A predictive control method, device and storage medium for a hybrid energy storage system

Abstract

The present invention relates to the field of hybrid energy storage system control technology, and in particular to a hybrid energy storage system predictive control method, device, equipment and computer storage medium. The present invention comprises: the hybrid energy storage system predictive control method of the present invention, which adopts the hybrid energy storage system state space model obtained by the subspace identification algorithm as the prediction model. Compared with the traditional topology-based model, its design reduces the sensitivity of the model predictive control to parameters and has higher accuracy; the present invention adopts a combination of offline identification and online identification. Compared with the traditional offline identification, its design takes into account the changes in system parameters and better simulates the actual working scenario. When the system parameters suddenly change, robust model predictive control is adopted to suppress the influence of parameter changes on control performance, thereby improving the robustness of the system.

Description

Technical Field

[0001] The present invention relates to the field of hybrid energy storage system control technology, and in particular to a hybrid energy storage system predictive control method, device, equipment and computer storage medium. Background Art

[0002] A hybrid energy storage system is a combination of at least two energy storage systems. Its primary goal is to optimally utilize the advantages of available energy storage. A common combination is batteries and supercapacitors, due to their complementary energy and power characteristics. Batteries have high energy density and provide long-term energy buffering; correspondingly, supercapacitors have high power density and are suitable for rapid power response. This combination can extend battery life without sacrificing overall system performance. Therefore, hybrid energy storage systems composed of batteries and supercapacitors have found widespread application in fields such as electric vehicles and renewable energy generation. To ensure good dynamic performance in hybrid energy storage systems, various control strategies have been proposed. Previous work can be broadly divided into two categories: rule-based and optimization-based. Optimization-based control strategies can be categorized into offline and online methods, with model predictive control (MPC) being a prominent online optimization management strategy. MPC is an advanced model-based control method whose performance depends significantly on the parameters used and the accuracy of the constructed model. However, the multivariable, strongly coupled, and nonlinear characteristics of hybrid energy storage systems make obtaining accurate models challenging and challenging. Furthermore, when system parameters change, model mismatch can occur, compromising control performance. Summary of the Invention

[0003] To this end, the technical problem to be solved by the present invention is to overcome the problem in the prior art that it is difficult to accurately construct a prediction model for the hybrid energy storage system model predictive control and there is a model mismatch problem when the system parameters suddenly change, resulting in poor control performance.

[0004] To solve the above technical problems, the present invention provides a predictive control method for a hybrid energy storage system, comprising:

[0005] Based on offline input and output data, a state space model of the hybrid energy storage system is established based on subspace identification and used as a prediction model for the predictive control strategy.

[0006] When the system parameters are fixed, a model predictive control strategy is used to control the operation of the hybrid energy storage system. The model predictive control strategy is as follows: taking minimizing cost as the optimization goal, establishing a first objective function and a first constraint, measuring the current system state and obtaining an optimal solution, and applying it to the system;

[0007] When the system parameters suddenly change, the robust model predictive control strategy is switched to control the operation of the hybrid energy storage system. The online input and output data generated in real time during the control process are collected, and the online state space model of the hybrid energy storage system at the corresponding moment is obtained using subspace identification technology. The robust predictive control strategy takes minimizing cost as the optimization goal and establishes H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system.

[0008] When the preset conditions are met, the robust model predictive control strategy is switched back to the model predictive control strategy, and the online state space model at this moment is used as the prediction model of the model predictive control. The preset condition is that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than the preset threshold.

[0009] Preferably, the establishing of the first objective function and the first constraint with minimizing cost as the optimization goal includes:

[0010] Building a cost function

[0011] Among them, the first part is the cost term of the state variable deviation, the second part is the cost term of the input variable deviation, Q s >0 and R s >0 is a symmetric weighted matrix, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, and u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k;

[0012] The first objective function is established with minimizing the cost function as the optimization goal:

[0013]

[0014] stx(k+i+1|k)=Ax(k+i|k)+Bu(k+i|k)

[0015]

[0016] in, is a constant matrix, is the maximum value of each input, m is the number of columns of matrix B;

[0017] Obtain the upper bound of the cost function and minimize the upper bound using the state feedback control law, converting the first objective function into finding Q, F, and Z to minimize γ:

[0018]

[0019] Among them, γ is a positive scalar, Q, F is a matrix, and Z is a symmetric matrix;

[0020] The first constraint is:

[0021]

[0022]

[0023]

[0024] Among them, x(k|k) represents the state prediction corresponding to the time instant k based on time k, I is a unit matrix, Y is a matrix, the symbol * represents a square in a symmetrical position, and Z jj is the j-th diagonal element of the matrix Z.

