MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation

Abstract

The invention discloses an MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation, and belongs to the technical field of automatic control. The method mainly solves the problem that frequency fluctuation is affected by various uncertain factors, complex unknown disturbance and the like in micro-grid (MGs) system frequency control. The control scheme comprises the following steps: designing an adaptive model prediction controller; designing a model-free adaptive controller based on the input and output data; and establishing a model-free adaptive control compensation MGs system adaptive model prediction frequency control scheme. According to the MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation disclosed by the invention, better control precision and stronger anti-interference capability can be realized when the frequency of a micro-grid system is influenced by unknown disturbance, and the precision requirement of micro-grid frequency control is met.

Description

Technical Field

[0001] The present invention belongs to the field of automatic control technology, and more specifically, relates to an adaptive model predictive frequency control (AMPC) method of an MGs system with model-free adaptive control (MFAC) compensation. Background Art

[0002] In recent years, with the rapid development of microgrids (MGs) mainly composed of renewable energy such as wind power generation and solar power generation, higher control requirements have been put forward for the frequency control of increasingly complex microgrids to ensure high-quality power supply.

[0003] For the research on microgrid frequency control, the following two issues need to be considered: 1. There are constantly changing unknown disturbances and power load consumption in the microgrid. The load frequency controller needs to effectively handle the frequency changes of the microgrid system caused by unknown disturbances and power load changes to stabilize the frequency control process; 2. The load frequency controller improves the accuracy of control error tracking while maintaining stable operation.

[0004] At present, in order to respond to the carbon neutral development goal and achieve the purpose of sustainable development, renewable energy with solar power generation (PV) and wind power generation (WT) as the main forms has been rapidly developed. Various new, discrete and complex microgrid systems with renewable energy as the main form of power generation have emerged to meet the growing power demand. At the same time, due to the various changing power loads and unknown factors in the microgrid system that seriously affect the microgrid frequency, higher requirements are also put forward for the frequency control of the microgrid. Therefore, for the load frequency controller of the microgrid, it should have higher control accuracy, greater flexibility and elasticity to meet the complex and changeable frequency control requirements of the microgrid system. In order to improve the frequency control effect of the microgrid system, domestic and foreign scholars have done a lot of research, such as proportional integral differential (PID) frequency control based on neural network learning parameter setting, PID load frequency control strategy based on optimization algorithm parameter setting, model prediction frequency control algorithm based on optimization algorithm parameter setting, robust load frequency control algorithm, etc. Most of these methods are based on operator experience and require a large amount of data to train the neural network, which challenges the real-time and reliability requirements of the controller. Therefore, how to design a control method with strong robustness, high control accuracy, high flexibility and elasticity to solve the frequency control in increasingly complex microgrids is an urgent problem to be solved in this field. Summary of the invention

[0005] The present invention discloses an MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation. The technical problem to be solved is to ensure that the frequency control of the increasingly complex and strongly nonlinear microgrid system can achieve high-precision control while enhancing the anti-interference ability of the controller to make it have better robustness, flexibility and elasticity.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] The present invention discloses an adaptive model prediction frequency control method for an MGs system based on model-free adaptive control compensation, which is mainly aimed at the problems of unknown disturbances in complex and strongly nonlinear microgrid systems and the impact of power load consumption on system frequency. It has good control accuracy and strong anti-interference ability, and meets the requirements of frequency control accuracy of the microgrid system.

