Multi-distributed optical storage converter grid-connected reactive power layered collaborative optimization distribution method

Abstract

The invention discloses a hierarchical collaborative optimization distribution method for grid-connected reactive power of a multi-distributed optical storage converter, and belongs to the field of grid-connected control of the multi-distributed optical storage converter. The method comprises local layer control and optimal distribution control of grid-connected reactive power support power of a network layer, and the local layer control realizes a conventional voltage control type virtual synchronous generator algorithm and a virtual impedance algorithm according to parameters of each distributed optical storage converter. By adding the virtual impedance algorithm, the system stability can be improved, the design range of network layer control parameters is enlarged, and the reactive power optimization distribution effect is improved; the network layer control realizes grid-connected reactive power optimal distribution of the multi-distributed optical storage converter on the basis of the local layer control; through cooperation of local layer control and network layer control, grid-connected reactive power sharing of the multi-distributed optical storage converter, improvement of voltage control precision and enhancement of system stability are jointly realized.

Description

Technical Field

[0001] The present invention belongs to the field of grid-connected control of multiple distributed photovoltaic storage converters, and in particular relates to a hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters. Background Art

[0002] With the large number of distributed renewable energy power generation equipment, loads based on power electronic interfaces, and single-phase loads being connected to the distribution network in a dispersed and disorderly manner, the voltage fluctuation of the distribution network may be large. Conventional reactive power compensation devices are installed in small numbers and concentrated in the distribution network, which makes it difficult to meet the reactive power compensation and voltage support requirements of the system under the access of distributed renewable energy. In recent years, with the development of energy storage technology and its widespread application in new power systems, photovoltaic storage grid-connected equipment that combines energy storage with photovoltaic power generation has become an effective means to solve the power quality problems in low-voltage areas and improve the stability of grid connection. The control of photovoltaic storage converters based on the voltage-controlled virtual synchronous generator algorithm has become a research hotspot at home and abroad.

[0003] The reactive loop of the conventional voltage-controlled virtual synchronous generator algorithm adopts reactive-voltage droop control, which has the ability to actively support the grid voltage. When the voltage amplitude of the distribution network to which it is connected fluctuates, the photovoltaic storage grid-connected converter absorbs or emits reactive power according to the droop characteristics for compensation, so as to reduce voltage fluctuations and maintain voltage stability. However, due to the different capacities of each distributed photovoltaic storage grid-connected device and the different lengths of the lines connected to the distribution network, it is difficult for each distributed photovoltaic storage converter to reasonably absorb or emit reactive compensation power according to its rated capacity. Even photovoltaic storage converters with small rated capacity are overcurrent protected due to excessive power generation, which reduces the stability of system operation.

[0004] At present, the research on reactive power distribution of multiple renewable energy converters at home and abroad mainly focuses on the reactive power sharing problem of inverters in isolated microgrid systems. There are few studies on the reasonable power distribution problem of distributed grid-connected converters under voltage support conditions. In addition, the reactive power sharing strategies of converters in isolated microgrids are mostly adaptive virtual impedance algorithms or secondary power regulation based on centralized communication and distributed communication, such as:

[0005] 1. The article entitled "Improved droop control strategy for reactive power balancing of inverters in low-voltage microgrids" by Bai Xiaodan, Miao Hong, Zeng Chengbi, and Most, "High Voltage Technology", Vol. 46, No. 04, 2020, pp. 1310-1318, constructed a centralized communication network in which the central controller communicated with each inverter and calculated the average reactive power of the system. Each inverter adaptively adjusted the virtual impedance value according to the average reactive power and its own output reactive power to achieve accurate distribution of reactive power. However, the centralized communication network architecture has problems such as information flow congestion, large communication delay, poor reliability and redundancy.

[0006] 2. In the article entitled “An Adaptive Virtual Impedance control for improving Power Sharing among Inverters in Islanded AC Microgrids,” 2020 IEEE Energy Conversion Congress and Exposition (ECCE), 2020, 4621-4626, the virtual impedance is adaptively adjusted according to the output current value of each inverter in the microgrid to reduce the reactive sharing error.

[0007] 3. "Research on reactive power distribution of isolated AC microgrid based on improved droop control" Yan Li, Mi Yang, Sun Wei2, Cai Hangyi, Niu Qingsong, Qi Fujun, Acta Energiae Solaris Sinica, Vol. 42, No. 08, 2021, pp. 07-15. The article adopts a distributed consensus algorithm and a virtual impedance strategy. The distributed inverters in the isolated microgrid adaptively adjust their own virtual impedance values ​​according to the reactive signals transmitted by neighboring nodes, and finally realize the reasonable distribution of reactive power output of all node inverters. The consensus algorithm based on the distributed communication network reduces the system's dependence on global communication and has higher reliability and scalability.