[0025] Preferably, the optimization goal is to minimize the cost and establish H ∞ The second objective function and the second constraint of the type include:

[0026] Build H ∞ Type cost function:

[0027] Among them, the first part is the cost term of the state variable deviation, and the second part is the cost term of the input variable deviation. Q s >0 and R s >0 is a symmetric weight matrix, μ is a positive scalar, is the external interference, q is the dimension of the interference vector, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k, and d(k+i|k) represents the interference variable corresponding to the time instant k+i based on time k;

[0028] To minimize H ∞ The cost function is used to establish the second objective function for the optimization goal:

[0029]

[0030] Obtain the H ∞ The upper bound of the cost function is obtained by using the state feedback control law to minimize the upper bound, and the second objective function is converted into finding Q, F, γ and λ so that Minimum:

[0031]

[0032] Among them, γ, λ is a positive scalar, Q, F is a matrix;

[0033] The second constraint is:

[0034]

[0035]

[0036] γ≤γ0

[0037]

[0038]

[0039] Among them, x(k) is the state variable at the current moment, ζ,ξ are positive scalars, γ0 is the upper limit of γ, X is a matrix, and I is an identity matrix.

[0040] Preferably, the state feedback control law is to transmit the state variable of the system to the input end through a proportional link for feedback, and its expression is:

[0041] u(k+i∣k)=Kx(k+i∣k),i≥0

[0042] Where K is the control gain.

[0043] Preferably, the step of establishing a hybrid energy storage system state space model based on subspace identification according to offline input and output data and using it as a prediction model for the predictive control strategy includes:

[0044] Get offline input and output data:

[0045] u=(u(0),u(1),...,u(s+N-2))

[0046] y=(y(0),y(1),...,y(s+N-2))

[0047] Where u(s+N-2) is the s+N-2th input data, y(s+N-2) is the s+N-2th output data, s is a constant strictly greater than the dimension n of the state vector, and N is the number of columns of the Hankel matrix;

[0048] Construct the Hankel matrix based on the offline input and output data:

[0049]

[0050]

[0051] Among them, m is the dimension of the input variable, and p is the dimension of the output variable;

[0052] The Hankel matrix is ​​converted into a block lower triangular matrix with 0 in the upper right corner by LQ decomposition technique:

[0053]

[0054] Among them, L 11 ,L 22 is a lower triangular matrix, L 21 is a general matrix, is an orthogonal matrix;

[0055] The lower triangular matrix Converted to:

[0056]

[0057] in, is a matrix, Σ1 is a diagonal matrix, V1, V2 are matrices;

[0058] According to the transformed lower triangular matrix Get the system estimated extended observation matrix

[0059] According to the extended observation matrix, the system matrices A, B, C, and D are calculated by the subspace identification algorithm to obtain the state space model of the hybrid energy storage system based on subspace identification:

[0060] x(k+1)=Ax(k)+Bu(k)

[0061] y(k)=Cx(k)+Du(k), k=0,1…

[0062] Among them, u(k) and y(k) are the input and output of the current moment respectively, x(k+1) is the input of the next moment, and k represents the kth moment.

[0063] Preferably, the online input and output data are collected to update the state space model:

[0064] x(k+1)=A(k)x(k)+B(k)u(k)

[0065] Among them, A(k) and B(k) are real-time system matrices obtained based on online identification technology, x(k) is the state variable at the current moment, and u(k) is the input at the current moment.

[0066] Preferably, the preset condition is:

[0067]

[0068] in, is the load voltage controlled by the online state space model, and x5 is the actual load voltage of the system.