[0008] The present invention discloses an MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation, the method comprising the following steps:

[0009] Step 1: Design an adaptive model predictive controller, which mainly includes the following steps:

[0010] Step 1.1: Design the following state-space equations for the internal model of model predictive control:

[0011]

[0012] in,

[0013] x AMPC (t) = Δf(t) is the frequency deviation of the microgrid at time t; y AMPC (t) is the microgrid regional control error predicted at time t; A, B, C are the internal model parameters of model predictive control; u AMPC (t) is the power generation control signal of the model predictive control at time t;

[0014] Step 1.2: Define Δx AMPC (t) = x AMPC (t)-x AMPC (t-1), forming the extended state space equation of model predictive control:

[0015]

[0016] in,

[0017] x -AMPC (t) = [Δx AMPC (t) y AMPC (t)] T is the expansion state quantity at time t; is the expansion matrix; Δu AMPC (t) is the power generation control signal deviation of the model predictive control at time t;

[0018] Step 1.3: Design the control law for model predictive control:

[0019] u AMPC (t) = u AMPC (t-1)+G AMPC Δu AMPC (t);

[0020] in,

[0021] G AMPC =-(Φ T Φ+S) -1 Φ T F is the feedback gain of model predictive control;

[0022]

[0023] Np is the prediction interval of model predictive control; Nc is the control interval of model predictive control; S is the weight;

[0024] Step 1.4: Design the Unscented Kalman Filter (UKF) to predict and update the parameters A and B of the internal model:

[0025] Define x UKF (t) = [y(t)θ(t)] T ,θ(t)=[A(t)B(t)],θ 1 (t) = A(t), θ 2 (t) = B(t), the following UKF state space equation is designed for prediction update:

[0026]

[0027] in,

[0028] x UKF (t) is the state quantity of UKF at time t; y UKF (t) is the predicted output of UKF at time t; I is the unit matrix.

[0029] Step 2: Design a model-free adaptive controller based on input and output data, characterized in that: design a model-free adaptive controller algorithm:

[0030]

[0031] in,

[0032] λ is a positive weight coefficient; ρ is a step size factor greater than 0; y * (t+1) is the expected regional control error of the microgrid system at time t+1; y(t) represents the regional control error of the microgrid system at time t; u MFAC (t) represents the power generation control signal of the model-free adaptive controller at time t; u MFAC (t-1) represents the power generation control signal of the model-free adaptive controller at time t-1; φ(t) is the defined pseudo partial derivative.

[0033] Step 3: Establishing an adaptive model prediction frequency control scheme for the MGs system without model adaptive control compensation, mainly including the following steps:

[0034] Establish the total power generation control signal of the microgrid system:

[0035] u(t)=u AMPC (t)+u MFAC (t);

[0036] in,

[0037] u(t) is the power generation control signal of the microgrid system at time t.

[0038] Beneficial effects:

[0039] The present invention discloses an adaptive model prediction frequency control method for an MGs system based on model-free adaptive control compensation. The model-free adaptive control algorithm is used for the accuracy compensation of the adaptive model predictive control, and a dual-loop compensation control scheme based on model-free predictive control and adaptive model predictive control is constructed, which improves the control accuracy and flexibility of the control algorithm, and has good robustness, meeting the requirements of the microgrid system frequency control on control accuracy and robustness.

[0040] After reading the specific embodiments of the present invention in conjunction with the accompanying drawings, other features and advantages of the present invention will become more clear. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of the adaptive model prediction frequency control of the MGs system based on model-free adaptive control compensation proposed by the present invention;

[0042] Figure 2 This is a simplified structural diagram of the working principle of the adaptive model prediction frequency control of the MGs system based on model-free adaptive control compensation proposed by the present invention;

[0043] Figure 3 This is a structural diagram of a microgrid according to an embodiment of the present invention;

[0044] Figure 4This is a frequency control performance diagram of the MGs system adaptive model predictive frequency control based on model-free adaptive control compensation proposed in the present invention;

[0045] Figure 5 It is the frequency control performance diagram based on the single-loop adaptive model predictive control method without compensation;

[0046] Figure 6 It is the frequency control performance diagram based on the single-loop proportional integral derivative control method without compensation; DETAILED DESCRIPTION

[0047] In order to better illustrate the purpose and advantages of the present invention, a comparison is made below with a single-loop proportional-integral-differential and adaptive model predictive control method based on no compensation, and the content of the invention is further described in detail in conjunction with the corresponding drawings.

