Weak communication distributed photovoltaic inverter reactive power adaptive control method

Abstract

The invention discloses a weak communication distributed photovoltaic inverter reactive power adaptive control method, which relates to the field of electric power, and comprises the following steps: S1, defining problems and parameters: determining power limitation, power factor limitation and voltage limitation of a photovoltaic inverter; s2, discretizing the output level: discretizing the photovoltaic output level into a plurality of nodes to form a first-order spline; s3, constructing an optimization problem: defining an objective function, minimizing the use of reactive power and the reduction amount of active power and adding constraint conditions, S4, solving the optimization problem: solving the problem by using a convex optimization solver to obtain nodes of a piecewise linear control function; s5, application of the control function: applying the optimized control function to the photovoltaic inverter; and S6, regularly updating parameters. The method is quick in response, reactive power use is optimized, the convex optimization problem is solved, the method adapts to an unbalanced power grid through central coordination control, voltage support is improved, and system loss is reduced.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of electric power, and specifically to a reactive adaptive control method for a weak-communication distributed photovoltaic inverter. Background Art

[0002] The increasing penetration of photovoltaics (PV) poses challenges to the operation of future power grids. During periods of low demand, high active power injection from PV inverters may cause transformer overload or grid voltage to rise to critical levels. Voltage rise is often the main limiting factor preventing the increase of PV power generation in low voltage (LV) networks. PV inverters provide their rated or maximum power for a small portion of the time, so the remaining power capacity in the inverter can be used for voltage control. Voltage control through reactive power control and active power reduction has proven to be an economically advantageous solution. However, most of the current reactive power control relies on remote instructions from the controller for control, which is difficult to achieve real-time control when communication is interrupted or communication conditions are poor. Therefore, a set of local reactive power control methods is an effective solution to the voltage limit problem caused by PV access under weak communication conditions.

[0003] A photovoltaic inverter control method and system with adaptive reactive power compensation is provided according to application number CN202210872199.1. The electrical part of the present invention is connected to the photovoltaic DC-AC converter by a photovoltaic array through a cable. The photovoltaic DC-AC converter converts the direct current output by the photovoltaic array into alternating current and transmits it to the AC bus; the control part is a self-learning optimization controller that communicates with the environmental data detection system to obtain the environmental irradiance and ambient temperature data collected in real time by the environmental detection system; the current and voltage data output by the photovoltaic array are obtained through current and voltage transformers; the voltage data of the AC bus is obtained through a voltage transformer; the self-learning optimization controller communicates voltage and power information with the data acquisition device of the adjacent network node as the calculation input data of the self-learning optimization controller, and the calculation result is sent to the photovoltaic DC-AC converter to control the operation. It can effectively avoid the problems of voltage over-limit and excessive deviation of controller parameters.

[0004] In the prior art, methods are based on local voltage measurement, local power production measurement or a combination of both; local measurements are used to define the voltage reference of the reactive voltage controller; when the local parameters are not adjusted to the grid structure, these local control methods may lead to a large amount of unnecessary reactive power consumption or insufficient voltage support; most of these local controllers are characterized by only one parameter and cannot achieve the best under different conditions of the system. Summary of the invention

[0005] Based on this, the purpose of the present invention is to provide a reactive adaptive control method for a weak communication distributed photovoltaic inverter to solve the technical problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A reactive power adaptive control method for a weak communication distributed photovoltaic inverter, the reactive power adaptive control method for the photovoltaic inverter comprising the following steps:

[0008] S1. Define the problem and parameters: Determine the power limit, power factor limit and voltage limit of the PV inverter;

[0009] S2, discretized output level: discretize the photovoltaic output level into several nodes to form a first-order spline;

[0010] S3. Construct the optimization problem: define the objective function to minimize the use of reactive power and the reduction of active power; add constraints, including power limit, power factor limit and voltage limit;

[0011] S4. Solve the optimization problem: use a convex optimization solver to solve the problem and obtain the nodes of the piecewise linear control function;

[0012] S5, applying the control function: applying the optimized control function to the photovoltaic inverter;

[0013] S6. Update parameters regularly: Set the update time and update the control parameters regularly to optimize the PV inverter in time.

[0014] Preferably, the objective function in said S3, constructing the optimization problem, is as follows:

[0015]

[0016] in,

[0017] represents the PV power cut at node p by the inverter at node i;

[0018] represents the reactive power at node p generated by the inverter at node i;

[0019] w: represents the weight factor of the reactive power term in the objective function;

[0020] P: represents the node index of the control function;

[0021] n nodes : Represents the total number of nodes in the distribution network.

