A distribution network voltage control method based on distributed photovoltaic active-reactive coordination

Abstract

This invention discloses a distribution network voltage control method based on distributed photovoltaic active-reactive coordination. By establishing a sensitivity matrix to analyze the impact of reactive / active injection on voltage stability, critical control gain constraints are derived to ensure the static stability of the distribution network voltage. A general model for photovoltaic inverter active / reactive coordinated control, applicable to different scenarios, is constructed based on a two-stage framework of intraday scheduling and real-time control. During the intraday scheduling phase, the control curve parameters of each photovoltaic inverter node are optimized with the goal of minimizing network loss costs, active power reduction costs, inverter voltage regulation costs, and voltage deviation. During the real-time control phase, active power reduction and reactive power regulation are dynamically coordinated based on the optimal control curve, with reactive power margin being prioritized to suppress voltage rise. When reactive power capacity is saturated, active power reduction is used to directly regulate power flow distribution. This invention balances grid security and economic efficiency, providing theoretical support and practical solutions for voltage control in high-penetration photovoltaic distribution networks.

Description

Technical Field

[0001] The present invention relates to the field of distribution network voltage control, and in particular to a distribution network voltage control method based on distributed photovoltaic active-reactive coordination. Background Art

[0002] As a clean and sustainable form of energy, photovoltaic power generation has seen its share of the power system increase annually. However, the output characteristics of photovoltaic power generation are significantly intermittent and fluctuating. This is especially true during midday on clear days, when photovoltaic active power reaches its peak, potentially causing reverse flow in the distribution network. This reverse flow alters the power distribution of traditional distribution networks, significantly changing the voltage profile and causing grid voltage overshoot. Voltage overshoot not only affects power quality but also threatens the safe operation of grid equipment. In severe cases, it can even cause equipment damage or grid collapse.

[0003] Traditional voltage regulation methods primarily rely on reactive power control, such as adjusting the inverter's reactive output or switching capacitor banks to maintain voltage stability. However, during midday, when photovoltaic power generation is high, the inverter's reactive power regulation capability is often limited, resulting in insufficient reactive power margin. Reactive power control alone cannot effectively address voltage overshooting. The effectiveness of reactive power control is particularly limited in distribution networks with a high proportion of photovoltaic integration. Therefore, relying solely on reactive power control is no longer sufficient to meet the requirements for safe grid operation, and more flexible and diverse control strategies must be considered.

[0004] Active power curtailment is gaining increasing attention as an effective voltage control measure. Active power curtailment proactively reduces the active power output of photovoltaic (PV) systems, minimizing the impact of adverse currents on the distribution network and thereby alleviating voltage overshoots. Compared to reactive power control, active power curtailment can directly alter the power flow distribution within the grid, resulting in a more significant voltage regulation effect. However, its implementation presents numerous challenges, such as determining the mapping relationship between inverter reactive power output and node voltage. Therefore, optimizing active power curtailment strategies and balancing grid security and economic efficiency have become key areas of research in PV grid-connected technology. Summary of the Invention

[0005] Purpose of the invention: The present invention aims to provide a distribution network voltage control method based on distributed photovoltaic active-reactive coordination to solve the voltage control problem of high-penetration photovoltaic distribution networks.

[0006] Technical solution: The distribution network voltage control method based on distributed photovoltaic active-reactive coordination described in the present invention includes the following steps:

[0007] (1) Based on the small disturbance analysis method, the sensitivity matrix of reactive / active power injection increment to the change of node voltage amplitude is derived according to the Jacobian matrix, and the control gain constraint conditions of each inverter are determined;

[0008] (2) Based on the distribution network voltage control framework of the PV inverter's active power (PV) and reactive power (QV) coordination in the intraday scheduling stage and the local real-time control stage, a general model for the PV and QV coordinated control of the PV inverter is constructed;

[0009] (3) Based on the critical control gains of each inverter obtained in step (1) and the general model of PV and QV coordinated control of photovoltaic inverters obtained in step (2), while considering the distribution network flow constraints and the life loss of photovoltaic inverters participating in reactive power compensation, with the goal of minimum network loss cost, minimum voltage deviation, minimum active power reduction and minimum reactive power regulation cost of the distribution network, a distribution network voltage control optimization model based on distributed photovoltaic active-reactive coordination is established to realize voltage regulation of photovoltaic inverters.

