Modularized direct current converter hybrid control method based on model prediction and current stress optimization
By introducing a hybrid control method combining model prediction and current stress optimization into the MDCC system, and combining it with a dual-phase-shifting current stress optimization algorithm, the problems of slow dynamic response and high current stress in the MDCC system are solved, achieving faster dynamic response and lower current stress, thus improving the stability and efficiency of the system.
Patent Information
- Application Number
- CN202411119300.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-15
- Publication Date
- 2024-11-26
AI Technical Summary
Existing MDCC systems suffer from slow dynamic response and high current stress. Traditional proportional-integral control is not robust, and single-phase-shift modulation control lacks sufficient degrees of freedom.
A hybrid control method for modular DC-DC converters based on model prediction and current stress optimization is adopted. Combined with a dual phase-shifting current stress optimization algorithm, the optimal phase-shifting control quantity is solved by rolling optimization through the construction of a cost function. The drive signal is embedded in the MDCC converter to balance the sub-module voltage.
It improves the dynamic response speed and robustness of the MDCC system, reduces current stress, achieves capacitor voltage balance, enhances system flexibility and stability, and reduces system losses.
Smart Images

Figure CN119030349A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medium and low voltage DC power grid control, and in particular to a modular DC converter hybrid control method based on model prediction and current stress optimization. Background Art
[0002] With the access of a large number of DC distributed power sources, energy storage devices and loads to the power grid, DC transmission and distribution technology has developed rapidly. As a key device in the DC distribution network, high-frequency DC converters have been widely used due to their small size, high power and light weight. In medium and low voltage DC distribution networks, MDCC converters combine the advantages of modular multi-level converters and dual active bridge converters, namely high fault tolerance, high pressure bearing capacity and easy soft switching of power devices. In addition, the dv / dt of the transformer can be effectively reduced by using quasi-square wave modulation, so it has broad application prospects.
[0003] At present, most MDCCs use a combination of traditional proportional-integral control and single-phase shift modulation. The proportional-integral control of the output voltage has the problems of slow dynamic response and weak robustness; the single-phase shift modulation has only one control degree of freedom, which simplifies the mathematical model but also leads to greater system current stress. Summary of the invention
[0004] Purpose of the invention: In order to improve the dynamic response speed of the MDCC system and optimize the current stress at the same time, a hybrid control method of modular DC converter based on model prediction and current stress optimization is proposed. This method adds a dual phase shift current stress optimization algorithm to the model predictive control strategy, and obtains the optimal solution by constructing a cost function, thereby further improving the performance of the MDCC system, thereby effectively solving the above-mentioned problems existing in the prior art.
[0005] The present invention proposes a hybrid control method for a modular DC converter based on model prediction and current stress optimization, the control method comprising the following steps:
[0006] The mathematical model of modular multilevel DC / DC converter (MDCC) under double phase shift and quasi-square wave modulation is established, and the per-unit value of system transmission power and current stress are obtained under two states.
[0007] The per-unit value of the current stress is used as the objective function, and the per-unit value of the transmission power is used as the constraint equation to obtain a transfer ratio relationship under the optimal condition of the current stress;
[0008] A state space model is established for the output side of the MDCC converter and discretized using the forward Euler method.
[0009] A prediction model is established on the basis of a discretized state space model, and a cost function for a control target is constructed. The cost function is solved by a rolling optimization method to obtain an optimal phase shift control amount.
[0010] The optimal phase shift control amount is embedded in the driving signal, and on this basis, the capacitor voltage data of the bridge arm submodule is collected and the voltage of each submodule is balanced to generate a new driving signal to act on the MDCC converter.
[0011] In a further embodiment, the transmission power per unit value P * The calculation formula is as follows:
[0012]
[0013] The current stress per unit value I * The calculation formula is as follows:
[0014] I * =2[(k-1)(1-d)+2(1-k)D 1 +4D 2 ]
[0015] Where d is the phase shift angle between adjacent submodules; n is the transformer ratio; k is the voltage transfer ratio, k = U 1 / nU 2 , U 1 , U 2 are the input and output side voltages respectively; D 1 Compared with the inward shift, D 2 is the external shift ratio; P is the rated transmission power; P b is the normalized transmission power.
[0016] In a further embodiment, the shift ratio relationship under the optimal current stress condition is as follows:
[0017]
[0018] In the formula, k is the voltage transfer ratio, k = U 1 / nU 2 , U1 and U2 are the input and output side voltages respectively; D 1 Compared with the inward shift, D 2 Compared with outward migration.