[0022] Preferably, the power limit in the constraint condition in S3, constructing the optimization problem is as follows:

[0023]

[0024] in,

[0025] represents the photovoltaic power generated by the inverter at node p at node i;

[0026] represents the nominal apparent power rating of the inverter at node i;

[0027] Active power reduction constraints:

[0028] Preferably, the power factor restriction in the constraint conditions in the S3, constructing the optimization problem is as follows:

[0029]

[0030] in,

[0031] α: Constant defining the permissible ratio between reactive power and active power.

[0032] Preferably, the voltage limit in the constraint condition in the S3, constructing the optimization problem is as follows:

[0033] V min ≤|V p.i |≤V max

[0034] in,

[0035] |V p.i |: represents the voltage at node i;

[0036] V min : Indicates the minimum allowable voltage;

[0037] V max : Indicates the maximum allowable voltage.

[0038] Preferably, the voltage estimation formula is as follows:

[0039]

[0040] in,

[0041] represents the voltage at node i due to the load in the previous time step;

[0042] represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the photovoltaic power generated by the inverter at node p at node m;

[0043] represents the PV power cut at node p by the inverter at node m;

[0044] represents the reactive power at node p generated by the inverter at node m.

[0045] Preferably, the active power reduction constraint further includes a fair active power reduction constraint:

[0046]

[0047] in,

[0048] represents the nominal apparent power rating of the inverter at node j;

[0049] represents the PV power curtailed at node p by the inverter at node j.

[0050] Preferably, a constraint of maximum apparent power is also included, and the formula is as follows:

[0051]

[0052] in,

[0053] represents the active power of household load connected to node i;

[0054] represents the reactive power of household loads connected to node i;

[0055] S max : Indicates the maximum apparent power passing through the transformer.

[0056] Preferably, the goal of constructing the optimization problem in S3 is to minimize the reduction of active power by adjusting local control parameters, while using reactive power to support voltage control when necessary.

[0057] Preferably, in the objective function, i represents a node index; the number of nodes of the function is limited and is set to 10 to ensure that the computational complexity is controllable.

[0058] In summary, the present invention mainly has the following beneficial effects:

[0059] The reactive power adaptive control method of photovoltaic inverter is compared with the existing technology.

[0060] Fast response: The local controller makes adjustments based on real-time PV power measurements and can quickly respond to changes in PV power generation, especially in the case of rapid changes caused by cloud movement;

[0061] Optimize reactive power usage: Regularly (every 15 minutes) readjust local control parameters through a central optimization algorithm to ensure near-optimal use of reactive power and minimize active power reduction when PV power generation changes;

[0062] Convex optimization problem: The optimization problem is formulated as a convex optimization problem, which can be solved efficiently and ensures that the optimal solution is found quickly under complex grid conditions;

[0063] Reduced system losses: Compared with methods that rely solely on local control, central optimization can more reasonably allocate reactive power and active power reduction, reducing system losses;

[0064] Adapting to unbalanced power grid: In a three-phase four-wire low-voltage distribution network, the proposed method can take into account the neutral point drift effect, better regulate the voltage, and avoid the problem of other phase voltage increases due to reactive power absorption;

[0065] Improve voltage support: Through central coordinated control, reactive power can be injected or absorbed in one phase to avoid overvoltage problems in another phase, making reactive voltage support more uniform and reducing system losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 It is an overall flow chart of the reactive power adaptive control method of the photovoltaic inverter of the present invention;

[0067] Figure 2 It is a schematic diagram comparing active power reduction under the condition of no control of the present invention and under the control strategy of the present patent;

[0068] Figure 3 The present invention is a diagram showing the reduction of active power of each household under reactive voltage control and unfair conditions. DETAILED DESCRIPTION

[0069] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0070] Example

[0071] like Figure 1 , 2 As shown in , 3, the photovoltaic inverter reactive power adaptive control method comprises the following steps:

[0072] S1. Define the problem and parameters: Determine the power limit, power factor limit and voltage limit of the PV inverter;

[0073] S2, discretized output level: discretize the photovoltaic output level into several nodes to form a first-order spline;

[0074] S3. Construct the optimization problem: define the objective function to minimize the use of reactive power and the reduction of active power; add constraints, including power limit, power factor limit and voltage limit;

[0075] S4. Solve the optimization problem: use a convex optimization solver to solve the problem and obtain the nodes of the piecewise linear control function;

