are shown in Table 3. In Table 3, |ACE| is the absolute value of the system ACE signal, |Δf| is the absolute value of the system frequency deviation, and CPS1 is the control performance standard (CPS) indicator used to evaluate the frequency control performance of the control area. p = 0.55 and k i =0.015, the participation factor of each unit is proportional to its capacity. Figure 4 As shown in (b)(d)(f) in the figure, the proposed method can distribute the power output of all units according to the capacity ratio. For DPI method 1, the fast response unit bears more AGC requirements due to its fast response capability. After the load changes from 0 to 30MW, the capacity of the fast response unit is not released. Therefore, in the second load change process from 30MW to 55MW, since there is no capacity reserved for the fast response unit, its fast response capability cannot be better released, which is not conducive to the next round of frequency adjustment. Figure 3 As shown in Table 3, the method proposed in the present invention is obviously superior to the other two methods in various performances.

[0112] Table 3 Statistical results of different AGC methods

[0113]

[0114] Example 2

[0115] This embodiment provides a distributed cooperative automatic power generation control system for heterogeneous frequency modulation units based on local linearization and sliding mode control, including:

[0116] The power system is composed of multiple heterogeneous frequency-regulated AGC units distributed;

[0117] Each heterogeneous frequency regulation AGC unit is described by the single-area power system AGC dynamic response model;

[0118] Each heterogeneous frequency modulation AGC unit is provided with a controller, wherein the controller includes a super local linear model state observer, an AGC controller based on sliding mode control and a distributed proportional load distribution algorithm controller;

[0119] The distributed cooperative automatic power generation control method of heterogeneous frequency modulation units based on local linearization and sliding mode control described in Example 1 is adopted to perform automatic power generation control.

[0120] The above contents are only for explaining the technical idea of ​​the present invention and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the claims of the present invention.

Claims

1. A method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control, which is used in a distributed system containing multiple heterogeneous frequency modulation AGC units, comprises the following steps: (1) A single-region power system AGC dynamic response model is used to describe each heterogeneous frequency regulation AGC unit; (2) The sum of known constants and unknown disturbances F of each heterogeneous frequency-regulated AGC unit in a single-region power system 1,i and regional control error ε i , respectively perform super local linear model state observations and obtain the corresponding estimated values and i represents the i-th heterogeneous frequency modulation AGC unit; Among them, the regional control error ε i The ACE signals of each heterogeneous frequency modulation AGC unit are obtained by distributed estimation of the ACE discovery algorithm; (3) Output power and estimated value of each heterogeneous frequency modulation AGC unit and estimated values Adopting AGC algorithm based on sliding mode control to control the regional control error of each heterogeneous frequency modulation AGC unit to zero; (4) In the final stage of frequency regulation, each heterogeneous frequency regulation AGC unit is assigned the frequency regulation task in proportion to the available capacity allocated by the set distributed proportional load allocation algorithm.

2. According to the method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control according to claim 1, the steps for establishing the AGC dynamic response model of the single-region power system are: 1-1) The first-order transfer function is used to represent the dynamic response model of the governor-turbine controller of each heterogeneous frequency modulation AGC unit: in, Subscript i is the unit number, T t,i is the time constant; 1-2) Define the regional control error ε(t): ε(t)=β·Δf(t) Among them, Δf is the system frequency deviation, β is the frequency deviation coefficient; 1-3) Primary control output power of each AGC frequency modulation unit Calculated as: Among them, R i represents the droop coefficient of the i-th AGC frequency regulation unit; 1-4) For the regional control error, the state equation is: Among them, Δp t,i represents the output power of the i-th unit, Δp d represents the power disturbance, T ps and K ps Represent the system time constant and gain constant respectively; The relationship between the regional control error and the output power of each frequency modulation unit is: Will With Δp d Considered as a disturbance.

3. According to the method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control according to claim 2, the method for performing super local linear model state observation is: 2-1) Design of AGC super local linear model for single-region power system: in, η1 and ψ1 are constant coefficients, F1 represents the sum of known constants and unknown disturbances in the single-region power system, and ξ1 is expressed as the differential of F1; 2-2) For the super local linear model, its state observer is designed as: in, and Represent the estimated values ​​of ε and F1, μ 1,i represents the sliding mode control function, λ1 represents the gain parameter; 2-3) The observation error of the state observer is: 2-4) The sliding surface of the state observer is: Among them, the sliding mode parameter α 1,i >0; 2-5) The sliding mode control function of the state observer is: The sliding mode parameter β 1,i , β 2,i >0, 2-6) The calculation formula is:

4. According to the method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control according to claim 3, the AGC algorithm based on sliding mode control is: 3-1) Output power of sliding mode controller The output power Δp of the heterogeneous frequency modulation AGC unit t,i The relationship is: in, The second item of the secondary frequency regulation output power of the AGC frequency regulation unit; 3-2) The sliding surface of the sliding mode controller is: S εi =∫εdt+α 2,i ·e Among them, the sliding mode parameter α 2,i >0; 3-3) Considering Δw as a disturbance, the AGC control law based on sliding mode control is designed as: Among them, the sliding mode parameter β 3,i , 5. According to the method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control according to claim 4, the distributed ACE discovery algorithm is expressed as: in, ε(t) is the actual ACE value of the system, and are the ACE estimation values ​​of the i-th and j-th heterogeneous frequency modulation AGC units respectively, and the gain coefficient ρ1>

0. If the i-th heterogeneous frequency modulation AGC unit is connected to the control center, then c i =1, otherwise c i =0; Through the distributed ACE discovery algorithm, the ACE signals of each heterogeneous frequency modulation AGC unit are estimated in a distributed manner to obtain the ACE estimation value to replace the actual ACE value ε(t), thereby avoiding centralized communication.

6. According to the method for distributed coordinated automatic power generation control of heterogeneous frequency modulation units based on local linearization and sliding mode control according to claim 5, the distributed proportional load distribution algorithm is: The second term of the secondary frequency modulation output power of the i-th heterogeneous frequency modulation AGC unit for: in, Gain coefficient ρ4>0; w i,j is an element of the symmetric matrix W, defined as: In the formula, the subscripts i and j represent the AGC unit number, n i 、n j Indicates the number of adjacent AGC units, γ>0; Δp pu,i (t) is the per-unit value of the actual output power of the i-th AGC unit, defined as: Where, Δpt,i(t) is the actual output power of the i-th AGC unit, Δp max,i and Δp min,i They are respectively the upper and lower limits of the AGC unit's adjustment capability.

7. A distributed cooperative automatic power generation control system for heterogeneous frequency-regulated units based on local linearization and sliding mode control, comprising: The power system consists of multiple distributed heterogeneous frequency-regulated AGC units; Each heterogeneous frequency regulation AGC unit is described by the single-area power system AGC dynamic response model; Each heterogeneous frequency modulation AGC unit is provided with a controller, wherein the controller includes a super local linear model state observer, an AGC controller based on sliding mode control and a distributed proportional load distribution algorithm controller; Automatic power generation control is performed by adopting the distributed collaborative automatic power generation control method of heterogeneous frequency modulation units based on local linearization and sliding mode control as described in any one of claims 1 to 6.

Images