A Model-Free Predictive Voltage Control Method for a Magnetic Network Power Router

Abstract

The present invention discloses a model-free predictive voltage control method for a magnetic network power router. First, a discrete model of the output voltage of the magnetic network power router is established, and then the output voltage gradient of the magnetic network power router is constructed. Next, according to the output voltage of the magnetic network power router and the output voltage reference, an adaptive phase-shift angle range is designed to provide a phase-shift angle for the output voltage prediction calculation. Then, a voltage gradient look-up table is established to establish the voltage gradient relationship of different phase-shift angles and update the voltage gradients of all phase-shift angles. Finally, according to the voltage gradient and its corresponding phase-shift angle, the output voltage at the future moment is predicted to achieve model-free predictive control, and the value function is used to evaluate the optimal phase-shift angle and apply it in the next control cycle to realize the regulation of the output voltage of the magnetic network power router. This control method improves the robustness of the voltage control of the magnetic network power router and enhances the scalability of the voltage control method in the application of magnetic network energy routers with different parameters.

Description

Technical Field

[0001] The invention belongs to the field of isolated power electronic converters, and particularly relates to a model-free predictive voltage control method for a magnetic network power router. Background Art

[0002] With the rapid development of renewable energy, the scale and application scope of DC microgrids are constantly expanding. In order to achieve flexible configuration and optimal utilization of renewable energy in DC microgrids, the application of DC-DC converters has been promoted, providing support and guarantee for the development of DC microgrids. Among them, magnetic network power routers have received extensive attention and research due to their abilities such as bidirectional power transmission, wide voltage conversion range, and zero voltage switching. When a magnetic network power router is applied to a DC microgrid, it must effectively cope with operating conditions such as load mutations and voltage mutations. Therefore, the robustness and dynamic performance of the control system are important indicators for the efficient control of magnetic network power routers.

[0003] In recent years, MPC (model predictive control) has been applied and studied in magnetic network power routers due to its advantages such as rapid dynamic response, simple control principle, and strong multi-objective control ability, realizing stable and efficient output voltage control of magnetic network power routers. MPC uses the discrete model of the magnetic network power router to predict the output voltage at future moments. However, the control performance of MPC depends on the accurate discrete model of the system. During the actual operation of the magnetic network power router system, the actual parameters of inductors and capacitors will change with temperature changes, hardware aging, and changes in operating conditions, resulting in errors between the actual parameters and the model parameters, thereby reducing the performance of the magnetic network power router.

[0004] To improve the parameter robustness of the MPC output voltage control of magnetic network power routers, researchers have proposed various MPC methods based on parameter identification. However, before implementing the MPC method based on parameter identification, a relatively accurate system model still needs to be established to obtain the relationship between system identification variables and input-output variables. At the same time, the above parameter identification methods ignore the influence of parasitic parameters, and when identification errors exist, they will still affect the output voltage performance of the magnetic network power router. Summary of the Invention

[0005] The technical problem to be solved by the present invention is: aiming at the problems existing in the prior art, a model-free predictive voltage control method for a magnetic network power router is provided. By establishing and updating the output voltage gradient of different phase-shift angles of the magnetic network power router, model-free predictive voltage control is realized, solving the technical problem of the influence of traditional leakage inductance and capacitance parameter changes on traditional control, and achieving the invention purpose of robust, fast, and accurate voltage control of the magnetic network power router.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A model-free predictive voltage control method for a magnetic network power router, characterized by comprising the following steps:

[0007] S1. According to the magnetic network power router model, establish a discrete model of the output voltage of the magnetic network power router, and then construct the output voltage gradient of the magnetic network power router;

[0008] S2. According to the output voltage of the magnetic network power router and the output voltage reference, calculate the adaptive phase-shifting angle range of the magnetic network power router to provide a phase-shifting angle for the output voltage prediction calculation;

[0009] S3. According to the adaptive phase-shifting angle range of the magnetic network power router, establish a voltage gradient look-up table, and establish the voltage gradient relationship of different phase-shifting angles, and update the voltage gradients of all phase-shifting angles;

[0010] S4. According to the voltage gradient of the magnetic network power router and its corresponding phase-shifting angle, predict the output voltage at the future moment to achieve model-free predictive control; use the value function to evaluate the optimal phase-shifting angle and apply it in the next control cycle to realize the regulation of the output voltage of the magnetic network power router.