[0010] Furthermore, when considering the stability constraint of the PV inverter QV control, ignoring the active power increment, the sensitivity matrix of the reactive power injection increment to the node voltage amplitude change is obtained: for

[0011] ;

[0012] in, is the amplitude change of the node voltage, is the reactive injection increment of the photovoltaic inverter, N, M, K, and L represent the Jacobian matrix Elements in

[0013] ;

[0014] Introducing diagonal matrices , the reactive injection increment of the PV inverter for

[0015] ;

[0016] Among them, the diagonal matrix The diagonal elements of , diagonal elements Indicates the impact of reactive power changes on the local node voltage amplitude; is the inverter reactive-voltage control gain, is the voltage change;

[0017] Determine the critical control gain during reactive injection based on the small disturbance analysis method ;

[0018] When reactive power is injected,i The control gain of each inverter The constraints are as follows:

[0019] ;

[0020] in, for or .

[0021] Furthermore, ignoring the effect of reactive power and considering the active power of the photovoltaic inverter on voltage stability, the sensitivity matrix of the active power injection increment to the node voltage amplitude change is calculated. for

[0022] ;

[0023] in, is the amplitude change of the node voltage, is the active power injection increment of the photovoltaic inverter, N, M, K, and L represent the Jacobian matrix Elements in

[0024] ;

[0025] Introducing diagonal matrices , the active power injection increment of the PV inverter for

[0026] ;

[0027] Among them, the diagonal matrix The diagonal elements of , diagonal elements Indicates the impact of active power changes on the local node voltage amplitude; is the deviation between the real-time measured voltage and the expected voltage of the node at time t;

[0028] When active power is injected, the critical control gain is determined according to the small disturbance analysis method. ;

[0029] When active power is injected, i The control gain of each inverter The constraints are as follows:

[0030] ;

[0031] in, for or .

[0032] Furthermore, in step (2), during the intraday scheduling phase, based on the photovoltaic and load output data, network topology and parameter information of each scheduling period, with the goal of minimizing network loss and average node voltage deviation, the parameters of the local control curve of each photovoltaic inverter are optimized;

[0033] During the local real-time control phase, within each scheduling interval, the inverter adjusts its active / reactive output values ​​in real time according to the assigned PV and QV control curves, reducing active output while releasing reactive power margin to suppress voltage over-limit.

[0034] Furthermore, in step (2), the general model of the coordinated control of the photovoltaic inverter PV and QV is as follows:

[0035] Active power output value of photovoltaic inverter for

[0036] ;

[0037] in, The set value of photovoltaic active output, that is, the active output value after reduction; For the i The control gains of the inverters; the diagonal elements Indicates the impact of active power changes on the local node voltage amplitude; is the dead zone range; is the threshold; is the voltage amplitude of the i-th node at time t in scenario r.

[0038] Furthermore, the reactive power output value of the photovoltaic inverter for

[0039] ;

[0040] in, and are the upper and lower limits of the inverter reactive output respectively; The reactive output value of the inverter when it is within the voltage dead zone range; and Represent the control gains in the droop region of the curve; the diagonal elements Indicates the impact of reactive power changes on the local node voltage amplitude; is the dead zone range; is the voltage amplitude of the i-th node at time t in scenario r.

[0041] Furthermore, the photovoltaic active output setting value , upper limit of inverter reactive output , the lower limit of the inverter reactive output and the reactive output value of the inverter when the voltage dead zone range is The following constraints are met:

[0042] ;

[0043] ;

[0044] ;

[0045] ;

[0046] ;

[0047] in, is the inverter capacity; To predict photovoltaic output; The active power quota for photovoltaic output; is the load active power.

[0048] Furthermore, in step (3), the objective functions of minimizing network loss cost, minimizing active power reduction, minimizing inverter voltage regulation cost, and minimizing voltage deviation are as follows:

[0049] ;

[0050] ;

[0051] ;

[0052] ;

[0053] The cost of the life loss of the photovoltaic inverter due to its participation in reactive power compensation is calculated as follows:

[0054] ;

[0055] ;

[0056] ;

[0057] ;

[0058] in, is the network loss cost, It is a meritorious cost reduction. is the inverter voltage regulation cost, represents the average node voltage offset for all scenarios, is the weight factor; The total number of preset scenes for each time period; is the electricity price at time t; Indicates branch resistance; Indicates rated voltage; is the total number of nodes; and They represent the network active power loss and PV active power reduction respectively; 、 、 is a constant; and Line Active power and reactive power flowing through it; for Voltage amplitude of node t at time r.