[0048] The present invention has good control accuracy and strong anti-interference ability under the frequency control of the microgrid system, meeting the requirements of the frequency control accuracy of the microgrid system. An adaptive model prediction frequency control method for MGs system based on model-free adaptive control compensation is proposed. Below, a specific implementation method of an adaptive model prediction frequency control method for MGs system based on model-free adaptive control compensation is described in detail.

[0049] See also Figure 1 As shown, the present embodiment discloses an MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation, comprising the following steps:

[0050] Step S1: Design an adaptive model predictive controller.

[0051] Step S11: Design the following first-order state space equation for the internal model of model predictive control:

[0052]

[0053] in,

[0054] x AMPC (t) = Δf(t) is the frequency deviation of the microgrid at time t; y AMPC (t) is the microgrid regional control error predicted at time t; A, B, C are the internal model parameters of model predictive control; u AMPC (t) is the power generation control signal of the model predictive control at time t;

[0055] Define Δx AMPC (t) = x AMPC (t)-x AMPC (t-1), forming the extended state space equation of model predictive control:

[0056]

[0057] in,

[0058] x -AMPC (t) = [Δx AMPC (t) y AMPC (t)] T is the expansion state quantity at time t; is the expansion matrix; Δu AMPC (t) is the power generation control signal deviation of the model predictive control at time t;

[0059] The state quantity is solved by iterative prediction of the expanded state equation:

[0060]

[0061] The output is solved by iterative prediction of the expanded state equation:

[0062]

[0063] in,

[0064] Np is the prediction interval of model predictive control; Nc is the control interval of model predictive control;

[0065] Step S12: Design the objective function of model predictive control:

[0066]

[0067] in,

[0068] h is the time interval; Q and S are weights;

[0069] Step S13: Design the control law of model predictive control:

[0070] u AMPC (t) = u AMPC (t-1)+G AMPC Δu AMPC (t);

[0071] in,

[0072] G AMPC =-(Φ T Φ+S) -1 Φ T F is the feedback gain of model predictive control;

[0073]

[0074] Step S14: Design an unscented Kalman filter (UKF) to predict and update the parameters A and B of the internal model:

[0075] Define x UKF (t) = [y(t)θ(t)] T ,θ(t)=[A(t)B(t)],θ 1 (t) = A(t), θ 2 (t) = B(t), the following UKF state space equation is designed for prediction update:

[0076]

[0077] in,

[0078] is the state quantity of UKF at time t; y UKF (t) is the predicted output of UKF at time t; I is the unit matrix.

[0079] Step S15: UKF generates sample points.

[0080]

[0081]

[0082] in,

[0083] x UKFt is the sample point at time t; UKF is the scaling parameter; n is the dimension; P is the covariance; W is the weight factor;

[0084] Step S16: The sample points are transferred through the state equation.

[0085]

[0086] in,

[0087] v(·) is the state equation of UKF in step S14;

[0088] Step S17: Calculate the prior covariance of the predicted state quantity at time t.

[0089]

[0090] The sample points are passed through the output state equation designed in step S14:

[0091]

[0092] in,

[0093] z(·) is the predicted output state equation of UKF in step S14;

[0094] Step S18: Calculate the prior covariance of the predicted output at time t.

[0095]

[0096]

[0097] Step S19: Calculate the UKF gain value based on the prior covariance.

[0098]

[0099] in,

[0100] σ is the system noise; g t+1 is the UKF gain at time t+1;

[0101] Step S110: Calculate and update the UKF prediction state quantity.

[0102]

[0103] Step S2: Design a model-free adaptive controller based on input and output data.

[0104] Step S21: Establishing a nonlinear mathematical relationship between the input and output data of the controlled system.