[0008] However, the introduction of the virtual impedance algorithm inevitably causes the inverter output voltage to further decrease, thereby reducing the voltage control accuracy and even causing system voltage collapse. In order to reduce the voltage deviation, additional voltage secondary regulation is required. Summary of the invention

[0009] The technical problem to be solved by the present invention is the problem of reasonable allocation of reactive power output by each photovoltaic storage inverter when the photovoltaic storage grid-connected inverter adopts a voltage-controlled virtual synchronous machine algorithm to actively support the voltage in a scenario where multiple distributed photovoltaic storage devices are connected to a low-voltage distribution network. The present invention provides a hierarchical collaborative optimization allocation method for the reactive power of multiple distributed photovoltaic storage inverters connected to the grid, so that each distributed photovoltaic storage inverter can reasonably allocate output power according to its own rated capacity in the dynamic and steady-state process of outputting reactive power to support the grid voltage, and has high voltage control accuracy and system stability. In addition, the method provided by the present invention has high reliability and scalability.

[0010] To achieve the above-mentioned object, the present invention provides a hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters, wherein the hierarchical collaborative optimization allocation method involves N photovoltaic storage converters and a distribution network, wherein each photovoltaic storage converter is connected to a photovoltaic module and an energy storage battery on the DC side, the photovoltaic module and the energy storage battery are connected in parallel to provide a DC power supply for the photovoltaic storage converter, and the photovoltaic storage converter outputs a three-phase AC voltage on the AC side and is connected to the distribution network through a line impedance;

[0011] The hierarchical collaborative optimization allocation method includes local layer control and network layer grid-connected reactive support power optimization allocation control; the local layer control implements a voltage-controlled virtual synchronous generator algorithm and a virtual impedance algorithm according to the parameters of each photovoltaic storage converter itself; the network layer control, based on the local layer control, realizes the optimization allocation of grid-connected reactive power of multiple distributed photovoltaic storage converters by constructing a distributed communication network of multiple photovoltaic storage converters and a reactive optimization allocation strategy based on multi-agent consistency; any one of the N photovoltaic storage converters is recorded as photovoltaic storage converter i, i∈[1,N], and the hierarchical collaborative optimization allocation method includes the following steps:

[0012] Step 1: construct local layer control for the photovoltaic storage converter i based on its own sampling parameters and voltage-controlled virtual synchronous generator algorithm;

[0013] The local layer control includes an output power calculation module, a voltage-controlled virtual synchronous generator algorithm module, a virtual impedance module, a voltage-current dual closed-loop control module, and a drive signal generation module;

[0014] Step 1.1: The output power calculation module samples the three-phase output voltage u obtained by the photovoltaic storage converter i. ioa ,u iob ,u ioc And the three-phase output current i ioa ,i iob ,i ioc The output voltage dq component U in the synchronous rotating coordinate system is obtained by Clark transformation and Park transformation iod ,U ioq and the output current dq component I iod ,I ioq ; Then calculate the output instantaneous active power P ie And output instantaneous reactive power Q ie , and are filtered by a first-order low-pass filter to obtain the output average active power P ie_av And the output average reactive power Q ie_av ;

[0015] Step 1.2: The voltage-controlled virtual synchronous generator algorithm module outputs average active power P ie_av And the output average reactive power Qie_av Perform active loop and reactive loop control. Specifically, calculate the reference angular frequency ω of the output voltage of the current beat according to the active loop control equation. i After that, the output voltage reference phase angle θ of the current beat is obtained by integration. i ,θ i =∫ω i ; The output voltage reference angular frequency ω of the current beat i and the reference phase angle θ of the output voltage i Used for the next beat calculation and form a closed loop; the voltage amplitude d-axis command U is calculated according to the reactive loop control equation ird ;

[0016] Step 1.3, let the voltage amplitude q-axis command U irq =0, change U ird ,U irq and I iod ,I ioq As the input of the virtual impedance module, the dq axis voltage amplitude reference value U of the output voltage and current dual closed-loop control module is obtained through virtual impedance calculation. iRd ,U iRq ;

[0017] Step 1.4: The voltage and current dual closed-loop control module uses the output voltage reference phase angle θ i And dq axis voltage amplitude reference value U iRd ,U iRq Perform double closed-loop control of the output voltage and output current of the photovoltaic storage converter i, and send its output as a modulation wave to the drive signal generation module to generate a drive signal for the power switch tube through a modulation strategy;

[0018] Step 2: Multiple distributed photovoltaic storage converters build a network-layer grid-connected reactive power optimization allocation strategy based on a distributed communication network architecture and a consistency algorithm;

[0019] The network layer grid-connected reactive power optimization allocation strategy includes a signal sending module, a signal receiving module and a grid-connected reactive power optimization allocation module of a photovoltaic storage converter, and includes the following steps:

[0020] Step 2.1: According to the situation of distributed photovoltaic storage converters connected to the distribution network, a distributed communication network including N photovoltaic storage converters is constructed based on multi-agent distributed communication and consistency control theory. Specifically, each photovoltaic storage converter i in the distributed communication network can transmit its own reactive power-voltage droop coefficient k iU And the output average reactive power Q ie_av Transmitted to one or more neighboring photovoltaic storage converters j, and at the same time, the reactive power-voltage droop coefficient k sent by one or more neighboring photovoltaic storage converters j can be received through the signal receiving module jUAnd the output average reactive power Q je_av ; Assume that there are M neighboring solar-storage converters, j is the serial number of the neighboring solar-storage converter, j∈[1,M], M<N;