[0069] The present invention also provides a hybrid energy storage system predictive control device, comprising:

[0070] The prediction model building module is used to establish a state space model of the hybrid energy storage system based on subspace identification according to offline input and output data, and use it as a prediction model for the predictive control strategy;

[0071] A model predictive control module is configured to control the operation of the hybrid energy storage system using a model predictive control strategy when system parameters are fixed. The model predictive control strategy is to establish a first objective function and a first constraint with minimizing cost as the optimization goal, measure the current system state, and obtain an optimal solution, which is then applied to the system.

[0072] The robust model predictive control module is used to switch to the robust model predictive control strategy to control the operation of the hybrid energy storage system when the system parameters suddenly change. It collects the online input and output data generated in real time during the control process and uses the subspace identification technology to obtain the online state space model of the hybrid energy storage system at the corresponding moment. The robust predictive control strategy is to minimize the cost as the optimization goal and establish H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system.

[0073] The predictive control switching module is used to switch from the robust model predictive control strategy back to the model predictive control strategy when a preset condition is met, and at the same time use the online state space model at this moment as the prediction model of the model predictive control. The preset condition is that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than a preset threshold.

[0074] The present invention also provides a hybrid energy storage system predictive control device, comprising:

[0075] Memory for storing computer programs;

[0076] A processor is used to implement the steps of the above-mentioned hybrid energy storage system predictive control method when executing the computer program.

[0077] The present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned predictive control method for a hybrid energy storage system are implemented.

[0078] The above technical solution of the present invention has the following advantages over the prior art:

[0079] The hybrid energy storage system predictive control method described in the present invention adopts the hybrid energy storage system state space model obtained by the subspace identification algorithm as the prediction model. Compared with the traditional topology-based model, its design reduces the sensitivity of model predictive control to parameters and has higher accuracy; the present invention adopts a combination of offline identification and online identification. Compared with traditional offline identification, its design takes into account the changes in system parameters and better simulates the actual working scenario. When the system parameters suddenly change, robust model predictive control is used to suppress the impact of parameter changes on control performance, thereby improving the robustness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0081] Figure 1 This is a flow chart of the implementation of a predictive control method for a hybrid energy storage system of the present invention.

[0082] Figure 2 This is a flow chart for implementing a predictive control method for a hybrid energy storage system provided by an embodiment of the present invention.

[0083] Figure 3 This is a structural diagram of a predictive control method for a hybrid energy storage system provided by an embodiment of the present invention.

[0084] Figure 4 This is a flow chart of the state space model of the hybrid energy storage system established in the present invention.

[0085] Figure 5 A structural block diagram of a hybrid energy storage system predictive control device of the present invention. DETAILED DESCRIPTION

[0086] The core of the present invention is to provide a hybrid energy storage system predictive control method, device, equipment and computer storage medium, which improves the system predictive control performance.

[0087] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0088] Please refer to Figure 1 , Figure 2 and Figure 3 , Figure 1 This is a flow chart of an implementation of a predictive control method for a hybrid energy storage system provided by the present invention. Figure 2This is a flowchart of a predictive control method for a hybrid energy storage system according to an embodiment of the present invention. Figure 3 This is a structural diagram of a predictive control method for a hybrid energy storage system provided by an embodiment of the present invention; the specific operating steps are as follows:

[0089] S101: Based on offline input and output data, a state space model of the hybrid energy storage system based on subspace identification is established, and used as a prediction model for the predictive control strategy;

[0090] The specific process is as follows Figure 4 As shown:

[0091] Get offline input and output data:

[0092] u=(u(0),u(1),...,u(s+N-2))

[0093] y=(y(0),y(1),...,y(s+N-2))

[0094] Where u(s+N-2) is the s+N-2th input data, y(s+N-2) is the s+N-2th output data, s is a constant strictly greater than the dimension n of the state vector, and N is the number of columns of the Hankel matrix;

[0095] Construct the Hankel matrix based on the offline input and output data:

[0096]

[0097]

[0098] Among them, m is the dimension of the input variable, and p is the dimension of the output variable;

[0099] The Hankel matrix is ​​converted into a block lower triangular matrix with 0 in the upper right corner by LQ decomposition technique:

[0100]