[0059] Furthermore, step (3) satisfies the following constraints:

[0060] Inverter capacity constraints:

[0061] ;

[0062] in, is the reactive output value of the inverter, is the active power output value of the photovoltaic inverter, is the capacity of the PV inverter;

[0063] The active power balance equation is

[0064] ;

[0065] The reactive power balance equation is

[0066] ;

[0067] in, and Line Active power and reactive power flowing through it; is the active power from other branches flowing into node i, is the reactive power flowing into node i from other branches; and are active load and reactive load respectively; is the set of branches where power flows into node i;

[0068] The voltage relationship between the buses is as follows:

[0069] ;

[0070] in, is the voltage amplitude of node j, represents the branch resistance, is the branch reactance, Indicates rated voltage;

[0071] Node voltage amplitude constraint:

[0072] ;

[0073] in, and are the upper and lower limits of the node voltage amplitude respectively;

[0074] Constraints on line active power transmission capacity:

[0075] ;

[0076] in, is the maximum transmission capacity of the line active power;

[0077] PV control stability constraints:

[0078] ;

[0079] QV control stability constraints:

[0080] .

[0081] Beneficial effects: Compared with the existing technology, the significant advantages of the present invention are: 1. The present invention aims at the voltage control problem of high-penetration photovoltaic distribution network. By constructing a Jacobian matrix sensitivity model, the critical gain threshold of distributed photovoltaic active / reactive control is derived, and the voltage stability constraint mechanism is embedded in the optimization process, which solves the voltage instability dilemma caused by excessive control gain in traditional methods; 2. The present invention proposes a two-stage collaborative framework of "intraday scheduling-real-time control", which realizes the optimal inverter control curve parameter pre-decision for the economic efficiency of the entire network at the minute-level optimization layer, and breaks through the capacity limitation of a single regulation means through the collaborative strategy of reactive power priority-active power reduction at the second-level execution layer. Actual measurements show that the voltage over-limit suppression efficiency is improved, and the photovoltaic absorption rate increases year-on-year; 3. Intraday scheduling In the first stage, the goal is to minimize the network loss cost, active power reduction cost, inverter voltage regulation cost and voltage deviation, and optimize the control curve parameters of the photovoltaic inverters at each node; in the real-time control stage, active power reduction and reactive power regulation are dynamically coordinated based on the optimal control curve, and reactive power margin is preferentially used to suppress voltage rise. When the reactive power capacity is saturated, the power flow distribution is directly adjusted by active power reduction; 4. The present invention constructs a PV-QV collaborative model suitable for different scenarios, and uses piecewise linearization modeling to generate a universal control curve that takes into account fairness and dynamic response capabilities. It still ensures control accuracy in scenarios with drastic fluctuations in light intensity, reduces voltage over-limit compared to the traditional fixed slope mode, and constructs a four-dimensional objective function including network loss cost, active power reduction cost, equipment life loss and voltage deviation, and first passes the weighting coefficient Achieve Pareto optimization of safe and economical operation; 5. The present invention forms a complete technical chain from mechanism innovation, architecture design, model construction to engineering application, providing a voltage control paradigm with stability, economy and adaptability for the grid connection of a high proportion of new energy under the new power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 Schematic diagram of the structure of the distribution network voltage control framework of the photovoltaic inverter PV and QV collaboration of the present invention;

[0083] Figure 2 This is a graph showing the coordinated voltage control of the photovoltaic inverter PV and QV according to the present invention;

[0084] Figure 3 Schematic diagram of the network topology and photovoltaic locations of the test system of the embodiment;

[0085] Figure 4 The voltage comparison chart of Case 1 to Case 3 is shown below. DETAILED DESCRIPTION

[0086] The present invention will be further described below with reference to the accompanying drawings.