[0076] S5, applying the control function: applying the optimized control function to the photovoltaic inverter;

[0077] S6. Update parameters regularly: set the update time, update the control parameters regularly, and optimize the photovoltaic inverter in time;

[0078] It should be noted that, in this embodiment, the present invention is a centralized optimization algorithm that optimizes the parameters of the local controllers every 15 minutes; after the local controllers are optimized, these local controllers can achieve near-optimal performance even if PV production changes due to cloud movement;

[0079] An existing three-phase four-wire radial distribution system in a western city is used as the simulation network; it has a TT grounding arrangement; the network consists of 62 users and is a semi-urban reference network; it has been used to study the impact of increased photovoltaic power generation in public low-voltage distribution networks; the study shows that it can cope with PV installations installed on 50% of the roofs without transformer overload; all main feeder cables are of type EAXVB 1kV4×150mm2, and the cable between node 2 and node 4 is of type EAXVB1kV4×95mm2;

[0080] The impedance values ​​are calculated assuming an operating temperature of 45°C, all households have single-phase connections, and the nominal line-to-neutral voltage is 230 V, evenly distributed among the three phases; considering that the voltage on the secondary side of the transformer is 235 V when there is no load, this can be considered as a typical low-voltage transformer tap setting to avoid low voltage at the end of the feeder; statistically representative residential load curves are provided for load flow simulation; a constant power load model is assumed, and the process of generating these load curves is described in the literature; no reactive power consumption data of the loads is provided, so the reactive power of the loads is ignored in the simulation.

[0081] like Figure 1 , 2 As shown in , S3, the objective function in the optimization problem is as follows:

[0082]

[0083] in,

[0084] represents the PV power cut at node p by the inverter at node i;

[0085] represents the reactive power at node p generated by the inverter at node i;

[0086] w: represents the weight factor of the reactive power term in the objective function;

[0087] P: represents the node index of the control function;

[0088] n nodes : represents the total number of nodes in the distribution network;

[0089] S3. The power limit in the constraints of the optimization problem is as follows:

[0090]

[0091] in,

[0092] represents the photovoltaic power generated by the inverter at node p at node i;

[0093] represents the nominal apparent power rating of the inverter at node i;

[0094] Active power reduction constraints: S3. The power factor limit in the constraints of the optimization problem is as follows:

[0095]

[0096] in,

[0097] α: constant defining the permissible ratio between reactive power and active power;

[0098] S3. The voltage limit in the constraint conditions of the optimization problem is as follows:

[0099] V min ≤|V p.i |≤V max

[0100] in,

[0101] |V p.i |: represents the voltage at node i;

[0102] V min : Indicates the minimum allowable voltage;

[0103] V max : Indicates the maximum allowable voltage;

[0104] The voltage estimation formula is as follows:

[0105]

[0106] in,

[0107] represents the voltage at node i due to the load in the previous time step;

[0108] represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the photovoltaic power generated by the inverter at node p at node m; represents the PV power cut at node p by the inverter at node m;

[0109] represents the reactive power at node p by the inverter at node m;

[0110] The active power reduction constraint also includes fair active power reduction constraints:

[0111]

[0112] in,

[0113] represents the nominal apparent power rating of the inverter at node j;

[0114] represents the PV power cut by the inverter at node p at node j;

[0115] It also includes the constraint of maximum apparent power, which is given by the following formula:

[0116]

[0117] in,

[0118] represents the active power of household load connected to node i;

[0119] represents the reactive power of household loads connected to node i;

[0120] S max : Indicates the maximum apparent power passing through the transformer.

[0121] It should be noted that in this example, the PV curve is measured in a fixed rooftop PV installation with a 1-minute time step, assuming that all PV panels are in the same direction, the nominal capacity of the PV power generation unit is 5kVA, and it is not evenly distributed in the three phases. All households connected to the U phase have a PV installation, while only 50% of the other households have a PV installation, and the voltage is limited to ±10% of the nominal voltage. The time step used in the simulation is 1 minute;

[0122] According to the partial solution code:

[0123] import cvxpy as cp

[0124] #Define problem data and parameters

[0125] n=10#Number of nodes

[0126] P_PV=cp.Parameter(n,nonneg=True)#Photovoltaic power generation

[0127] S_Nom = cp.Parameter (n, positive = True) # nominal apparent power

[0128] V_min, V_max = 0.95, 1.05 # voltage limit

[0129] #Define decision variables

[0130] P_curt = cp.Variable (n, nonneg = True) # Reduced active power

[0131] Q_PV = cp.Variable(n) #Generated reactive power

[0132] #Define the objective function

[0133] objective=cp.Minimize(cp.sum_squares(Q_PV)+cp.sum(P_curt))