[0101] Among them, L 11 ,L 22 is a lower triangular matrix, L 21 is a general matrix, is an orthogonal matrix;

[0102] The lower triangular matrix Converted to:

[0103]

[0104] in, is a matrix, Σ1 is a diagonal matrix, V1, V2 are matrices;

[0105] According to the transformed lower triangular matrix Get the system estimated extended observation matrix

[0106] According to the extended observation matrix, the system matrices A, B, C, and D are calculated by the subspace identification algorithm to obtain the state space model of the hybrid energy storage system based on subspace identification:

[0107] x(k+1)=Ax(k)+Bu(k)

[0108] y(k)=Cx(k)+Du(k), k=0,1…

[0109] Among them, x(k) is the state variable at the current moment, u(k) is the input at the current moment, y(k) is the output at the current moment, and k represents the kth moment.

[0110] S102: When system parameters are fixed, a model predictive control strategy is used to control the operation of the hybrid energy storage system. The model predictive control strategy is as follows: minimizing cost as an optimization goal, establishing a first objective function and a first constraint, measuring the current system state and obtaining an optimal solution, and applying the optimal solution to the system;

[0111] S103: When the system parameters suddenly change, the robust model predictive control strategy is switched to control the operation of the hybrid energy storage system, the online input and output data generated in real time during the control process are collected, and the online state space model of the hybrid energy storage system at the corresponding moment is obtained using the subspace identification technology. The robust predictive control strategy is to minimize the cost as the optimization goal, and establish H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system.

[0112] The parameter mutation includes a load resistance mutation;

[0113] The input and output data sets generated in real time during the acquisition control process are used to obtain the online state space expression using subspace identification technology:

[0114] x(k+1)=A(k)x(k)+B(k)u(k)

[0115] Among them, A(k) and B(k) are real-time system matrices obtained based on online identification technology, x(k) is the state variable at the current moment, and u(k) is the input at the current moment.

[0116] S104: When the preset conditions are met, the robust model predictive control strategy is switched back to the model predictive control strategy, and the online state space model at this moment is used as the prediction model of the model predictive control. The preset conditions are that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than the preset threshold.

[0117] The preset conditions are:

[0118]

[0119] in, is the load voltage controlled by the online state space model, and x5 is the actual load voltage of the system.

[0120] The hybrid energy storage system predictive control method described in the present invention adopts the hybrid energy storage system state space model obtained by the subspace identification algorithm as the prediction model. Compared with the traditional topology-based model, its design reduces the sensitivity of model predictive control to parameters and has higher accuracy; the present invention adopts a combination of offline identification and online identification. Compared with traditional offline identification, its design takes into account the changes in system parameters and better simulates the actual working scenario. When the system parameters suddenly change, robust model predictive control is used to suppress the impact of parameter changes on control performance, thereby improving the robustness of the system.

[0121] Based on the above embodiments, this embodiment designs a control algorithm based on predictive control according to control requirements, including a model predictive control and a robust model predictive control. Robust model predictive control is used to suppress the impact of parameter changes on control performance and improve system robustness:

[0122] The design concept of the model predictive control is to measure the current system state and solve the optimization problem to obtain the optimal solution of the cost function, and then apply the optimal solution to the system. Specifically,

[0123] Taking minimizing cost as the optimization goal, establishing the first objective function and the first constraint includes:

[0124] Building a cost function

[0125] The cost function is in the form of a sum of positive definite functions of infinite time, and consists of two parts. The first part is the cost term of the state variable deviation, and the second part is the cost term of the input variable deviation. Q s >0 and R s >0 is a symmetric weighted matrix, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, and u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k;

[0126] The first objective function is established with minimizing the cost function as the optimization goal:

[0127]

[0128] stx(k+i+1|k)=Ax(k+i|k)+Bu(k+i|k)

[0129]

[0130] in, is a constant matrix, is the maximum value of each input, m is the number of columns of matrix B;

[0131] Obtain the upper bound of the cost function and minimize the upper bound using the state feedback control law, converting the first objective function into finding Q, F, and Z to minimize γ:

[0132]

[0133] Among them, γ is a positive scalar, Q, F is a matrix, and Z is a symmetric matrix;

[0134] The first constraint is:

[0135]

[0136]

[0137]

[0138] Among them, x(k|k) represents the state prediction corresponding to the time instant k based on time k, I is a unit matrix, Y is a matrix, the symbol * represents a square in a symmetrical position, and Z jj is the j-th diagonal element of the matrix Z.