[0087] The distribution network voltage control method based on distributed photovoltaic active-reactive coordination according to the present invention comprises the following steps:

[0088] (1) Based on the small disturbance analysis method, the sensitivity matrices of reactive / active power injection increments to node voltage amplitude changes are derived according to the Jacobian matrix, and the control gain constraints of each inverter are determined.

[0089] The relationship between the power injection increment of the PV inverter and the node voltage change can be expressed as follows:

[0090] ;

[0091] When considering the stability constraint of reactive power-voltage QV control of photovoltaic inverter, the sensitivity matrix of reactive power injection increment to node voltage amplitude change can be expressed as

[0092] ;

[0093] Where, The diagonal elements of Represents the effect of reactive power change on the local node voltage amplitude; Ignoring the effect of active power increment on voltage fluctuation, according to formula (1) we get . Introducing the diagonal matrix , the diagonal elements of this matrix are The sensitivity matrix of the PV node reactive power injection increment to the node voltage amplitude change is: .

[0094] The reactive power injection increment of the PV inverter is

[0095] ;

[0096] Where, is the inverter reactive-voltage control gain, considering that the number of nodes with PV inverters is , so we can Simplified to , for dimensional column vector:

[0097] ;

[0098] Where, is the unit matrix. If we want to ensure the system voltage stability, All eigenvalues ​​of should lie within the unit circle. Defined as The e-th eigenvalue of for The e-th eigenvalue of . According to the characteristics of the eigenvalue, the following relationship is obtained:

[0099] ;

[0100] Where, It can be expressed as

[0101] ;

[0102] To ensure , reactive-voltage control gain The following constraints must be satisfied:

[0103] ;

[0104] Define a critical control gain , Optimize the limit of the reactive-voltage control function gain. If the control gain is greater than This will cause the distribution network voltage to become unstable. Before optimization, the critical control gain is calculated as

[0105] ;

[0106] To satisfy constraint (7), the control gain of each inverter is Should be less than the critical control gain Therefore, the voltage stability constraint can be expressed as

[0107] ;

[0108] Inequality (9) is a constraint in the optimization model, where Is a very small positive number, usually or . Critical control gain It can be calculated in advance and is a constant in the constraints.

[0109] Similarly, considering only the active power of the photovoltaic inverter on the voltage stability constraint, the sensitivity matrix of the active power injection increment to the node voltage amplitude change can be expressed as

[0110] ;

[0111] Where, The diagonal elements of represents the effect of active power change on the local node voltage amplitude; ignoring the effect of reactive power increment on voltage fluctuation, according to formula (1) we get The voltage stability constraint only considers the nodes where PV systems are installed. The number of PV nodes is counted as Therefore, we introduce the diagonal matrix , the diagonal elements of this matrix are The sensitivity matrix of the PV node active power injection increment to the node voltage amplitude change is: .

[0112] At time t, the deviation between the real-time measured voltage of the node and the expected voltage is It can be expressed as

[0113] ;

[0114] Where, is the measured voltage, is the expected voltage.

[0115] The active power injection increment of the photovoltaic inverter is

[0116] ;

[0117] Where, is the inverter active-voltage control gain, considering that the number of nodes with PV inverters is , so we can Simplified to , for -dimensional column vector.

[0118] ;

[0119] In order to maintain the network voltage stability, All eigenvalues ​​of should lie within the unit circle. Defined as The e-th eigenvalue of for The e-th eigenvalue of . According to the characteristics of the eigenvalue, the following relationship is obtained:

[0120] ;

[0121] Where, It can be expressed as

[0122] ;

[0123] To ensure , active power-voltage control gain The following constraints must be satisfied:

[0124] ;

[0125] Define a critical control gain , To optimize the limit of the active power-voltage control function gain, if the control gain is greater than This will cause the distribution network voltage to become unstable. Before optimization, the critical control gain is calculated as

[0126] ;

[0127] To ensure that equation (16) is satisfied, the control gain of each inverter is Should be less than the critical control gain Therefore, the voltage stability constraint can be expressed as

[0128] ;

[0129] In the formula is a very small positive number, usually or , the active-voltage stability constraint can be brought into the optimization model.

[0130] (2) Based on the distribution network voltage control framework of photovoltaic inverter PV and QV coordination in the intraday scheduling stage and local real-time control stage, a general model of photovoltaic inverter PV and QV coordinated control is constructed.