[0021] Step 2.2, the grid-connected reactive power optimization allocation module of the photovoltaic storage converter i calculates the reactive power-voltage droop coefficient k of each neighboring photovoltaic storage converter j received jU And the output average reactive power Q je_av The product of ΔU cij , and calculate the total reactive power deviation ΔU of M neighboring solar-storage converters j cj , the calculation formulas are:

[0022]

[0023] Step 2.3: The grid-connected reactive power optimization allocation module of the photovoltaic storage converter i is based on its own reactive power-voltage droop coefficient k. iU And the output average reactive power Q ie_av , the total reactive power deviation ΔU obtained in step 2.2 cj Calculate the reactive power distribution compensation value ΔU ci , and the reactive power distribution compensation value ΔU ci It is superimposed into the reactive loop control equation in the local layer control of the photovoltaic storage converter i;

[0024] The reactive power distribution compensation value ΔU ci The calculation formula is:

[0025]

[0026] where k ipc is the proportional coefficient of the grid-connected reactive power optimization allocation algorithm, k iic is the integral coefficient of the grid-connected reactive power optimization allocation algorithm, and s is the Laplace operator;

[0027] Step 2.4, perform the calculations of steps 2.2 to 2.3 on each of the N photovoltaic storage converters i to obtain N reactive power distribution compensation values ​​ΔU ci , and are superimposed on the reactive loop control equation in the local layer control of the photovoltaic storage converter i, so as to optimize the distribution of the grid-connected reactive support power of multiple distributed photovoltaic storage converters according to their own rated capacity.

[0028] Preferably, the transformation formulas of the Clark transformation and the Park transformation in step 1.1 are respectively:

[0029]

[0030] Among them, θ i′ is the reference phase angle of the output voltage of the photovoltaic storage converter obtained in the previous shot;

[0031] The output instantaneous active power P ie And output instantaneous reactive power Q ie The calculation formulas are:

[0032]

[0033] The output average active power P ie_av And the output average reactive power Q ie_av The calculation formulas are:

[0034]

[0035] Among them, T if is the filtering time constant of the first-order low-pass filter.

[0036] Preferably, the active loop control equation in step 1.2 is:

[0037]

[0038] Where ω0 is the rated angular frequency of the system voltage, ω i ′ is the reference angular frequency of the output voltage of the photovoltaic converter obtained in the previous shot, k iω is the primary frequency modulation coefficient, P iref is the grid-connected active power instruction, D i is the damping coefficient, J i is the virtual inertia;

[0039] The reactive loop control equation is:

[0040] U ird =U0-k iU (Q ie_av -Q iref )+ΔU ci

[0041] Among them, U0 is the rated output voltage amplitude of the system, k iU is the reactive power-voltage droop coefficient, Q iref is the grid-connected reactive power instruction, ΔU ci The reactive power distribution compensation value calculated by the network layer grid-connected reactive power optimization distribution strategy.

[0042] Preferably, the dq axis voltage amplitude reference value U in step 1.3 is iRd ,U iRq The calculation formula is:

[0043]

[0044] Among them, Li is the virtual inductance value of the photovoltaic storage converter i, and ω0 is the rated angular frequency of the system voltage.

[0045] Preferably, step 1.4 is implemented as follows:

[0046] First, the output voltage and output current of the photovoltaic storage converter i are controlled in a double closed loop. Specifically, the output I of the dq axis output voltage outer loop control is calculated by the dq axis output voltage outer loop control equation. iRd ,I iRq , the output U of the dq axis output current inner loop control equation is obtained ipwmd ,U ipwmq ;

[0047] The dq axis output voltage outer loop control equation and the dq axis output current inner loop control equation are respectively:

[0048]

[0049]

[0050] Among them, k ipv is the proportional coefficient of the output voltage outer loop proportional-integral controller, k iiv is the integral coefficient of the output voltage outer loop proportional integral controller, k ipi is the proportional coefficient of the output current inner loop proportional-integral controller, k iii is the integral coefficient of the output current inner loop proportional-integral controller, s is the Laplace operator;

[0051] Then the three-phase voltage modulation wave u is calculated ipwma ,u ipwmb ,u ipwmc , the calculation formula is:

[0052]

[0053] Compared with the existing reactive power allocation scheme based on centralized communication network architecture and the reactive power allocation scheme based on adaptive virtual impedance algorithm, the hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters disclosed in the present invention has the following beneficial effects:

[0054] 1. This hierarchical collaborative optimization allocation method includes building local layer control based on the voltage-controlled virtual synchronous generator algorithm, and network layer grid-connected reactive power optimization allocation control based on the distributed communication network architecture and consistency algorithm. Through the mutual coordination of network layer and local layer control, the grid-connected reactive power support power of multiple distributed photovoltaic storage converters is optimized and allocated according to their own rated capacity, and at the same time has high reliability and scalability;

[0055] 2. In order to enable each distributed photovoltaic storage inverter to reasonably allocate output power according to its own rated capacity in the dynamic and steady-state process of outputting reactive power and to have higher system stability, this hierarchical collaborative optimization allocation method adds a virtual impedance algorithm to the local layer control of the photovoltaic storage inverter. The virtual impedance value is a fixed value and is designed only according to its own control parameters without the need for adaptive changes, which reduces the complexity of parameter design. The introduction of the virtual impedance algorithm will not reduce the voltage control accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is an electrical connection diagram and distributed network architecture diagram of N distributed photovoltaic storage devices connected to a power distribution network according to an embodiment of the present invention.