[0011] Further, in the aforementioned step S1, establish a discrete model of the output voltage of the magnetic network power router as follows:

[0012]

[0013] where T s is the control cycle of the magnetic network power router; v o (k) is the output voltage of the magnetic network power router at the k-th moment; i o (k) is the output current of the magnetic network power router at the k-th moment; v o (k + 1) is the output voltage of the magnetic network power router at the (k + 1)-th moment; v i (k) is the input voltage of the magnetic network power router at the k-th moment; D(k) is the phase-shifting angle between the input side and the output side of the magnetic network power router at the k-th moment; C o is the output capacitance of the magnetic network power router; n is the turns ratio of the transformer of the magnetic network power router; L r is the leakage inductance of the transformer of the magnetic network power router.

[0014] Further, in the aforementioned step S1, construct the output voltage gradient of the magnetic network power router as follows:

[0015] Δv o (k) = v o (k) - v o (k - 1),

[0016] where, Δv o (k) is the output voltage gradient of the magnetic network power router, which corresponds to the applied phase shift angle D(k - 1).

[0017] Furthermore, in the aforementioned step S2, the adaptive phase shift angle range of the magnetic network power router is as follows:

[0018] {D(k) - 2*ΔD a (k), D(k) - ΔD a (k), D(k), D(k) + ΔD a (k), D(k) + 2*ΔD a (k)}, where, ΔD a (k) = ΔD(k)(1 + εΔV 2 (k)), ΔDa(k) is the adaptive phase shift angle, v m is the maximum power error, ε is the adjustment coefficient, ΔV(k) is the output voltage error, v oref (k) is the output voltage reference of the magnetic network power router.

[0019] Furthermore, in the aforementioned step S3, the specific voltage gradient relationship for different phase shift angles is as follows:

[0020]

[0021] Furthermore, in the aforementioned step S4, the prediction of the output voltage at the future moment is specifically: Based on the phase shift angle range obtained in step S2 set as {D(k) - 2*ΔD a (k), D(k) - ΔD a (k), D(k), D(k) + ΔD a (k), D(k) + 2*ΔD a (k)} and the output voltage gradients {Δv o1 (k), Δv o2 (k), Δv o3 (k), Δv o4 (k), Δv o5 (k)} corresponding to the phase shift angles obtained in step S3, combined with the output voltage v o (k) at time k, the output voltage v ox (k + 1) at time (k + 1) can be obtained as: v ox (k + 1) = v o (k) + Δv ox (k) x ∈ {1, 2, 3, 4, 5}.

[0022] Further, in the aforementioned step S4, the optimal phase-shifting angle is evaluated using the value function and applied in the next control cycle to achieve the regulation of the output voltage of the magnetic network power router. The value function is as follows:

[0023] G = α1G1 + α2G2, where G1 is the value function for reference voltage tracking, G2 is the value function for ensuring the steady-state performance of the output voltage of the magnetic network power router, and α1 and α2 are the weight factors of G1 and G2, respectively.

[0024] Compared with the prior art, the beneficial technical effects of the present invention adopting the above technical solutions are as follows:

[0025] (1) The model-free predictive voltage control method for the magnetic network power router proposed by the present invention uses the output voltage gradient of the magnetic network power router to replace the traditional mathematical model, realizes model-free predictive voltage control, and improves the robustness of the output voltage control of the magnetic network power router.