[0134] #Add constraints

[0135] constraints=[

[0136] cp.square(P_PV-P_curt)+cp.square(Q_PV)<=

[0137] cp.square(S_Nom),

[0138] V_min<=V_nom+cp.matmul(V_sens,P_PV-P_curt+Q_PV)<=V_max,

[0139] #More constraints... ]

[0141] #Define the optimization problem

[0142] prob=cp.Problem(objective,constraints)

[0143] #Specify the solver and solve the problem

[0144] result=prob.solve(solver=cp.MOSEK)

[0145] print("Optimal cost:",result)

[0146] print("Optimal curtailment:",P_curt.value)

[0147] print("Optimal reactive power:",Q_PV.value).

[0148] like Figure 1 , 2 As shown in 3, preferably, S3, the goal of constructing the optimization problem is to minimize the reduction of active power by adjusting local control parameters, while using reactive power to support voltage control when necessary; in the objective function, i represents the node index; the number of nodes of the spline function is limited,

[0149] The number is set to 10 to ensure that the computational complexity is controllable.

[0150] It should be noted that in this embodiment, the proposed distributed photovoltaic inverter reactive power local control function is defined as a piecewise linear function (spline function); the control parameters of each inverter are readjusted every 15 minutes by the central optimization algorithm to ensure near-optimal performance under varying PV power generation levels;

[0151] Comparison of active power reduction without control and with the patented control strategy, as well as active power reduction per household under reactive voltage control and unfairness,

[0152] It can be found that one advantage of the present methods is that they can require reactive power injection or absorption in one phase to avoid excessive voltage in another phase; therefore, the reactive power is more evenly distributed in the three phases, resulting in reduced losses. At the same time, the active power reduction for users is fair, avoiding the occurrence of three-phase imbalance problems in the substation area.

[0153] The working principle of the present invention is:

[0154] The proposed distributed PV inverter reactive power local control function is defined as a piecewise linear function (spline function); the control parameters of each inverter are readjusted every 15 minutes by a central optimization algorithm to ensure near-optimal performance under varying PV power generation levels;

[0155] The present invention is a centralized optimization algorithm that optimizes the parameters of local controllers every 15 minutes; after the local controllers are optimized, these local controllers can achieve near-optimal performance even if PV production changes due to cloud movement;

[0156] An existing three-phase four-wire radial distribution system in a western city is used as the simulation network; it has a TT grounding arrangement; the network consists of 62 users and is a semi-urban reference network; it has been used to study the impact of increased photovoltaic power generation in public low-voltage distribution networks; the study shows that it can cope with PV installations installed on 50% of the roofs without transformer overload; all main feeder cables are of type EAXVB 1kV4×150mm2, and the cable between node 2 and node 4 is of type EAXVB1kV4×95mm2;

[0157] The impedance values ​​are calculated assuming an operating temperature of 45°C. All households have single-phase connections and a nominal line-to-neutral voltage of 230 V, evenly distributed among the three phases. Considering that the voltage on the secondary side of the transformer is 235 V when there is no load, this can be considered a typical low-voltage transformer tap setting to avoid low voltage at the end of the feeder. Statistically representative residential load curves are provided for load flow simulation. A constant power load model is assumed, and the process of generating these load curves is described in the literature. No reactive power consumption data of the loads is provided, so the reactive power of the loads is ignored in the simulation.

[0158] The PV curves are measured in a fixed rooftop PV installation with a 1-minute time step, assuming that all PV panels are oriented in the same direction, and the nominal capacity of the PV generating unit is 5kVA, which is not evenly distributed in the three phases. All households connected to the U phase have a PV installation, while only 50% of the other households have a PV installation, and the voltage is limited to ±10% of the nominal voltage. The time step used in the simulation is 1 minute;

[0159] According to the partial solution code:

[0160] import cvxpy as cp

[0161] #Define problem data and parameters

[0162] n=10#Number of nodes

[0163] P_PV=cp.Parameter(n,nonneg=True)#Photovoltaic power generation

[0164] S_Nom = cp.Parameter (n, positive = True) # nominal apparent power

[0165] V_min, V_max = 0.95, 1.05 # voltage limit

[0166] #Define decision variables

[0167] P_curt = cp.Variable (n, nonneg = True) # Reduced active power

[0168] Q_PV = cp.Variable(n) #Generated reactive power

[0169] #Define the objective function

[0170] objective=cp.Minimize(cp.sum_squares(Q_PV)+cp.sum(P_curt))