[0057] Figure 2 This is a local layer control based on the voltage-controlled virtual synchronous generator algorithm and a network layer grid-connected reactive power optimization distribution control structure diagram of an embodiment of the present invention.

[0058] Figure 3 The photovoltaic storage converter of the embodiment of the present invention adopts the local layer control based on the voltage-controlled virtual synchronous generator algorithm, and does not introduce the virtual impedance algorithm (ie, L i =0mH), the grid voltage suddenly drops 5% at 1.5s, and the output average reactive power Q under U0 working condition ie_av and output voltage amplitude d-axis command U ird waveform.

[0059] Figure 4 The embodiment of the present invention adopts the local layer control based on the voltage-controlled virtual synchronous generator algorithm and the network layer grid-connected reactive power optimization distribution control, and does not introduce the virtual impedance algorithm. When the grid voltage suddenly drops by 5% at 1.5s, the average reactive power Q is output under the U0 working condition. ie_av and output voltage amplitude d-axis command U ird waveform.

[0060] Figure 5 In the embodiment of the present invention, the proportional coefficient k of the grid-connected reactive power optimization allocation algorithm is ipc The waveform of the output reactive power when it increases to 3.5;

[0061] Figure 6 for Figure 5 When the grid voltage suddenly drops by 5% after 1.5s of operation, the average reactive power Q under U0 condition is ie_av and output voltage amplitude d-axis command U ird Waveform diagram of . DETAILED DESCRIPTION

[0062] The present embodiment is described in detail below with reference to the accompanying drawings.

[0063] Figure 1 The electrical connection diagram and distributed network architecture diagram of N distributed photovoltaic storage devices connected to the distribution network in the embodiment of the present invention are as follows: Figure 1 It can be seen that the hierarchical collaborative optimization allocation method involves N photovoltaic storage inverters and a distribution network, wherein each photovoltaic storage inverter is connected to a photovoltaic module and an energy storage battery on the DC side, and the photovoltaic module and the energy storage battery are connected in parallel to provide a DC power supply for the photovoltaic storage inverter. The photovoltaic storage inverter outputs a three-phase AC voltage on the AC side and is connected to the distribution network through the line impedance.

[0064] In this embodiment, the DC side voltage of the photovoltaic storage inverter is 600V, and the rated phase voltage of the output three-phase AC voltage is 220V / 50Hz.

[0065] In this embodiment, the number of distributed photovoltaic and energy storage grid-connected devices N = 2 constitutes the minimum system, and the rated capacity of each photovoltaic and energy storage converter is S r =20KVA, the switching frequency of the power switch tube and the control sampling frequency are both 10kHz. The photovoltaic storage converter 1 sets its own reactive power-voltage droop coefficient k 1U And the output average reactive power Q 1e_av The signal is transmitted to the neighboring photovoltaic storage converter 2 through the signal sending module, and at the same time, the reactive power-voltage droop coefficient k sent by the neighboring photovoltaic storage converter 2 can be received through the signal receiving module. 2U And average reactive power information Q 2e_av .

[0066] Figure 2 This is a local layer control based on the voltage-controlled virtual synchronous generator algorithm and a network layer grid-connected reactive power optimization distribution control structure diagram of the embodiment of the present invention. Figure 2 It can be seen that the hierarchical collaborative optimization allocation method includes local layer control and network layer grid-connected reactive support power optimization allocation control; the local layer control implements the voltage-controlled virtual synchronous generator algorithm and the virtual impedance algorithm according to the parameters of each photovoltaic storage inverter itself, and the network layer control is based on the local layer control, by constructing a distributed communication network of multiple photovoltaic storage inverters, and a reactive power optimization allocation strategy based on multi-agent consistency, to achieve the optimal allocation of grid-connected reactive power of multiple distributed photovoltaic storage inverters.

[0067] Any one of the N photovoltaic storage converters is denoted as photovoltaic storage converter i, i∈[1,N], and the hierarchical collaborative optimization allocation method comprises the following steps:

[0068] Step 1: construct local layer control for the photovoltaic storage converter i based on its own sampling parameters and voltage-controlled virtual synchronous generator algorithm;

[0069] The local layer control includes an output power calculation module, a voltage-controlled virtual synchronous generator algorithm module, a virtual impedance module, a voltage-current dual closed-loop control module, and a drive signal generation module;

[0070] Step 1.1: The output power calculation module samples the three-phase output voltage u obtained by the photovoltaic storage converter i. ioa ,u iob ,u ioc And the three-phase output current i ioa ,i iob ,i ioc The output voltage dq component U in the synchronous rotating coordinate system is obtained by Clark transformation and Park transformation iod ,U ioq and the output current dq component I iod ,I ioq ; Then calculate the output instantaneous active power P ie And output instantaneous reactive power Q ie , and are filtered by a first-order low-pass filter to obtain the output average active power P ie_av And the output average reactive power Q ie_av .