[0019] In a further embodiment, a state space model is established for the output side of the MDCC converter, and the state space model is based on the output capacitor C 0 Express:
[0020]
[0021] In the formula, C0 is the output capacitor; R is the output load; P is the transmission power; U 2 is the output side voltage;
[0022] The state space model expression is discretized by the forward Euler method, and the expression is as follows:
[0023]
[0024] The state space model expression after discretization is as follows:
[0025]
[0026] Where, T is the switching period; U 2 (t k ) and U 2 (t k+1 ) are the output voltage value at the current moment and the output voltage value at the next moment respectively; d(t k ) is the input control variable in the model predictive control algorithm; f s is the switching frequency; L is the inductance value.
[0027] In a further embodiment, the process of constructing the prediction model includes:
[0028] Assume that the control quantity is from t k N steps of change start to occur at the moment, and the predicted future output voltage in the time domain P can be expressed as:
[0029]
[0030] Where U 2 (t k+i |t k ) is the voltage value U at the current moment 2 (t k ) to predict t k+i The predicted voltage value at time i = 1, 2, ..., P; d(t k+j ) is the control quantity of the predicted j-th step, j = 0, 1, ..., N-1;
[0031] Convert the output voltage expression in the prediction time domain P into a prediction model in matrix form:
[0032] U 2 =FU 2 (t k )+BD
[0033] In the formula, the matrix
[0034] In a further embodiment, according to the two control objectives of rapid output response and stable regulation process, a cost function J(k) for measuring the output performance of predictive control is constructed:
[0035]
[0036] In the formula, is the reference value of the system output voltage at time k+i; r j is the weight coefficient of the control quantity;
[0037] The rolling optimization method is used to solve the cost function J(k). The optimization goal is to minimize the cost function. The expression is:
[0038]
[0039] In the formula, R represents the control quantity weight coefficient matrix; It represents the predicted output voltage reference value at time P in the future under the action of the predicted control amount.
[0040] In a further embodiment, the optimization objective is substituted into the prediction model and only the variables related to the control quantity D are considered, and the cost function is further simplified as follows:
[0041]
[0042] Without considering the control quantity constraint, the simplified cost function J is minimized to obtain the optimal phase shift control quantity D(t k ):
[0043]
[0044] Take the first phase shift relationship of the control sequence as the input control quantity and set the weight coefficient r of the control quantity j Set to 0, the current optimal phase shift control amount d(t k ) is expressed as:
[0045]
[0046] In the formula, It represents the predicted output voltage reference value at time P in the future under the action of the predicted control amount.
[0047] In a further embodiment, embedding the optimal phase shift control amount into a driving signal specifically includes:
[0048] Generate a drive signal based on the MDCC converter switching frequency, and convert the optimal shift phase to D 1 and D 2 Embedded in it, reflected in the half cycle T hsThe phase shift of the two sets of switch tube drive signals inside the medium voltage side and the low voltage side is D 1 T hs , and the phase shift of the external switch tube drive signal between the two sides is D 2 T hs .
[0049] In a further embodiment, the process of collecting the capacitor voltage data of the bridge arm submodule and balancing the voltage of each submodule specifically includes:
[0050] The current submodule voltage and the voltage increment of each submodule calculated compared with the previous cycle are sorted respectively, and the driving signal with the largest voltage increment is allocated to the submodule with the lowest current voltage, and the driving signal with the smallest voltage increment is allocated to the submodule with the highest current voltage.
[0051] In addition, the present invention further proposes a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the process of the above-mentioned modular DC converter hybrid control method can be automatically run.
[0052] Compared with the prior art, the present invention can achieve the following beneficial effects:
[0053] (1) The quasi-square wave and double shift combined modulation method adopted by the present invention can not only effectively reduce the dv / dt of the transformer, but also further reduce the current stress.
[0054] (2) Compared with the traditional proportional-integral control, the model predictive control adopted in the present invention has faster dynamic response speed, excellent steady-state performance and stronger robustness.
[0055] (3) The control method combining model prediction and current stress optimization proposed in the present invention can simultaneously reduce the current stress of the system and improve the dynamic response speed of the system. By embedding the current stress optimization algorithm under dual phase shift control into the model predictive control, the rolling optimization method is used to calculate the optimal solution based on the construction of the cost function.
[0056] (4) The capacitor voltage double sorting algorithm adopted by the present invention is not limited by voltage gain and load conditions, and the voltage equalization effect is better and more excellent. BRIEF DESCRIPTION OF THE DRAWINGS
[0057] Figure 1 It is the overall topological structure diagram of the MDCC converter in the present invention.