[0026] (2) Through the model-free predictive voltage control method of the magnetic network power router of the present invention, by establishing the relationship between the output voltage gradients of different phase-shifting angles, the real-time update of the output voltage gradients of all phase-shifting angles is realized, and the accuracy of the model-free predictive voltage control of the magnetic network power router is improved.

[0027] (3) The principle of the model-free predictive voltage control method for the magnetic network power router proposed by the present invention is simple and is easily extended to magnetic network power routers with different parameters, different power levels, and different ports, and has high practical value. Description of the Drawings

[0028] Figure 1 is a control block diagram of a model-free predictive voltage control method for a magnetic network power router proposed by the present invention.

[0029] Figure 2 is a schematic diagram of a model-free predictive voltage control method for a magnetic network power router proposed by the present invention.

[0030] Figure 3 is a flowchart of a model-free predictive voltage control method for a magnetic network power router proposed by the present invention. Detailed Embodiments

[0031] In order to better understand the technical content of the present invention, specific embodiments are hereby given and described in conjunction with the accompanying drawings as follows.

[0032] Aspects of the present invention are described with reference to the accompanying drawings, in which a number of illustrative embodiments are shown. Embodiments of the present invention are not limited to those described in the drawings. It should be understood that the present invention can be implemented by any one of the various concepts and embodiments introduced above, and those described in detail below, because the concepts and embodiments disclosed in the present invention are not limited to any particular embodiment. Additionally, some aspects disclosed in the present invention can be used alone or in any suitable combination with other aspects disclosed in the present invention.

[0033] The present invention provides a model-free predictive voltage control method for a magnetic network power router, and the control block diagram is as Figure 1 shown. The specific process refers to Figure 3 , and includes the following steps:

[0034] S1. According to the magnetic network power router model, establish a discrete model of the output voltage of the magnetic network power router, and then construct the output voltage gradient of the magnetic network power router;

[0035] S2. According to the output voltage and output voltage reference of the magnetic network power router, calculate the adaptive phase shift angle range of the magnetic network power router to provide a phase shift angle for the output voltage prediction calculation;

[0036] S3. According to the adaptive phase shift angle range of the magnetic network power router, establish a voltage gradient lookup table, establish the voltage gradient relationship for different phase shift angles, and update the voltage gradients for all phase shift angles;

[0037] S4. According to the voltage gradient of the magnetic network power router and its corresponding phase shift angle, predict the output voltage at a future moment to achieve model-free predictive control; evaluate the optimal phase shift angle using a value function and apply it in the next control cycle to achieve the regulation of the output voltage of the magnetic network power router.

[0038] In step S1, the structure of the magnetic network power router is as Figure 1 shown. According to the structure of the magnetic network power router, the magnetic network power router model shown in Figure 2 is obtained. Among them, the secondary side output voltage and output current of the magnetic network power router can be expressed as:

[0039]

[0040] In the formula: C o is the output capacitor of the magnetic network power router, v o is the output voltage of the magnetic network power router, i s is the output current of the H-bridge on the output side of the magnetic network power router, and i o is the output current of the magnetic network power router.

[0041] By using the forward Euler method, formula (1) is discretized and expressed as:

[0042]

[0043] Where: T s is the control period of the magnetic network power router; v o (k) is the output voltage of the magnetic network power router at time k; i o (k) is the output current of the magnetic network power router at time k; i s (k) is the output current of the H-bridge on the output side of the magnetic network power router; v o (k + 1) is the output voltage of the magnetic network power router at time k + 1.

[0044] Under the single-phase shift modulation method, the transmission power P of the magnetic network power router is:

[0045]

[0046] Where: v1 is the voltage on the input side of the transformer of the magnetic network power router; D is the phase shift angle between the input side and the output side of the magnetic network power router, and its range is [-0.5 to 0.5]; n is the turns ratio of the transformer of the magnetic network power router; L r is the leakage inductance of the transformer of the magnetic network power router.