[0071] In this embodiment, the transformation formulas of the Clark transformation and the Park transformation are respectively:

[0072]

[0073] Among them, θ i ′ is the reference phase angle of the output voltage of the photovoltaic storage converter obtained in the previous shot;

[0074] The output instantaneous active power P ie And output instantaneous reactive power Q ie The calculation formulas are:

[0075]

[0076] The output average active power P ie_av And the output average reactive power Q ie_av The calculation formulas are:

[0077]

[0078] Among them, T if is the filtering time constant of the first-order low-pass filter, and s is the Laplace operator. In this embodiment, the cutoff frequency of the first-order low-pass filter is set to 100 Hz, and the corresponding T if =0.0016s.

[0079] Step 1.2: The voltage-controlled virtual synchronous generator algorithm module outputs average active power P ie_av And the output average reactive power Q ie_av Perform active loop and reactive loop control. Specifically, calculate the reference angular frequency ω of the output voltage of the current beat according to the active loop control equation. i After that, the output voltage reference phase angle θ of the current beat is obtained by integration. i ,θ i =∫ω i ; The output voltage reference angular frequency ω of the current beat i and the reference phase angle θ of the output voltage i Used for the next beat calculation and form a closed loop; the voltage amplitude d-axis command U is calculated according to the reactive loop control equation ird .

[0080] In this embodiment, the active loop control equation in step 1.2 is:

[0081]

[0082] Where ω0 is the rated angular frequency of the system voltage, ω i ′ is the reference angular frequency of the output voltage of the photovoltaic converter obtained in the previous shot, k iω is the primary frequency modulation coefficient, P iref is the grid-connected active power instruction, D i is the damping coefficient, J i is the virtual inertia. In this embodiment, ω0=314.159rad / s, k iω =6366.2W·s / rad,P iref =0kW,D i =6366.2W·s / rad,J i =4kg·m 2 .

[0083] The reactive loop control equation is:

[0084] U ird =U0-k iU (Q ie_av -Q iref )+ΔU ci

[0085] Among them, U0 is the rated output voltage amplitude of the system, k iU is the reactive power-voltage droop coefficient, Q iref is the grid-connected reactive power instruction, ΔU ci It is the reactive power distribution compensation value calculated by the network layer grid-connected reactive power optimization distribution strategy. In this embodiment, U0 = 311V, k iU =0.0016V / kVA, Qiref =0kvar.

[0086] Step 1.3, let the voltage amplitude q-axis command U irq =0, change U ird ,U irq and I iod ,I ioq As the input of the virtual impedance module, the dq axis voltage amplitude reference value U of the output voltage and current dual closed-loop control module is obtained through virtual impedance calculation. iRd ,U iRq .

[0087] In this embodiment, the dq axis voltage amplitude reference value U iRd ,U iRq The calculation formula is:

[0088]

[0089] Among them, L i is the virtual inductance value of the photovoltaic storage converter i.

[0090] The virtual impedance in the present invention mainly plays the role of active and reactive decoupling when the photovoltaic storage converter is connected to the low-voltage distribution network, and enhances the system stability while improving the reactive optimization distribution effect of the multi-distributed photovoltaic storage converter. Therefore, the virtual inductance value L i The design is based on the active and reactive power decoupling control requirements and the parameters related to the network layer's grid-connected reactive power optimization allocation strategy, and is a fixed value. i =0.02mH.

[0091] Step 1.4: The voltage and current dual closed-loop control module uses the output voltage reference phase angle θ i And dq axis voltage amplitude reference value U iRd ,U iRq The output voltage and output current of the photovoltaic storage inverter i are controlled in a double closed loop, and its output is sent to the drive signal generation module as a modulation wave, and the drive signal of the power switch tube is generated through the modulation strategy.

[0092] In this embodiment, the implementation steps of step 1.4 are as follows:

[0093] First, the output voltage and output current of the photovoltaic storage converter i are controlled in a double closed loop. Specifically, the output I of the dq axis output voltage outer loop control is calculated by the dq axis output voltage outer loop control equation. iod ,I ioq , the output U of the dq axis output current inner loop control equation is obtained ipwmd ,U ipwmq ;

[0094] The dq axis output voltage outer loop control equation and the dq axis output current inner loop control equation are respectively:

[0095]

[0096]

[0097] Among them, k ipv is the proportional coefficient of the output voltage outer loop proportional-integral controller, k iiv is the integral coefficient of the output voltage outer loop proportional integral controller, k ipi is the proportional coefficient of the output current inner loop proportional-integral controller, k iii is the integral coefficient of the output current inner loop proportional-integral controller, s is the Laplace operator;

[0098] Then the three-phase voltage modulation wave u is calculated ipwma ,u ipwmb ,u ipwmc , the calculation formula is:

[0099]

[0100] In this embodiment, k ipv =10,k iiv =100,k ipi =0.15, k iii =0.