[0139] The design idea of ​​the robust model predictive control is to combine the advantages of the three original characteristics of predictive control with robust control and further optimize the model predictive control strategy.

[0140] Specifically, it uses H ∞ Type cost function, by measuring the state to solve the optimal problem, specifically:

[0141] Taking minimizing cost as the optimization goal, establish H ∞ The second objective function and the second constraint of the type include:

[0142] Build H ∞ Type cost function:

[0143] Among them, the first part is the cost term of the state variable deviation, and the second part is the cost term of the input variable deviation. Q s >0 and R s >0 is a symmetric weight matrix, μ is a positive scalar, is the external interference, q is the dimension of the interference vector, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k, and d(k+i|k) represents the interference variable corresponding to the time instant k+i based on time k;

[0144] To minimize H ∞ The cost function is used to establish the second objective function for the optimization goal:

[0145]

[0146] Obtain the H ∞ The upper bound of the cost function is obtained by using the state feedback control law to minimize the upper bound, and the second objective function is converted into finding Q, F, γ and λ so that Minimum:

[0147]

[0148] Among them, γ, λ is a positive scalar, Q, F is a matrix;

[0149] The second constraint is:

[0150]

[0151]

[0152] γ≤γ0

[0153]

[0154]

[0155] Among them, x(k) is the state variable at the current moment, ζ,ξ are positive scalars, γ0 is the upper limit of γ, X is a matrix, and I is an identity matrix.

[0156] The above state feedback control law is a feedback method in which the state variable of the system is transmitted to the input end through the proportional link. Its expression is:

[0157] u(k+i∣k)=Kx(k+i∣k),i≥0

[0158] Where K is the control gain.

[0159] The present invention uses a subspace identification algorithm to obtain a state-space model of the hybrid energy storage system. Compared to traditional topology-based models, its design does not require a tedious derivation process or the necessary expertise. The model building process is simple, and it can solve the problem of difficulty in obtaining an accurate model due to the complex structure of the hybrid energy storage system. Its design reduces the sensitivity of model predictive control to parameters and has high accuracy, thereby improving the control effect of predictive control. More importantly, taking into account the situation of parameter changes, the combination of offline and online identification is adopted to better simulate the actual working scenario. When the system parameters suddenly change, robust model predictive control is used to suppress the impact of parameter changes on control performance, thereby improving the robustness of the system.

[0160] Please refer to Figure 5 , Figure 5 This is a block diagram of a predictive control device for a hybrid energy storage system provided by an embodiment of the present invention; the specific device may include:

[0161] The prediction model building module 100 is used to establish a hybrid energy storage system state space model based on subspace identification according to offline input and output data, and use it as a prediction model for the predictive control strategy;

[0162] A model predictive control module 200 is configured to control the operation of the hybrid energy storage system using a model predictive control strategy when system parameters are fixed. The model predictive control strategy is to minimize cost as an optimization goal, establish a first objective function and a first constraint, measure the current system state, and obtain an optimal solution, which is then applied to the system.

[0163] The robust model predictive control module 300 is used to switch to the robust model predictive control strategy to control the operation of the hybrid energy storage system when the system parameters suddenly change, collect the online input and output data generated in real time during the control process, and use the subspace identification technology to obtain the online state space model of the hybrid energy storage system at the corresponding moment. The robust predictive control strategy is to minimize the cost as the optimization goal, and establish H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system.

[0164] The predictive control switching module 400 is used to switch from the robust model predictive control strategy back to the model predictive control strategy when a preset condition is met, and at the same time use the current online state space model as the predictive model of the model predictive control. The preset condition is that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than a preset threshold.