[0058] Figure 2 It is a waveform diagram of the voltage between the bridge arms on both sides and the inductor current when the MDCC converter in the present invention operates in two modes.
[0059] Figure 3It is a flow chart of the model prediction and current stress optimization hybrid control method proposed in an embodiment of the present invention.
[0060] Figure 4 1 is an output side equivalent circuit diagram in an embodiment of the present invention.
[0061] Figure 5 4 is a schematic diagram of a submodule capacitor voltage double sorting algorithm used in an embodiment of the present invention. DETAILED DESCRIPTION
[0062] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features known in the art are not described.
[0063] This embodiment discloses a modular DC converter hybrid control method based on model prediction and current stress optimization, which specifically includes the following steps:
[0064] Step 1: Analyze the working principle of the MDCC converter under double phase shift and quasi-square wave modulation and establish a mathematical model. Figure 1 As shown, the medium voltage side of the MDCC converter is a single-phase MMC structure, in which each bridge arm consists of N half-bridge sub-modules SM and a bridge arm inductor L s Composition: The low-voltage side is a DAB converter, and the two sides are connected through a high-frequency transformer and an inductor L. This topology combines the advantages of the MMC converter and the DAB converter, and has the characteristics of high fault tolerance, high voltage bearing capacity, and easy soft switching of power devices. In addition, the dv / dt of the transformer is effectively reduced through quasi-square wave modulation. Compared with single phase shift, double phase shift modulation can further reduce current stress by adding a control degree of freedom. The specific analysis and establishment of the mathematical model include the following sub-steps:
[0065] Step 1.1, according to the relationship between the internal and external shifts, the operation mode of the MDCC converter can be divided into two types: Figure 2 (a) and (b) are respectively 0≤D 1 ≤D 2 ≤1 and 0≤D 2 ≤D 1 ≤1 waveforms of voltage between bridge arms and inductor current in two modes, where D 1 Compared with the inward shift, D 2 For the convenience of analysis, the voltage steps under quasi-square wave modulation are simplified to oblique lines.
[0066] Step 1.2: Combination Figure 2And the MDCC power transfer characteristics, the power transfer expression is obtained:
[0067]
[0068] Where d is the phase shift angle between adjacent submodules; n is the transformer ratio; U 1 , U 2 are the input and output voltages respectively; f s is the switching frequency; L is the inductance value.
[0069] Step 1.3: Define P b =nU 1 U 2 / 8f s L is the power reference value, and the transmission power per unit value can be obtained as:
[0070]
[0071] Step 1.4: Define the inductor current reference value I N =nU 2 / 8f s L, the maximum current stress and its per-unit value can be expressed as:
[0072]
[0073] I * =2[(k-1)(1-d)+2(1-k)D 1 +4D 2 ]
[0074] Where, T hs is half a switching cycle; where k = U 1 / nU 2 is the voltage transfer ratio.
[0075] Step 2: Use the Lagrange multiplier method to solve the shift ratio combination that can minimize the current stress under fixed transmission power. Compared with the traditional step-by-step method, it is more conducive to simplifying the calculation. Taking the per-unit value of current stress as the objective function and the per-unit value of transmission power as the constraint equation, the Lagrangian function can be expressed as:
[0076] L(D 1 ,D 2 ,λ)=I * (D 1 ,D 2 )+λ(P * (D 1 ,D 2 )-P 0N )
[0077] Where λ is the constraint coefficient, P0N is the output normalized power reference value.
[0078] By finding the Lagrangian function with respect to D 1 and D 2 The partial derivative of the current stress can be expressed as:
[0079]
[0080] Step 3: Figure 3 The model prediction and current stress optimization hybrid control flow chart shown can be divided into the following sub-steps:
[0081] Step 3.1: Figure 4 The output side equivalent circuit diagram shown in the figure establishes the state space model, which is expressed as follows:
[0082]
[0083] Substituting the transmission power expression into the formula, we can get:
[0084]
[0085] In the formula, d(D 1 ,D 2 ) is the combined function of the internal and external phase shift angles, and its expression is:
[0086]
[0087] Step 3.2: Discretize the state space model expression by the forward Euler method. The formula of the forward Euler method can be expressed as:
[0088]
[0089] Where, T is the control period; U 2 (t k ) and U 2 (t k+1 ) are the output voltage value at the current moment and the output voltage value at the next moment respectively.
[0090] The discretized state space model expression after forward Euler method is:
[0091]
[0092] In the formula, d(t k ) is the input control variable in the model predictive control algorithm.