[0171] #Add constraints

[0172] constraints=[

[0173] cp.square(P_PV-P_curt)+cp.square(Q_PV)<=

[0174] cp.square(S_Nom),

[0175] V_min<=V_nom+cp.matmul(V_sens,P_PV-P_curt+Q_PV)<=V_max,

[0176] #More constraints... ]

[0178] #Define the optimization problem

[0179] prob=cp.Problem(objective,constraints)

[0180] #Specify the solver and solve the problem

[0181] result=prob.solve(solver=cp.MOSEK)

[0182] print("Optimal cost:",result)

[0183] print("Optimal curtailment:",P_curt.value)

[0184] print("Optimal reactive power:",Q_PV.value);

[0185] The results are as follows Figure 2 and 3 As shown, the comparison of active power reduction under no control and the patent control strategy, as well as the reduction of active power of each household under reactive voltage control and unfair conditions,

[0186] It can be found that one advantage of the present methods is that they can require reactive power injection or absorption in one phase to avoid excessive voltage in another phase; therefore, the reactive power is more evenly distributed in the three phases, resulting in reduced losses. At the same time, the active power reduction for users is fair, avoiding the occurrence of three-phase imbalance problems in the substation area.

[0187] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A method for adaptive reactive power control of a weak communication distributed photovoltaic inverter, characterized in that ,The photovoltaic inverter reactive adaptive control method comprises the following steps: S1. Define the problem and parameters: Determine the power limit, power factor limit and voltage limit of the PV inverter; S2, discretized output level: discretize the photovoltaic output level into several nodes to form a first-order spline; S3. Construct the optimization problem: define the objective function to minimize the use of reactive power and the reduction of active power; add constraints, including power limit, power factor limit and voltage limit; S4. Solve the optimization problem: use a convex optimization solver to solve the problem and obtain the nodes of the piecewise linear control function; S5, applying the control function: applying the optimized control function to the photovoltaic inverter; S6. Update parameters regularly: Set the update time and update the control parameters regularly to optimize the PV inverter in time.

2. According to claim 1, a method for adaptive reactive power control of a weak communication distributed photovoltaic inverter is characterized in that: The objective function in constructing the optimization problem in S3 is as follows: in, represents the PV power cut at node p by the inverter at node i; represents the reactive power at node p generated by the inverter at node i; w: represents the weight factor of the reactive power term in the objective function; P: represents the node index of the control function; n nodes : Represents the total number of nodes in the distribution network.

3. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 1, characterized in that: The power limit in the constraint condition in S3, constructing the optimization problem is as follows: in, represents the photovoltaic power generated by the inverter at node p at node i; represents the nominal apparent power rating of the inverter at node i; Active power reduction constraints:

4. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 1, characterized in that: The power factor restriction in the constraint conditions of constructing the optimization problem in S3 is as follows: in, α: Constant defining the permissible ratio between reactive power and active power.

5. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 1, characterized in that: The voltage limit in the constraint condition in S3, constructing the optimization problem is as follows: in, |V p.i |: represents the voltage at node i; V min : Indicates the minimum allowable voltage; V max : Indicates the maximum allowable voltage.

6. A method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 5, characterized in that: The voltage estimation formula is as follows: in, represents the voltage at node i due to the load in the previous time step; represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the sensitivity of the active power injected into node m to the voltage amplitude of node i; represents the photovoltaic power generated by the inverter at node p at node m; represents the PV power cut at node p by the inverter at node m; represents the reactive power at node p generated by the inverter at node m.

7. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 3 is characterized in that: The active power reduction constraint also includes a fair active power reduction constraint: in, represents the nominal apparent power rating of the inverter at node j; represents the PV power curtailed at node p by the inverter at node j.

8. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 1, characterized in that: It also includes the constraint of maximum apparent power, which is given by the following formula: in, represents the active power of household load connected to node i; represents the reactive power of household loads connected to node i; S max : Indicates the maximum apparent power passing through the transformer.

9. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 1, characterized in that: The goal of constructing the optimization problem in S3 is to minimize the amount of active power reduction by adjusting local control parameters while using reactive power to support voltage control when necessary.

10. The method for adaptive reactive power control of a weak communication distributed photovoltaic inverter according to claim 2, characterized in that: In the objective function, i represents a node index; the number of nodes of the function is limited and is set to 10 to ensure that the computational complexity is controllable.

Images