[0101] Step 2: Multiple distributed photovoltaic storage converters build a network-layer grid-connected reactive power optimization allocation strategy based on a distributed communication network architecture and a consistency algorithm;

[0102] The network layer grid-connected reactive power optimization allocation strategy includes a signal sending module, a signal receiving module and a grid-connected reactive power optimization allocation module of a photovoltaic storage converter, and includes the following steps:

[0103] Step 2.1: According to the situation of distributed photovoltaic storage converters connected to the distribution network, a distributed communication network including N photovoltaic storage converters is constructed based on multi-agent distributed communication and consistency control theory. Specifically, each photovoltaic storage converter i in the distributed communication network can transmit its own reactive power-voltage droop coefficient k iU And the output average reactive power Q ie_av Transmitted to one or more neighboring photovoltaic storage converters j, and at the same time, the reactive power-voltage droop coefficient k sent by one or more neighboring photovoltaic storage converters j can be received through the signal receiving module jU And the output average reactive power Q je_av ; Assume that there are M neighboring solar-storage converters, j is the serial number of the neighboring solar-storage converter, j∈[1,M], M<N;

[0104] Step 2.2, the grid-connected reactive power optimization allocation module of the photovoltaic storage converter i calculates the reactive power-voltage droop coefficient k of each neighboring photovoltaic storage converter j received jU And the output average reactive power Q je_av The product of ΔU cij , and calculate the total reactive power deviation ΔU of M neighboring solar-storage converters j cj , the calculation formulas are:

[0105]

[0106] Step 2.3: The grid-connected reactive power optimization allocation module of the photovoltaic storage converter i is based on its own reactive power-voltage droop coefficient k. iU And the output average reactive power Q ie_av , the total reactive power deviation ΔU obtained in step 2.2 cj Calculate the reactive power distribution compensation value ΔU ci , and the reactive power distribution compensation value ΔU ci It is superimposed into the reactive loop control equation in the local layer control of the photovoltaic storage converter i;

[0107] The reactive power distribution compensation value ΔU ci The calculation formula is:

[0108]

[0109] where k ipc is the proportional coefficient of the grid-connected reactive power optimization allocation algorithm, k iic It is the integral coefficient of the grid-connected reactive power optimization allocation algorithm;

[0110] Step 2.4, perform the calculations of steps 2.2 to 2.3 on each of the N photovoltaic storage converters i to obtain N reactive power distribution compensation values ​​ΔU ci , and are superimposed on the reactive loop control equation in the local layer control of the photovoltaic storage converter i, so as to optimize the distribution of the grid-connected reactive support power of multiple distributed photovoltaic storage converters according to their own rated capacity.

[0111] In this embodiment, N=2, M=1, k ipc =3.5, k iic =15.

[0112] The invention in this embodiment is applicable to the scenario where multiple distributed photovoltaic storage devices are connected to a low-voltage distribution network, and when the photovoltaic storage grid-connected converter adopts a voltage-controlled virtual synchronous machine algorithm for active voltage support, the reactive power output by each photovoltaic storage converter is reasonably distributed.

[0113] To prove the technical effect of the present invention, Figure 2The control method shown in the figure is used to simulate the grid-connected system of two 20kVA photovoltaic storage inverters. The main circuit structure and control parameters of the two photovoltaic storage inverters are the same, but the grid-connected impedance of photovoltaic storage inverter 1 connected to the low-voltage distribution network is 0.001+j0.018Ω, and the grid-connected impedance of photovoltaic storage inverter 2 connected to the low-voltage distribution network is 0.005+j0.07Ω.

[0114] Figure 3 The two PV-storage converters only use the local layer control based on the voltage-controlled virtual synchronous generator algorithm, and do not introduce the virtual impedance algorithm (i.e., L i =0mH), the grid voltage suddenly drops 5% at 1.5s, and the output average reactive power Q under U0 working condition ie_av and output voltage amplitude d-axis command U ird It can be seen that due to the impedance difference between the two photovoltaic storage inverters connected to the low-voltage distribution network, the reactive support power output by photovoltaic storage inverter 1 during its active voltage support is three times that of photovoltaic storage inverter 2, and reaches 75% of its rated capacity. If the photovoltaic storage inverter needs to output active power at this time, it is very easy to cause overcurrent and trigger a protective shutdown. In addition, because photovoltaic storage inverter 1 outputs a large reactive power, its output voltage amplitude is low, close to the limit of 10% voltage amplitude deviation, and the voltage control accuracy is low.

[0115] Figure 4 The two photovoltaic storage converters both use local layer control based on the voltage-controlled virtual synchronous generator algorithm and network layer grid-connected reactive power optimization distribution control, and do not introduce the virtual impedance algorithm. The output average reactive power Q under the condition of the grid voltage suddenly dropping by 5% U0 at 1.5s is ie_av and output voltage amplitude d-axis command U ird The proportional coefficient k of the grid-connected reactive power optimization allocation algorithm ipc =1.5, integration coefficient k iic =15. The grid-connected reactive power optimization distribution control at the network layer effectively reduces the reactive power difference between the two photovoltaic storage converters, and the output reactive power remains consistent in steady state. The output voltage amplitude of photovoltaic storage converter 1 is relatively large. Figure 3 The output voltage amplitude deviation of the photovoltaic storage converter 2 is relatively large. Figure 3 It has been reduced by 1.32%. However, there is still unequal distribution in the dynamic process of the photovoltaic storage inverter outputting reactive power, and the maximum difference is about 2kvar.