[0131] The control framework consists of an intraday scheduling phase (minute-level) and a local real-time control phase (second-level). The intraday scheduling phase optimizes the parameters of each PV inverter's local control curve based on PV and load output data, network topology, and parameter information for each scheduling period, with the goal of minimizing network losses and average node voltage deviation. Within each scheduling interval, the inverter adjusts its active and reactive output values ​​in real time according to the assigned PV and QV control curves, reducing active output while releasing reactive power margin to further suppress voltage overshoots.

[0132] (3) Based on the critical control gains of each inverter obtained in step (1) and the general model of PV and QV coordinated control of photovoltaic inverters obtained in step (2), while considering the distribution network flow constraints and the life loss of photovoltaic inverters participating in reactive power compensation, with the goal of minimum network loss cost, minimum voltage deviation, minimum active power reduction and minimum reactive power regulation cost of the distribution network, a distribution network voltage control optimization model based on distributed photovoltaic active-reactive coordination is established to realize voltage regulation of photovoltaic inverters.

[0133] The relationship between node active power, line power and voltage phase angle is as follows:

[0134] ;

[0135] Where, and They represent the node injected active power and line active power respectively; is the voltage phase angle; and They are the node susceptance matrix and the line-node susceptance matrix, which are expressed as follows:

[0136] ;

[0137] Where N is the number of nodes in the distribution network, is the line reactance between node i and node j.

[0138] From formula (19), the relationship between the node injection power and the line power flow is as follows:

[0139] ;

[0140] Where P is the PTDF matrix and the dimension is the number of branches Number of nodes. Dimension is the number of branches The number of photovoltaic nodes indicates the influence factor coefficient of photovoltaic nodes on the power flow of each branch.

[0141] At noon, when distributed photovoltaic power generation is high, it causes reverse power flow in the distribution network. The Jacobian matrix in polar coordinates shows that the node voltage is affected by the node injection power. Therefore, this section introduces a voltage sensitivity matrix to reflect the impact of the reverse power flow caused by photovoltaics on the node voltage amplitude. The correction equation in polar coordinates is as follows:

[0142] ;

[0143] in, is the Jacobian matrix, and the photovoltaic node voltage-active power sensitivity matrix can be expressed as

[0144] ;

[0145] in, is the voltage amplitude of node m, is the active power deviation of node i; Indicates the phase angle deviation between node i and node j .

[0146] Introducing the impact matrix of photovoltaic access points on line flow and node voltage and :

[0147] ;

[0148] The approximate ideal solution is used to calculate the ranking of each distributed photovoltaic entity's impact on line flow and node voltage, and finally the weighted average is used to obtain the ranking of each distributed photovoltaic entity's comprehensive impact on the distribution network. The calculation process of this method is as follows:

[0149] Matrix normalization, Take the matrix as an example:

[0150] ;

[0151] Where K is the number of photovoltaic power plants in the distribution network; L is the number of branches. Matrix normalization yields matrix Z, elements The normalization formula and the normalized matrix are expressed as follows:

[0152] ;

[0153] Define the best and worst solutions respectively and :

[0154] ;

[0155] Calculate the distance between the i-th node and the optimal and worst solutions and :

[0156] ;

[0157] Calculate the overall impact score of each photovoltaic power plant on the distribution network flow :

[0158] ;

[0159] Similarly, the overall impact score of K photovoltaic entities on node voltage can be obtained: .

[0160] After obtaining the overall impact scores of K photovoltaic entities on the distribution network flow and node voltage, the average of these two score matrices is taken to obtain the overall comprehensive impact score matrix M of the photovoltaic entities on the distribution network:

[0161] ;

[0162] The above-mentioned comprehensive score matrix is ​​the basis for allocating the ratio of the power reverse transmission quota to the distributed photovoltaic entities, that is, the allocation is based on the proportion of the elements of the matrix M. When the allocated quota ensures that the distribution network does not have overvoltage at that moment, the reverse transmission power quota value Q of K distributed photovoltaic entities can be obtained:

[0163] ;

[0164] A node's reverse power quota is allocated based on its impact on power flow and voltage at other nodes. When a node's reverse power exceeds the quota, the system faces overvoltage risk, and the node bears the responsibility for the overvoltage. Furthermore, this reverse power quota is net reverse power, meaning it represents the power quota that enters the grid after self-production and sales. Therefore, the distribution network can control PV power curtailment to ensure that the reverse power does not exceed the quota.