[0093] Step 3.3: Further simplify the discretized state space model expression, where:
[0094]
[0095] The simplified state space model can be expressed as:
[0096] U 2 (t k+1 )=μ 1 U 2 (t k )+μ 2 d(t k )
[0097] Step 4: Based on the discrete state space model, the model predictive control is performed according to k The output voltage U of the MDCC converter at the moment 2 (t k ), and in addition, assuming that the phase shift control amount d(t k ), d(t k+1 ),…,d(t k+N-1 ) is the quantity after N steps of change and does not change afterwards. The output voltage in the predicted time domain P can be expressed as:
[0098]
[0099] Where P and N are the prediction time domain and control time domain respectively. 2 (t k+i |t k ) is the current voltage value U 2 (t k ) to predict t k+i The predicted voltage value at time i = 1, 2, ..., P; d(t k+j ) is the predicted control quantity of the jth step, j = 0, 1, ..., N-1. On this basis, solving the optimal shift control quantity can be divided into the following sub-steps:
[0100] Step 4.1: Convert the output voltage expression in the prediction time domain P into a prediction model in matrix form:
[0101] U 2 =FU 2 (t k )+BD
[0102] In the formula, the matrix
[0103] Step 4.2: Based on the two control objectives of rapid output response and stable regulation process, a cost function to measure the output performance of predictive control is constructed, and its expression is:
[0104]
[0105] Where U 2 * (t k+i ) is the reference value of the system output voltage at time k+i, r j is the weight coefficient of the control quantity. j When it is 0, the system output voltage error is small, and the size of the control quantity can be ignored. j If the value is large, the cost function will limit it.
[0106] Step 4.3: Use the rolling optimization method to solve the cost function. The optimization goal is to minimize the cost function. The expression is:
[0107]
[0108] Step 4.4, substituting it into the prediction model and considering only the variables related to the control quantity D, the cost function can be further simplified as:
[0109]
[0110] Without considering the control quantity constraint, the simplified cost function is minimized to obtain the optimal phase shift control quantity expression:
[0111]
[0112] Step 4.5: Take the first phase shift relation of the control sequence as the input control quantity and set r j Set to 0, the current optimized shift ratio expression is:
[0113]
[0114] Step 5: The optimal shift phase combination can be calculated by system sampling and combining the above steps. On this basis, the optimal shift phase is embedded in the modulation signal and the bridge arm submodule capacitor voltage data is collected. Figure 5 The capacitor voltage sorting algorithm shown completes the submodule voltage balancing. Its principle is to sort the current submodule voltage and the voltage increment of each submodule calculated by comparing with the previous cycle, and assign the drive signal with the largest voltage increment to the submodule with the lowest current voltage, and the drive signal with the smallest voltage increment to the submodule with the highest current voltage. Finally, a new drive signal is generated to act on the MDCC. This ensures that the output voltage is quickly responded to and the current stress is optimized under the premise of stable system operation.
[0115] The modular DC converter hybrid control method based on model prediction and current stress optimization proposed in the present invention does not need to calculate the optimal shift ratio by combining multiple equations. The calculation process is relatively simple and has better dynamic response speed and smaller current stress. It improves the flexibility and stability of MDCC when operating in medium and low voltage DC distribution networks, and also further reduces the loss of the system and improves efficiency.
[0116] The present embodiment also provides a computer-readable storage medium, in which at least one executable instruction is stored, and when the executable instruction is executed on an electronic device, the electronic device performs the operation of the modular DC converter hybrid control method based on model prediction and current stress optimization as described in the above embodiment. More specific examples of the computer-readable storage medium disclosed in the present embodiment may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.
[0117] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, which carry readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device.
[0118] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes may be made to it in form and detail without departing from the spirit and scope of the present invention as defined in the appended claims.
Claims
1. A modular DC converter hybrid control method based on model prediction and current stress optimization, characterized in that: The steps include: Calculate the per-unit value of the transmission power and the per-unit value of the current stress of the MDCC converter under double phase shift; The per-unit value of the current stress is used as the objective function, and the per-unit value of the transmission power is used as the constraint equation to obtain a transfer ratio relationship under the optimal condition of the current stress; A state space model is established for the output side of the MDCC converter and discretized using the forward Euler method. A prediction model is established on the basis of a discretized state space model, and a cost function for a control target is constructed. The cost function is solved by a rolling optimization method to obtain an optimal phase shift control amount. The optimal phase shift control amount is embedded in the driving signal, and on this basis, the capacitor voltage data of the bridge arm submodule is collected and the voltage of each submodule is balanced to generate a new driving signal to act on the MDCC converter.