[0116] In order to reduce the difference in reactive power output of the photovoltaic storage converter in the dynamic process and realize dynamic reactive power distribution, it is necessary to increase the proportional coefficient k of the grid-connected reactive power optimization distribution algorithm. ipc , k ipc In the initial stage of reactive power response, the reactive power distribution weight is larger, and the larger k ipcThe value of k can make the output reactive power have a better balancing effect in the initial response stage. ipc increases, the system stability decreases. Figure 5 The proportional coefficient k of the grid-connected reactive power optimization allocation algorithm is ipc When it increases to 3.5, the waveforms of the reactive power output of the two photovoltaic storage converters show that the output reactive power oscillates with an amplitude exceeding the rated capacity, and the system becomes unstable.

[0117] Both PV-storage converters use local layer control based on voltage-controlled virtual synchronous generator algorithm and network layer grid-connected reactive power optimization distribution control, and set virtual impedance algorithm L i =0.2mH, the proportional coefficient k of the grid-connected reactive power optimization distribution algorithm ipc =3.5, integration coefficient k iic =15. When the grid voltage suddenly drops by 5% at 1.5s, the average reactive power Q output by the two photovoltaic storage converters under the U0 condition is ie_av and output voltage amplitude d-axis command U ird The waveform is as Figure 6 As shown in the figure. The introduction of virtual impedance effectively improves the system stability, the output reactive power has no oscillation, and the power is evenly distributed in both dynamic and steady-state processes. In addition, the steady-state deviation of the output voltage amplitude of the two photovoltaic storage converters is about 1.05%, and the voltage control accuracy is high.

Claims

1. A hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters, the hierarchical collaborative optimization allocation method involving N photovoltaic storage converters and a distribution network, wherein: Each photovoltaic storage inverter is connected to a photovoltaic module and an energy storage battery on the DC side. The photovoltaic module and the energy storage battery are connected in parallel to provide a DC power supply for the photovoltaic storage inverter. The photovoltaic storage inverter outputs a three-phase AC voltage on the AC side and is connected to the distribution network through the line impedance. The method is characterized in that the hierarchical collaborative optimization allocation method includes local layer control and network layer grid-connected reactive support power optimization allocation control; the local layer control implements the voltage-controlled virtual synchronous generator algorithm and the virtual impedance algorithm according to the parameters of each photovoltaic storage converter itself; the network layer control is based on the local layer control, by constructing a distributed communication network of multiple photovoltaic storage converters, and a reactive optimization allocation strategy based on multi-agent consistency, to achieve the optimal allocation of grid-connected reactive power of multiple distributed photovoltaic storage converters; any one of the N photovoltaic storage converters is recorded as photovoltaic storage converter i, i∈[1,N], and the hierarchical collaborative optimization allocation method includes the following steps: Step 1: construct local layer control for the photovoltaic storage converter i based on its own sampling parameters and voltage-controlled virtual synchronous generator algorithm; The local layer control includes an output power calculation module, a voltage-controlled virtual synchronous generator algorithm module, a virtual impedance module, a voltage-current dual closed-loop control module, and a drive signal generation module; Step 1.1: The output power calculation module samples the three-phase output voltage u obtained by the photovoltaic storage converter i. ioa ,u iob ,u ioc And the three-phase output current i ioa ,i iob ,i ioc The output voltage dq component U in the synchronous rotating coordinate system is obtained by Clark transformation and Park transformation iod ,U ioq and the output current dq component I iod ,I ioq ; Then calculate the output instantaneous active power P ie And the output instantaneous reactive power Q ie , and are filtered by a first-order low-pass filter to obtain the output average active power P ie_av And the output average reactive power Q ie_av ; Step 1.2: The voltage-controlled virtual synchronous generator algorithm module outputs average active power P ie_av And the output average reactive power Q ie_av Perform active loop and reactive loop control. Specifically, calculate the reference angular frequency ω of the output voltage of the current beat according to the active loop control equation. i After that, the output voltage reference phase angle θ of the current beat is obtained by integration. i ,θ i =∫ω i ; The output voltage reference angular frequency ω of the current beat i and the reference phase angle θ of the output voltage i Used for the next beat calculation and form a closed loop; the voltage amplitude d-axis command U is calculated according to the reactive loop control equation ird ; Step 1.3, let the voltage amplitude q-axis command U irq =0, change U ird ,U irq and I iod ,I ioq As the input of the virtual impedance module, the dq axis voltage amplitude reference value U of the output voltage and current dual closed-loop control module is obtained through virtual impedance calculation. iRd ,U iRq ; Step 1.4: The voltage and current dual closed-loop control module uses the output voltage reference phase angle θ i And dq axis voltage amplitude reference value U iRd ,U iRq Perform double closed-loop control of the output voltage and output current of the photovoltaic storage converter i, and send its output as a modulation wave to the drive signal generation module to generate a drive signal for the power switch tube through a modulation strategy; Step 2: Multiple distributed photovoltaic storage converters build a network-layer grid-connected reactive power optimization allocation strategy based on a distributed communication network architecture and a consistency algorithm; The network layer grid-connected reactive power optimization allocation strategy includes a signal sending module, a signal receiving module and a grid-connected reactive power optimization allocation module of a photovoltaic storage converter, and includes the following steps: Step 2.1: According to the situation of distributed photovoltaic storage converters connected to the distribution network, a distributed communication network including N photovoltaic storage converters is constructed based on multi-agent distributed communication and consistency control theory. Specifically, each photovoltaic storage converter i in the distributed communication network can transmit its own reactive power-voltage droop coefficient k iU And the output average reactive power Q ie_av Transmitted to one or more neighboring photovoltaic storage converters j, and at the same time, the reactive power-voltage droop coefficient k sent by one or more neighboring photovoltaic storage converters j can be received through the signal receiving module jU And the output average reactive power Q je_av ; Assume that there are M neighboring solar-storage converters, j is the serial number of the neighboring solar-storage converter, j∈[1,M], M<N; Step 2.2, the grid-connected reactive power optimization allocation module of the photovoltaic storage converter i calculates the reactive power-voltage droop coefficient k of each neighboring photovoltaic storage converter j received jU And the output average reactive power Q je_av The product of ΔU cij , and calculate the total reactive power deviation ΔU of M neighboring solar-storage converters j cj , the calculation formulas are: Step 2.3: The grid-connected reactive power optimization allocation module of the photovoltaic storage converter i is based on its own reactive power-voltage droop coefficient k. iU And the output average reactive power Q ie_av , the total reactive power deviation ΔU obtained in step 2.2 cj Calculate the reactive power distribution compensation value ΔU ci , and the reactive power distribution compensation value ΔU ci It is superimposed into the reactive loop control equation in the local layer control of the photovoltaic storage converter i; The reactive power distribution compensation value ΔU ci The calculation formula is: where k ipc is the proportional coefficient of the grid-connected reactive power optimization allocation algorithm, k iic is the integral coefficient of the grid-connected reactive power optimization allocation algorithm, and s is the Laplace operator; Step 2.4, perform the calculations of steps 2.2 to 2.3 on each of the N photovoltaic storage converters i to obtain N reactive power distribution compensation values ​​ΔU ci , and are superimposed on the reactive loop control equation in the local layer control of the photovoltaic storage converter i, so as to optimize the distribution of the grid-connected reactive support power of multiple distributed photovoltaic storage converters according to their own rated capacity.