[0165] The general model of photovoltaic inverter PV and QV coordinated control is as follows:

[0166] The main parameters of the PV curve include the dead zone range and control gain , PV inverter active power output value for

[0167] ;

[0168] Where, The photovoltaic active output setting value, that is, the active output value after reduction. When the inverter further reduces the active power output to suppress the voltage rise. The maximum allowable voltage is generally 1.05pu.

[0169] Similarly, the QV control curve achieves voltage support by adjusting reactive output, and its parameters include dead band range , control gain and and voltage thresholds and .

[0170] ;

[0171] Where, and are the upper and lower limits of the inverter reactive output respectively; is the reactive output value of the inverter when it is in the voltage dead zone; these three parameters are variables to be optimized. Generally, the minimum allowable voltage is 0.95pu; is the parameter to be optimized. If the node voltage amplitude is less than The inverter generates reactive power at its maximum remaining capacity; To be optimized parameters, if the node voltage amplitude is greater than The inverter absorbs reactive power with its maximum remaining capacity; The maximum allowable voltage is generally taken as 1.05pu.

[0172] In addition, the PV and QV control parameters need to meet the inverter capacity constraints:

[0173] ;

[0174] ;

[0175] ;

[0176] ;

[0177] ;

[0178] Where, is the inverter capacity; To predict photovoltaic output; The active power quota for photovoltaic output; is the load active power.

[0179] By optimizing the local control curve parameters of the PV inverter, the system's network loss costs, active power reduction costs, inverter voltage regulation costs, and node voltage deviation are minimized, including:

[0180] ;

[0181] ;

[0182] ;

[0183] ;

[0184] ;

[0185] ;

[0186] ;

[0187] ;

[0188] ;

[0189] ;

[0190] ;

[0191] ;

[0192] ;

[0193] ;

[0194] ;

[0195] ;

[0196] The objective function (39) aims to minimize the network loss cost, the active power reduction, the inverter voltage regulation cost and the voltage deviation, where Network loss cost, It is a meritorious cost reduction. is the inverter voltage regulation cost, represents the average node voltage offset for all scenarios, is the weight factor. Equations (41)-(46) respectively give the calculation methods of various costs and voltage deviations in the objective function, where R is the total number of preset scenarios in each time period; is the electricity price at time t; Indicates branch resistance; Indicates the rated voltage; N is the total number of nodes; and denote the network active power loss and photovoltaic active power reduction respectively; Equation (43) is used to calculate the life loss cost of the photovoltaic inverter due to its participation in reactive power compensation, where 、 and is a constant; and are the active power and reactive power flowing through line ij respectively; is the voltage amplitude of node i at time t in scenario r. Equation (47) is the inverter capacity constraint; Equations (48) and (49) are the active power balance equation and reactive power balance equation, respectively. Predict active power output for photovoltaics; is the reactive output value of the inverter; and are active load and reactive load respectively; in the branch power loss in the Distflow model, the quadratic term is much smaller than the linear term, and the influence of the quadratic term can be ignored. Therefore, the voltage relationship between buses can be expressed as a linear model, as shown in Equation (50). Equation (51) is the node voltage amplitude constraint, and are the upper and lower limits of the node voltage amplitude respectively. Equation (52) is the constraint of the line active power transmission capacity, is the maximum transmission capacity of the line active power. Equations (53) and (54) are the PV and QV control stability constraints.

[0197] In order to verify the effect of the present invention, eight photovoltaic units were connected to the grid. The access capacity and location information of the photovoltaic and inverter units are shown in Table 1. Figure 3 To test the network topology and PV location of the system, set the capacity of the PV inverter to 1.1 times the rated active power of the PV. This way, the inverter can provide reactive power support even when the inverter outputs rated active power. The required voltage range is set to 0.95pu to 1.05pu. and The power factor is set to 0.95 pu and 1.05 pu, respectively. The weight coefficient is set to 0.5. The proposed method's advantages in improving system performance are evaluated by comparing key indicators such as voltage distribution, network losses, and active power reduction before and after optimization. The daily local PV and QV control parameter instruction scheduling interval is set to 1 hour, and the number of source and load scenarios within the scheduling interval is expected. Finally, the effectiveness of the proposed control strategy is verified through three different control cases.