2. The modular DC converter hybrid control method according to claim 1, characterized in that: The transmission power per unit value P * The calculation formula is as follows: The current stress per unit value I * The calculation formula is as follows: I * =2[(k-1)(1-d)+2(1-k)D1+4D2] Where, d is the micro-phase shift angle between adjacent submodules; n is the transformer ratio; k is the voltage transfer ratio, k = U1 / nU2, U1 and U2 are the input and output side voltages respectively; D1 is the internal phase shift ratio, D2 is the external phase shift ratio; P is the rated transmission power; P b is the normalized transmission power.
3. The modular DC converter hybrid control method according to claim 1, characterized in that: The shift ratio relationship under the optimal condition of current stress is as follows: Where, k is the voltage transfer ratio, k=U1 / nU2, U1 and U2 are the input and output side voltages respectively; D1 is the inward phase ratio, and D2 is the outward phase ratio.
4. The modular DC converter hybrid control method according to claim 1, characterized in that: A state space model is established for the output side of the MDCC converter. The state space model is expressed in terms of output capacitance C0: In the formula, C0 is the output capacitor; R is the output load; P is the transmission power; U2 is the output side voltage; The state space model expression is discretized by the forward Euler method, and the expression is as follows: The state space model expression after discretization is as follows: Where, T is the switching period; U2(t k ) and U2(t k+1 ) are the output voltage value at the current moment and the output voltage value at the next moment respectively; d(t k ) is the input control variable in the model predictive control algorithm; f s is the switching frequency; L is the inductance value.
5. The modular DC converter hybrid control method according to claim 4, characterized in that: The construction process of the prediction model includes: Assume that the control quantity is from t k N steps of change start to occur at the moment, and the predicted future output voltage in the time domain P can be expressed as: In the formula, U2(t k+i |t k ) is the voltage value U2(t k ) to predict t k+i The predicted voltage value at time i = 1, 2, ..., P; d(t k+j ) is the control quantity of the predicted j-th step, j = 0, 1, ..., N-1; Convert the output voltage expression in the prediction time domain P into a prediction model in matrix form: U2=FU2(t k )+BD In the formula, the matrix 6. The modular DC converter hybrid control method according to claim 5, characterized in that: According to the two control objectives of rapid output response and stable regulation process, the cost function J(k) for measuring the output performance of predictive control is constructed: In the formula, is the reference value of the system output voltage at time k+i; r j is the weight coefficient of the control quantity; The rolling optimization method is used to solve the cost function J(k). The optimization goal is to minimize the cost function. The expression is: In the formula, R represents the control quantity weight coefficient matrix; It represents the predicted output voltage reference value at time P in the future under the action of the predicted control amount.
7. The modular DC converter hybrid control method according to claim 6, characterized in that: Substituting the optimization objective into the prediction model and considering only the variables related to the control quantity D, the cost function is further simplified as: Without considering the control quantity constraint, the simplified cost function J is minimized to obtain the optimal phase shift control quantity D(t k ): Take the first phase shift relationship of the control sequence as the input control quantity and set the weight coefficient r of the control quantity j Set to 0, the current optimal phase shift control amount d(t k ) is expressed as: In the formula, It represents the predicted output voltage reference value at time P in the future under the action of the predicted control amount.
8. The modular DC converter hybrid control method according to claim 7, characterized in that: Embedding the optimal phase shift control amount into a driving signal specifically includes: The drive signal is generated according to the switching frequency of the MDCC converter, and the optimal shift phase D1 and D2 are embedded in it, which is reflected in the half cycle T hs The phase shift of the two sets of switch tube drive signals inside the medium voltage side and the low voltage side is D1T. hs , and the phase shift of the external switch tube drive signal between the two sides is D2T hs .
9. The modular DC converter hybrid control method according to claim 8, characterized in that: The collecting of bridge arm submodule capacitor voltage data and balancing the voltage of each submodule specifically includes: The current submodule voltage and the voltage increment of each submodule calculated compared with the previous cycle are sorted respectively, and the driving signal with the largest voltage increment is allocated to the submodule with the lowest current voltage, and the driving signal with the smallest voltage increment is allocated to the submodule with the highest current voltage.
10. A computer-readable storage medium having a program stored thereon, wherein when the program is executed by a processor, the modular DC converter hybrid control method according to any one of claims 1 to 9 is implemented.
Citation Information
Cited By
Driving control method of DAB converter and related equipment
CN120729023A