[0047] According to formula (3), the output current i s (k) of the H-bridge on the output side of the magnetic network power router at time k can be calculated and expressed as:

[0048]

[0049] Where: v i (k) is the input voltage of the magnetic network power router at time k; D(k) is the phase shift angle between the input side and the output side of the magnetic network power router at time k.

[0050] Substituting formula (4) into (2) gives:

[0051]

[0052] According to formula (5), it can be seen that the prediction of the output voltage of the magnetic network power router at time k + 1 depends on the model parameters L r of the leakage inductance of the transformer and C o of the output capacitance. When the model parameters L r of the leakage inductance of the transformer of the magnetic network power router and the model parameters C o of the output capacitance are compared with the actual parameters L r0 of the leakage inductance of the transformer and the actual parameters C o0When there is a mismatch, it will lead to a predicted voltage error. To eliminate the influence of model parameters on the output voltage prediction of the magnetic network power router, the output voltage gradient can be established, thereby gradienting the traditional discrete model. The established output voltage gradient can be expressed as:

[0053]

[0054] In the formula: Δv o (k) is the output voltage gradient of the magnetic network power router, which corresponds to the applied phase shift angle D(k - 1).

[0055] The adaptive phase shift angle range of the magnetic network power router in S2 is as follows:

[0056] In the magnetic network power router for bidirectional power transmission, the phase shift angle range is -0.5 to 0.5. Taking the magnetic network power router for unidirectional power transmission as an example, its phase shift angle range is 0 to 0.5, and the discrete form of the phase shift angle range can be expressed as {0, ΔD, 2ΔD, …, 0.5}. Among them, the phase shift angle ΔD can be calculated and expressed as:

[0057]

[0058] In the formula: T c is the clock period of the magnetic network power router controller.

[0059] To reduce the phase shift angle range {0, ΔD, 2ΔD, …, 0.5} and at the same time satisfy the output voltage regulation of the magnetic network power router, the phase shift angle range can be set as {D(k) - 2 * ΔD a (k), D(k) - ΔD a (k), D(k), D(k) + ΔD a (k), D(k) + 2 * ΔD a (k)}. Among them, D(k) is the optimal phase shift angle of the previous control period, D(k) = D(k - 1), and ΔD a (k) is the adaptive phase shift angle, which can be calculated and expressed as:

[0060] ΔD a (k) = ΔD(k)(1 + εΔV 2 (k)) (8)

[0061] Among them:

[0062]

[0063] In the formula: v m is the maximum power error, ε is the adjustment coefficient, ΔV(k) is the output voltage error, and v oref (k) is the output voltage reference of the magnetic network power router.

[0064] In S3, according to the adaptive phase-shifting angle range of the magnetic network power router, a voltage gradient lookup table is established, which can be expressed as:

[0065] Table 1 Lookup table of the magnetic network power router at time k

[0066]

[0067] The model-free predictive voltage control of the present invention is calculated based on the s exhale voltage gradient in the lookup table of each control period. When predicting the output voltage at time (k + 1), the output voltage gradient of the lookup table at time k needs to be used. Therefore, all gradients in the lookup table at time k are first updated within the kth control period. If the output voltage gradients corresponding to some phase-shifting angles are not updated within the kth control period, the output voltage gradients of past control periods may exist in the lookup table at time k, resulting in prediction calculation errors.

[0068] As shown in Table 1, five phase-shifting angles respectively correspond to five voltage gradients. Among them, since D(k) = D(k - 1), the voltage gradient Δv o3 (k) corresponding to the phase-shifting angle D(k) is calculated by formula (5). Therefore, formula (5) can be further expressed as:

[0069]

[0070] In order to update the output voltage gradients corresponding to the remaining four voltage gradients, the relationship of the output voltage gradients of different phase-shifting angles can be established. According to formula (10), it can be known that Δv o (k) = Δv o3 (k), which corresponds to the applied phase-shifting angle D(k). When the phase-shifting angle D(k) is replaced by the phase-shifting angle D(k) - 2 * ΔD a (k), formula (10) can be further expressed as:

[0071]

[0072] Taking the difference between formula (10) and formula (11) gives:

[0073]

[0074] According to formula (12), it can be known that the relationship between the phase-shifting angles of two voltage gradients can be used to update the voltage gradients of the remaining phase-shifting angles. At the same time, the model parameters L r of the leakage inductance of the transformer of the magnetic network power router and the model parameter C o of the output capacitor still exist in formula (12). In order to eliminate the influence of the model parameters on the voltage gradient update, formula (12) can be advanced by one control period, which can be expressed as:

[0075]

[0076] Dividing formula (12) by formula (13) gives:

[0077]

[0078] Therefore, Δv o1 (k) can be obtained through the phase-shifting formula (14) and expressed as:

[0079]

[0080] Similarly, replace the phase-shifting angle D(k) - 2*ΔD a (k) in formula (15) with the voltage gradients Δv o2 (k), Δv o4 (k) and Δv o5 (k) corresponding phase-shifting angles D(k) - ΔD a (k), D(k) + ΔD a (k), D(k) + 2*ΔD a (k). The voltage gradients Δv o2 (k), Δv o4 (k) and Δv o5 (k) can be updated and expressed respectively as:

[0081]

[0082] In S4, according to the voltage gradient of the magnetic network power router and its corresponding phase-shifting angle, the output voltage at the future moment is predicted to achieve model-free predictive control. The specific process is as follows:

[0083] Figure 2 It is a schematic diagram of the model-free predictive voltage control method for the magnetic network power router. In the (k - 1)th control cycle, by applying the phase-shifting angle D(k - 1), the output voltage changes from v o (k - 1) to v o (k). To achieve model-free predictive voltage control, at the beginning of the kth control cycle, all output voltage gradients in the look-up table at the kth moment experience the calculation period T c and then are updated. The update process is shown in step S3. By calculating the phase-shifting angle range and its corresponding output voltage gradient in S2 and evaluating the prediction error according to the value function, model-free predictive voltage control is achieved. The specific implementation process is as follows.

[0084] Based on the phase-shifting angle range obtained in S2, it is set as {D(k) - 2*ΔD a (k), D(k) - ΔD a(k), D(k), D(k) + ΔD a (k), D(k) + 2*ΔD a}(k)} and the output voltage gradient {Δv corresponding to the phase-shifting angle obtained in S3 o1 (k), Δv o2 (k), Δv o3 (k), Δv o4 (k), Δv o5}(k)}, combined with the output voltage v at time k o (k) can obtain the output voltage v at time (k + 1) ox (k + 1):

[0085] v ox (k + 1) = v o (k) + Δv ox (k) x ∈ {1, 2, 3, 4, 5} (19)

[0086] As shown in formulas (15) to (19), the proposed model-free predictive voltage calculation of the magnetic network power router does not require the use of any model parameters, realizes model-free predictive control, and improves the parameter robustness of the output voltage control of the magnetic network power router.

[0087] In order to select the optimal phase-shifting angle from {D(k) - 2*ΔD a (k), D(k) - ΔD a (k), D(k), D(k) + ΔD a (k), D(k) + 2*ΔD a (k)} and apply it in the next control period, this section defines the double-objective value function G as shown in formula (20):

[0088] G = α1G1 + α2G2 (20)

[0089] Wherein:

[0090]

[0091] In the formula: G1 is the value function of reference voltage tracking, G2 is the value function to ensure the steady-state performance of the output voltage of the magnetic network power router, and α1 and α2 are the weight factors of G1 and G2 respectively.

[0092] Although the present invention has been described above with preferred embodiments, it is not intended to limit the present invention. Those with ordinary knowledge in the technical field to which the present invention pertains can make various modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be determined by the claims.