[0198] Case 1: There is no local control strategy on the photovoltaic inverter, and the initial operating state is obtained only by "minute-level" scheduling.

[0199] Case 2: The optimization model is solved by calling the solver. The local control parameters of the photovoltaic inverter PV and QV are obtained by solving the optimization model to adjust the reactive power of the distributed photovoltaic.

[0200] Case 3: By solving the optimization model of the present invention, the distributed photovoltaic PV and QV local control parameters are obtained to adjust the active power and reactive power of the inverter to achieve global optimization.

[0201] Table 1

[0202] ;

[0203] The voltage of the test system is controlled by the method of the present invention, such as Figure 4 The voltage distribution for PV nodes 18 and 32 is shown. The high proportion of distributed PV connected to the distribution network resulted in frequent voltage fluctuations at the nodes connected to distributed PV in Case 1, with severe voltage deviations and even over-limits. Take node 18 as an example. Reverse power flow occurred across the entire feeder between 12:00 and 14:00, caused by high PV active power output and low load levels. This caused the voltage at node 18 to exceed 1.05 pu. Using the control strategy proposed in Option 2, the PV inverter adjusted reactive power in real time according to the local QV control curve, improving the system voltage distribution. However, during this period, the inverter's reactive power margin was insufficient, providing insufficient support for the system voltage. Voltage fluctuations were significant, and over-limits persisted. Option 3 employs a comprehensive objective function for the local control strategy of the distributed PV inverter, simultaneously considering minimizing network loss costs, minimizing active power curtailment costs, and minimizing average node voltage deviation. Therefore, when a relatively severe voltage deviation occurred at node 18, a PV active power curtailment strategy was implemented between 12:00 and 14:00 to achieve comprehensive optimization of the objective function. At the same time, the objective function takes into account the minimum active power reduction of distributed photovoltaics. When reactive power is sufficient, frequent active power reduction is avoided to achieve the maximum utilization of photovoltaic resources.

Claims

1. A distribution network voltage control method based on distributed photovoltaic active-reactive coordination, characterized in that: The following steps are involved: (1) Based on the small disturbance analysis method, the sensitivity matrix of reactive / active power injection increment to the change of node voltage amplitude is derived according to the Jacobian matrix, and the control gain constraint conditions of each inverter are determined; (2) Based on the distribution network voltage control framework of the PV and QV coordinated active power of the PV inverter in the intraday scheduling stage and the local real-time control stage, a general model for the PV and QV coordinated control of the PV inverter is constructed; (3) Based on the critical control gains of each inverter obtained in step (1) and the general model of PV and QV coordinated control of photovoltaic inverters obtained in step (2), while considering the power flow constraints of the distribution network and the life loss of photovoltaic inverters participating in reactive power compensation, with the goal of minimum network loss cost, minimum voltage deviation, minimum active power reduction and minimum reactive power regulation cost of the distribution network, a distribution network voltage control optimization model based on distributed photovoltaic active-reactive coordination is established to achieve voltage regulation of photovoltaic inverters; In step (2), the general model of cooperative control of photovoltaic inverter PV and QV is as follows: PV inverter active power output value for ； in, The set value of photovoltaic active output, that is, the active output value after reduction; For the i The control gains of the inverters; the diagonal elements Indicates the impact of active power changes on the local node voltage amplitude; is the dead zone range; is the threshold; is the voltage amplitude of the i-th node at time t in scenario r; is the maximum allowable voltage; Reactive power output value of photovoltaic inverter for ； in, and are the upper and lower limits of the inverter reactive output respectively; The reactive output value of the inverter when it is within the voltage dead zone range; and Represent the control gains in the droop region of the curve; the diagonal elements Indicates the impact of reactive power changes on the local node voltage amplitude; is the dead zone range; is the voltage amplitude of the i-th node at time t in scenario r; is the parameter to be optimized; Photovoltaic active output setting value , upper limit of inverter reactive output , the lower limit of the inverter reactive output and the reactive output value of the inverter when the voltage dead zone range is The following constraints are met: ； ； ； ； ； in, is the inverter capacity; To predict photovoltaic output; The active power quota for photovoltaic output; is the load active power; is the reverse power quota value of the i-th distributed photovoltaic entity; In step (3), the objective functions of minimizing network loss cost, active power reduction, inverter voltage regulation cost, and voltage deviation are as follows: ； ； ； ； The cost of the life loss of the photovoltaic inverter due to its participation in reactive power compensation is calculated as follows: ； ； ； ； in, is the network loss cost, It is a meritorious cost reduction. is the inverter voltage regulation cost, represents the average node voltage offset for all scenarios, is the weight factor; The total number of preset scenes for each time period; is the electricity price at time t; Indicates branch resistance; Indicates rated voltage; is the total number of nodes; and They represent the network active power loss and PV active power reduction respectively; 、 、 is a constant; and Line Active power and reactive power flowing through it; for The voltage amplitude of the node at time t in scenario r; Step (3) satisfies the following constraints: Inverter capacity constraints: ； in, is the reactive output value of the inverter, is the active power output value of the photovoltaic inverter, is the capacity of the PV inverter; The active power balance equation is ； The reactive power balance equation is ； in, and Line Active power and reactive power flowing through it; is the active power flowing into node i from other branches h, is the reactive power flowing into node i from other branches h; and are active load and reactive load respectively; is the set of branches where power flows into node i; The voltage relationship between the buses is as follows: ； in, is the voltage amplitude of node j, represents the branch resistance, is the branch reactance, Indicates rated voltage; Node voltage amplitude constraint: ; in, and are the upper and lower limits of the node voltage amplitude respectively; Constraints on line active power transmission capacity: ； in, is the maximum transmission capacity of the line active power; PV control stability constraints: ； QV control stability constraints: ； in, When active power is injected i The control gain of each inverter; is the critical control gain during active power injection; for or ; For the i The control gain of each inverter, It is the critical control gain during reactive power injection.