2. The hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters according to claim 1 is characterized in that: The transformation formulas of the Clark transformation and the Park transformation in step 1.1 are: Among them, θ i ′ is the reference phase angle of the output voltage of the photovoltaic storage converter obtained in the previous shot; The output instantaneous active power P ie And the output instantaneous reactive power Q ie The calculation formulas are: The output average active power P ie_av And the output average reactive power Q ie_av The calculation formulas are: Among them, T if is the filtering time constant of the first-order low-pass filter.

3. The hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters according to claim 1 is characterized in that: The active loop control equation in step 1.2 is: Where ω0 is the rated angular frequency of the system voltage, ω i ′ is the reference angular frequency of the output voltage of the photovoltaic converter obtained in the previous shot, k iω is the primary frequency modulation coefficient, P iref is the grid-connected active power instruction, D i is the damping coefficient, J i is the virtual inertia; The reactive loop control equation is: U ird =U0-k iU (Q ie_av -Q iref )+ΔU ci Among them, U0 is the rated output voltage amplitude of the system, k iU is the reactive power-voltage droop coefficient, Q iref is the grid-connected reactive power instruction, ΔU ci The reactive power distribution compensation value calculated by the network layer grid-connected reactive power optimization distribution strategy.

4. The hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters according to claim 1 is characterized in that: The dq axis voltage amplitude reference value U in step 1.3 iRd ,U iRq The calculation formula is: Among them, L i is the virtual inductance value of the photovoltaic storage converter i.

5. The hierarchical collaborative optimization allocation method for grid-connected reactive power of multiple distributed photovoltaic storage converters according to claim 1 is characterized in that: The implementation steps of step 1.4 are as follows: First, the output voltage and output current of the photovoltaic storage converter i are controlled in a double closed loop. Specifically, the output I of the dq axis output voltage outer loop control is calculated by the dq axis output voltage outer loop control equation. iRd ,I iRq , the output U of the dq axis output current inner loop control equation is obtained ipwmd ,U ipwmq ; The dq axis output voltage outer loop control equation and the dq axis output current inner loop control equation are respectively: Among them, k ipv is the proportional coefficient of the output voltage outer loop proportional-integral controller, k iiv is the integral coefficient of the output voltage outer loop proportional integral controller, k ipi is the proportional coefficient of the output current inner loop proportional-integral controller, k iii is the integral coefficient of the output current inner loop proportional-integral controller, s is the Laplace operator; Then the three-phase voltage modulation wave u is calculated ipwma ,u ipwmb ,u ipwmc , the calculation formula is:

Images