[0105] The actual microgrid system is a nonlinear system with time-varying parameters, multiple variables, and unknown disturbances, which can be expressed as the following general nonlinear function relationship between input and output:

[0106] y(t+1)=f(y(t),u AMPC (t),u MFAC (t),w(t));

[0107] in,

[0108] f(·) represents a nonlinear function; y(t+1) represents the regional control error of the microgrid system at time t+1; y(t) represents the regional control error of the microgrid system at time t; u AMPC (t) represents the power generation control signal of the adaptive model predictive controller at time t; u MFAC (t) represents the power generation control signal of the model-free adaptive controller at time t; w(t) represents the external disturbance of the microgrid at time t;

[0109] Step S22: converting the nonlinear system into an equivalent dynamic linearization data model.

[0110] Specifically, the general nonlinear system in step S21 satisfies:

[0111] The input and output of the system are observable and controllable, that is, for a bounded expected value y * (t+1), in the presence of unknown disturbances, there exists a bounded control input signal so that under the action of the control input signal, the output of the system is equal to the expected given signal.

[0112] The system satisfies the generalized Lipschitz condition, namely:

[0113] |f(x 1 (t 1 ),...,x n (t 1 ))-f(x 1 (t 2 ),...,x n (t 2 ))|≤b 1 |x 1 (t 1 )-x1(t 2 )|+...+b n |x n (t 1 )-x n (t 2 )|;

[0114] in,

[0115] n∈Z + ;x z is the parameter of the nonlinear function f(·); b z >0 is a positive constant; z∈{1,...,n}; t 1 ,t 2 For different moments;

[0116] Step S23: For the general nonlinear system in step S21, when |Δu MFAC When (t)|≠0, there exists a pseudo partial derivative φ(t), which transforms the system into:

[0117] Δy(t+1)=φ(t)Δu MFAC (t);

[0118] in,

[0119] φ(t) is a pseudo partial derivative by definition; |φ(t)|≤b.

[0120] Step S24: See Figure 2 As shown, a model-free adaptive controller is designed based on input and output data.

[0121] First consider the following input estimation criteria:

[0122] J(u MFAC(t))=|y * (t+1)-y(t+1)| 2 +λ|u MFAC (t)-u MFAC (t-1)| 2 ;

[0123] Solving equations The model-free adaptive control law algorithm can be obtained:

[0124]

[0125] in,

[0126] λ is a positive weight coefficient; ρ is a step size factor greater than 0, which is introduced into the control algorithm to make it more general; y * (t+1) is the expected regional control error of the microgrid system.

[0127] In order to make the algorithm more capable of tracking time-varying systems, a reset algorithm is added:

[0128] φ(t)=φ(1),φ(t)≤ε or |Δu MFAC (t-1)|≤ε;

[0129] in,

[0130] ε is a small positive number; φ(1) is the initial value of the pseudo partial derivative.

[0131] Step S3: Establishing an adaptive model prediction frequency control scheme for the MGs system without model adaptive control compensation.

[0132] Establish the total power generation control signal of the microgrid system:

[0133] u(t)=u AMPC (t)+u MFAC (t);

[0134] in,

[0135] u(t) is the power generation control signal of the microgrid system at time t.

[0136] In the embodiment, a microgrid system consisting of a diesel generator, a speed regulator, renewable energy, and power load consumption is used. Figure 3 shown.

[0137] For comparison, a single-loop model predictive control method without compensation ( Figure 5 ) and the single-loop proportional-integral-derivative control method without compensation ( Figure 6 ) adopts the same microgrid environment as above.

[0138] Depend on Figure 4 and Figure 5 It can be seen that: in terms of frequency control performance, the control method proposed by the present invention is compared with the adaptive model predictive control method based on a single loop without compensation. At 500 seconds, the overshoot of the control method proposed by the present invention is 0.0089 Hz, and the overshoot of the adaptive model predictive control method based on a single loop without compensation is 0.0218 Hz. Compared with the adaptive model predictive control method of a single loop without compensation, the overshoot performance is improved by about 59.1%. Similarly, at 800 seconds, the overshoot performance of the method proposed by the present invention is improved by about 66.1%. Comparison shows that the tracking performance under the control algorithm proposed by the present invention is significantly improved compared with the adaptive model predictive control method based on a single loop without compensation.