2. The distribution network voltage control method based on distributed photovoltaic active-reactive coordination according to claim 1, characterized in that: When considering the stability constraint of the photovoltaic inverter QV control, ignoring the active power increment, the sensitivity matrix of the reactive power injection increment to the node voltage amplitude change is obtained: for ； in, is the amplitude change of the node voltage, is the reactive injection increment of the photovoltaic inverter, N, M, K, and L represent the Jacobian matrix Elements in ； Introducing diagonal matrices , the reactive injection increment of the PV inverter for ； Among them, the diagonal matrix The diagonal elements of , diagonal elements Indicates the impact of reactive power changes on the local node voltage amplitude; is the inverter reactive-voltage control gain, is the voltage change; Determine the critical control gain during reactive injection based on the small disturbance analysis method ; When reactive power is injected, i The control gain of each inverter The constraints are as follows: ； in, for or .

3. The distribution network voltage control method based on distributed photovoltaic active-reactive coordination according to claim 1, characterized in that: Ignoring the effect of reactive power and considering the active power of the photovoltaic inverter on voltage stability, the sensitivity matrix of the active power injection increment to the node voltage amplitude change is calculated. for ； in, is the amplitude change of the node voltage, is the active power injection increment of the photovoltaic inverter, N, M, K, and L represent the Jacobian matrix Elements in ； Introducing diagonal matrices , the active power injection increment of the PV inverter for ； Among them, the diagonal matrix The diagonal elements of , diagonal elements Indicates the impact of active power changes on the local node voltage amplitude; is the deviation between the real-time measured voltage and the expected voltage of the node at time t; When active power is injected, the critical control gain is determined according to the small disturbance analysis method. ; is the active power-voltage control gain; When active power is injected, i The control gain of each inverter The constraints are as follows: ； in, for or .

4. The distribution network voltage control method based on distributed photovoltaic active-reactive coordination according to any one of claims 2 or 3, characterized in that: In step (2), during the intraday scheduling phase, the local control curve parameters of each photovoltaic inverter are optimized based on the photovoltaic and load output data, network topology and parameter information of each scheduling period, with the goal of minimizing network loss and average node voltage deviation; During the local real-time control phase, within each scheduling interval, the inverter adjusts its active / reactive output values ​​in real time according to the assigned PV and QV control curves, reducing active output while releasing reactive power margin to suppress voltage over-limit.

Images