[0165] The hybrid energy storage system predictive control device of this embodiment is used to implement the aforementioned hybrid energy storage system predictive control method. Therefore, the specific implementation methods of the hybrid energy storage system predictive control device can be seen in the embodiment part of the hybrid energy storage system predictive control method mentioned above. For example, the prediction model construction module 100, the model predictive control module 200, the robust model predictive control module 300, and the predictive control switching module 400 are respectively used to implement steps S101, S102, S103, and S104 in the aforementioned hybrid energy storage system predictive control method. Therefore, its specific implementation methods can refer to the descriptions of the corresponding embodiments of each part and will not be repeated here.

[0166] A specific embodiment of the present invention further provides a hybrid energy storage system predictive control device, comprising: a memory for storing a computer program; and a processor for implementing the steps of the above-mentioned hybrid energy storage system predictive control method when executing the computer program.

[0167] A specific embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned predictive control method for a hybrid energy storage system are implemented.

[0168] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0169] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0170] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0171] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0172] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will appreciate that other variations or modifications can be made based on the above description. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the present invention.

Claims

1. A predictive control method for a hybrid energy storage system, characterized in that: include: Based on offline input and output data, a state space model of the hybrid energy storage system is established based on subspace identification and used as a prediction model for the predictive control strategy. When the system parameters are fixed, a model predictive control strategy is used to control the operation of the hybrid energy storage system. The model predictive control strategy is as follows: taking minimizing cost as the optimization goal, establishing a first objective function and a first constraint, measuring the current system state and obtaining an optimal solution, and applying it to the system; When the system parameters suddenly change, the robust model predictive control strategy is switched to control the operation of the hybrid energy storage system. The online input and output data generated in real time during the control process are collected, and the online state space model of the hybrid energy storage system at the corresponding moment is obtained using subspace identification technology. The robust predictive control strategy takes minimizing cost as the optimization goal and establishes H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system. Among them, the optimization goal is to minimize the cost and establish H ∞ The second objective function and the second constraint of the type include: Build H ∞ Type cost function: Among them, the first part is the cost term of the state variable deviation, and the second part is the cost term of the input variable deviation. Q s >0 and R s >0 is a symmetric weight matrix, μ is a positive scalar, d∈R q is the external interference, q is the dimension of the interference vector, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k, and d(k+i|k) represents the interference variable corresponding to the time instant k+i based on time k; To minimize H ∞ The cost function is used to establish the second objective function for the optimization goal: Obtain the H ∞ The upper bound of the cost function is obtained by using the state feedback control law to minimize the upper bound, and the second objective function is converted into finding Q, F, γ and λ so that Minimum: Among them, γ, λ is a positive scalar, Q, F is a matrix; The second constraint is: γ≤γ0 Where x(k) is the state variable at the current moment, ζ,ξ are positive scalars, γ0 is the upper limit of γ, X is a matrix, and I is an identity matrix; When the preset conditions are met, the robust model predictive control strategy is switched back to the model predictive control strategy, and the online state space model at this moment is used as the prediction model of the model predictive control strategy. The preset condition is that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than the preset threshold.

2. The predictive control method for a hybrid energy storage system according to claim 1, characterized in that: The optimization goal of minimizing cost is to establish the first objective function and the first constraint, which includes: establishing a cost function Among them, the first part is the cost term of the state variable deviation, the second part is the cost term of the input variable deviation, Q s >0 and R s >0 is a symmetric weighted matrix, x(k+i|k) represents the state prediction corresponding to the time instant k+i based on time k, and u(k+i|k) represents the input variable corresponding to the time instant k+i based on time k; The first objective function is established with minimizing the cost function as the optimization goal: stx(k+i+1|k)=Ax(k+i|k)+Bu(k+i|k) Where A∈R n×n , B∈R n×m is a constant matrix, is the maximum value of each input, m is the number of columns of matrix B; Obtain the upper bound of the cost function and minimize the upper bound using the state feedback control law, converting the first objective function into finding Q, F, and Z to minimize γ: Among them, γ is a positive scalar, Q, F is a matrix, and Z is a symmetric matrix; The first constraint is: Among them, x(k|k) represents the state prediction corresponding to the time instant k based on time k, I is a unit matrix, Y is a matrix, the symbol * represents a square in a symmetrical position, and Z jj is the j-th diagonal element of the matrix Z.