Claims

1. A model-free predictive voltage control method for a magnetic network power router, characterized in that: The steps include: S1. According to the magnetic network power router model, a discrete model of the output voltage of the magnetic network power router is established, and then the output voltage gradient of the magnetic network power router is constructed; S2. According to the output voltage of the magnetic network power router and the output voltage reference, the adaptive phase shift angle range of the magnetic network power router is calculated to provide a phase shift angle for output voltage prediction calculation; S3. According to the adaptive phase shift angle range of the magnetic network power router, a voltage gradient query table is established, and the voltage gradient relationship of different phase shift angles is established, and the voltage gradients of all phase shift angles are updated; S4. According to the voltage gradient of the magnetic network power router and its corresponding phase shift angle, the output voltage at the future moment is predicted to realize model-free predictive control; the optimal phase shift angle is evaluated by using the value function and applied in the next control cycle to realize the output voltage regulation of the magnetic network power router.

2. A model-free predictive voltage control method for a magnetic network power router according to claim 1, characterized in that: In step S1, a discrete model of the output voltage of the magnetic network power router is established, as follows: Among them, T s is the control cycle of the magnetic network power router; v o (k) is the output voltage of the magnetic network power router at time k; i o (k) is the output current of the magnetic network power router at time k; v o (k+1) is the output voltage of the magnetic network power router at time k+1; v i (k) is the input voltage of the magnetic network power router at time k; D(k) is the phase shift angle between the input side and the output side of the magnetic network power router at time k; C o is the output capacitance of the magnetic network power router; n is the turns ratio of the transformer of the magnetic network power router; L r Transformer leakage inductance for the magnetic network power router.

3. The model-free predictive voltage control method for a magnetic network power router according to claim 1 is characterized in that: In step S1, the output voltage gradient of the magnetic network power router is constructed as follows: Δv o (k)=v o (k)-v o (k-1), Where Δv o (k) is the output voltage gradient of the magnetic network power router, which corresponds to the applied phase shift angle D(k-1).

4. The model-free predictive voltage control method for a magnetic network power router according to claim 1 is characterized in that: In step S2, the adaptive phase shift angle range of the magnetic network power router is as follows: {D(k)-2*ΔD a (k),D(k)-ΔD a (k),D(k),D(k)+ΔD a (k),D(k)+2*ΔD a (k)}, Where, ΔD a (k) = ΔD(k)(1+εΔV 2 (k)), ΔDa(k) is the adaptive phase shift angle, v m is the maximum power error, ε is the adjustment coefficient, ΔV(k) is the output voltage error, v oref (k) is the output voltage reference of the magnetic network power router.

5. A model-free predictive voltage control method for a magnetic network power router according to claim 4, characterized in that: In step S3, the voltage gradient relationship of different phase shift angles is established as follows: (D(k-1)(1-2D(k-1))-(D(k-1)-2D a (k-1))(1-2(D(k-1)-2D a (k-1)))) 6. A model-free predictive voltage control method for a magnetic network power router according to claim 1, characterized in that: In step S4, the output voltage at the future moment is predicted by: setting the phase shift angle range obtained in step S2 to {D(k)-2*ΔD a (k), D(k)-ΔD a (k), D(k), D(k)+ΔD a (k), D(k)+2*ΔD a (k)} and the output voltage gradient {Δv o1 (k), Δv o2 (k), Δv o3 (k), Δv o4 (k), Δv o5 (k)}, combined with the output voltage vo(k) at time k, the output voltage vox(k+1) at time (k+1) can be obtained as: v ox (k+1)=v o (k)+Δv ox (k)x∈{1,2,3,4,5}.

7. A model-free predictive voltage control method for a magnetic network power router according to claim 1, characterized in that: In step S4, the optimal phase shift angle is evaluated using a value function and applied in the next control cycle to achieve output voltage regulation of the magnetic network power router. The value function is: G=α1G1+α2G2,where G1 is the value function of reference voltage tracking, G2 is the value function of ensuring the steady-state performance of the output voltage of the magnetic network power router, and α1 and α2 are the weight factors of G1 and G2 respectively.

Images