[0139] Depend on Figure 4 and Figure 6 It can be seen that the control method proposed in the present invention can also significantly reduce the overshoot and oscillation amplitude compared with the single-loop proportional-integral-differential control method based on no compensation, and has a significant improvement in control performance.

[0140] This embodiment mainly uses the frequency deviation output of the microgrid as a performance parameter. Figure 4 The frequency deviation curve of the MGs system adaptive model prediction frequency control method based on model-free adaptive control compensation proposed by the present invention is given. It can be seen from the figure that compared with the model prediction control based on a single loop without compensation and the proportional integral differential control method based on a single loop without compensation, the control method provided by the present invention has better frequency control effect and smaller overshoot.

[0141] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the embodiments are only used to explain the invention and are not used to limit the scope of protection of the invention. Any modifications, equivalent replacements, improvements, etc. under the same principles and concepts as the invention should be included in the scope of protection of the invention.

Claims

1. An adaptive model prediction frequency control method for MGs system based on model-free adaptive control compensation, characterized in that: The method comprises the following steps: Step 1: Design an adaptive model predictive controller; Step 2: Design a model-free adaptive controller based on input and output data; Step 3: Establish an adaptive model prediction frequency control scheme for the MGs system without model adaptive control compensation.

2. The method according to claim 1, characterized in that: The design of the adaptive model predictive controller in step 1 mainly includes the following steps: Step 1.1: Design the following state-space equations for the internal model of model predictive control: in, x AMPC (t) = Δf(t) is the frequency deviation of the microgrid at time t; y AMPC (t) is the microgrid regional control error predicted at time t; A, B, C are the internal model parameters of model predictive control; u AMPC (t) is the power generation control signal of the model predictive control at time t; Step 1.2: Define Δx AMPC (t) = x AMPC (t)-x AMPC (t-1), forming the extended state space equation of model predictive control: in, x -AMPC (t) = [Δx AMPC (t)y AMPC (t)] T is the expansion state quantity at time t; is the expansion matrix; Δu AMPC (t) is the power generation control signal deviation of the model predictive control at time t; Step 1.3: Design the control law for model predictive control: u AMPC (t)=u AMPC (t-1)+G AMPC Δu AMPC (t); in, G AMPC =-(Φ T Φ+S) -1 Φ T F is the feedback gain of model predictive control; Np is the prediction interval of model predictive control; Nc is the control interval of model predictive control; S is the weight; Step 1.4: Design the Unscented Kalman Filter (UKF) to predict and update the parameters A and B of the internal model: Define x UKF (t) = [y(t)θ(t)] T , θ(t)=[A(t)B(t)], θ1(t)=A(t), θ2(t)=B(t), and design the following UKF state space equation for prediction update: in, x UKF (t) is the state quantity of UKF at time t; y UKF (t) is the predicted output of UKF at time t; I is the unit matrix.

3. The method according to claim 1, characterized in that: In step 2, a model-free adaptive controller is designed based on input and output data, which mainly includes the following steps: in, λ is a positive weight coefficient; ρ is a step size factor greater than 0; y * (t+1) is the expected regional control error of the microgrid system at time t+1; y(t) represents the regional control error of the microgrid system at time t; u MFAC (t) represents the power generation control signal of the model-free adaptive controller at time t; u MFAC (t-1) represents the power generation control signal of the model-free adaptive controller at time t-1; φ(t) is the defined pseudo partial derivative.

4. The method according to claim 1, characterized in that: In step 3, an adaptive model prediction frequency control scheme for the MGs system without model adaptive control compensation is established, which mainly includes the following steps: Establish the total power generation control signal of the microgrid system: u(t)=u AMPC (t)+u MFAC (t); in, u(t) is the power generation control signal of the microgrid system at time t.

Images