3. The predictive control method for a hybrid energy storage system according to claim 1, characterized in that: The state feedback control law is to transmit the state variable of the system to the input end through the proportional link for feedback, and its expression is: u(k+i∣k)=Kx(k+i∣k),i≥0 Where K is the control gain.

4. The predictive control method for a hybrid energy storage system according to claim 1, characterized in that: The process of establishing a hybrid energy storage system state space model based on subspace identification according to offline input and output data and using it as a prediction model for a predictive control strategy includes: Get offline input and output data: u=(u(0),u(1),...,u(s+N-2)) y=(y(0),y(1),...,y(s+N-2)) Where u(s+N-2) is the s+N-2th input data, y(s+N-2) is the s+N-2th output data, s is a constant strictly greater than the dimension n of the state vector, and N is the number of columns of the Hankel matrix; Construct the Hankel matrix based on the offline input and output data: Among them, m is the dimension of the input variable, and p is the dimension of the output variable; The Hankel matrix is ​​converted into a block lower triangular matrix with 0 in the upper right corner by LQ decomposition technique: Among them, L 11 ,L 22 is a lower triangular matrix, L 21 is a general matrix, Q1∈R N×sm ,Q2∈R N×sp is an orthogonal matrix; The lower triangular matrix L 22 ∈R sp×sp Converted to: Among them, U1∈R sp×n ,U2∈R sp×(sp-n) is a matrix, ∑1 is a diagonal matrix, V1, V2 are matrices; According to the transformed lower triangular matrix L 22 ∈R sp×sp Get the system estimated extended observation matrix According to the extended observation matrix, the system matrices A, B, C, and D are calculated by the subspace identification algorithm to obtain the state space model of the hybrid energy storage system based on subspace identification: x(k+1)=Ax(k)+Bu(k) y(k)=Cx(k)+Du(k), k=0,1… Among them, u(k) and y(k) are the input and output of the current moment respectively, x(k+1) is the input of the next moment, and k represents the kth moment.

5. The predictive control method for a hybrid energy storage system according to claim 1, characterized in that: Collect the online input and output data and update the state space model: x(k+1)=A(k)x(k)+B(k)u(k) Among them, A(k) and B(k) are real-time system matrices obtained based on online identification technology, x(k) is the state variable at the current moment, and u(k) is the input at the current moment.

6. The predictive control method for a hybrid energy storage system according to claim 1, characterized in that: The preset conditions are: in, is the load voltage controlled by the online state space model, and x5 is the actual load voltage of the system.

7. A predictive control device for a hybrid energy storage system, characterized in that: Implementing the predictive control method for a hybrid energy storage system according to claim 1 comprises: The prediction model building module is used to establish a state space model of the hybrid energy storage system based on subspace identification according to offline input and output data, and use it as a prediction model for the predictive control strategy; A model predictive control module is configured to control the operation of the hybrid energy storage system using a model predictive control strategy when system parameters are fixed. The model predictive control strategy is to establish a first objective function and a first constraint with minimizing cost as the optimization goal, measure the current system state, and obtain an optimal solution, which is then applied to the system. The robust model predictive control module is used to switch to the robust model predictive control strategy to control the operation of the hybrid energy storage system when the system parameters suddenly change. It collects the online input and output data generated in real time during the control process and uses the subspace identification technology to obtain the online state space model of the hybrid energy storage system at the corresponding moment. The robust predictive control strategy is to minimize the cost as the optimization goal and establish H ∞ The second objective function and second constraint of the type are used to measure the current system state and obtain the optimal solution, which is then applied to the system. The predictive control switching module is used to switch from the robust model predictive control strategy back to the model predictive control strategy when a preset condition is met, and at the same time use the online state space model at this moment as the prediction model of the model predictive control strategy. The preset condition is that the difference between the load voltage controlled by the online state space model and the actual load voltage of the system is not greater than a preset threshold.

8. A predictive control device for a hybrid energy storage system, characterized in that: include: Memory for storing computer programs; A processor is configured to implement the steps of a hybrid energy storage system predictive control method as claimed in any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a predictive control method for a hybrid energy storage system as claimed in any one of claims 1 to 6.

Images