Distributed photovoltaic voltage regulation method based on reactive voltage sensitivity

Abstract

The present invention discloses a distributed photovoltaic voltage regulation method based on reactive voltage sensitivity, comprising the following steps: step 1, local linearization processing is performed on the steady-state operation working point of the distribution network, and the active regulation node and reactive regulation node with the most significant impact on the voltage level of the distribution network are obtained; step 2, the distributed photovoltaic output in the distribution network is maximized, the objective function is established, and the constraint conditions are determined; step 3, the distributed photovoltaic automatic power generation control optimization model is solved by an improved immune bimodal particle swarm optimization algorithm; step 4, by obtaining the optimization control that meets the constraint conditions, if the optimization control result that meets the constraint conditions is not obtained, the node with the second largest sensitivity is added as the regulation node, and the step 1 is returned to optimize again. The present invention can reduce the active power loss of the distribution network operation, reduce the phenomenon of abandoned light, realize the friendly access of distributed photovoltaics, and improve the safety and economy of the distribution network operation.

Description

Technical Field

[0001] The invention belongs to the field of power system automation and relates to a distributed photovoltaic voltage regulation method based on reactive voltage sensitivity. Background Art

[0002] As the energy crisis becomes increasingly severe, renewable energy is being vigorously developed and has become a trend, and photovoltaic power generation is one of the important forms. At present, the distributed photovoltaic power generation system connected to the distribution network generally operates in the maximum power tracking mode as an uncontrollable power source. As the penetration rate of distributed photovoltaics in the distribution network continues to increase, due to the inherent randomness and intermittency of photovoltaic power generation, the distribution network flow and power supply voltage fluctuate frequently, the power quality deteriorates, the power supply reliability decreases, and the phenomenon of abandoned light becomes increasingly serious, resulting in huge waste of resources, bringing many challenges to the safe, economical and stable operation of the distribution network. At present, in order to solve the voltage fluctuation problem caused by the access of distributed photovoltaics to the distribution network, the main methods used are to adjust the transformer tap of the substation, switch on and off the shunt capacitor, and install a continuous reactive power compensation device SVC in the distribution network for adjustment. However, frequent adjustment of transformer taps and shunt capacitors will greatly reduce the service life of the equipment, and it is difficult to achieve smooth adjustment of the distribution network voltage, which is easy to cause impact on the distribution network. The installation cost of reactive power compensation equipment such as SVC is high and is not suitable for large-scale application. And when the traditional reactive power regulation means in the distribution network cannot effectively suppress voltage fluctuations, it is easy to cause large-scale abandoned light phenomenon, which is easy to cause waste of resources. Summary of the invention

[0003] In order to solve the above technical problems, the present invention provides a distributed photovoltaic voltage regulation method based on reactive voltage sensitivity, which can suppress the voltage instability and power flow fluctuation problems caused by the high penetration rate of photovoltaic power generation. At the same time, it can reduce the active power loss of the distribution network operation, reduce the phenomenon of abandoned light, realize distributed photovoltaic friendly access, and improve the safety and economy of the distribution network operation.

[0004] A distributed photovoltaic voltage regulation method based on reactive voltage sensitivity of the present invention comprises the following steps:

[0005] Step 1: Perform local linearization on the steady-state operation point of the distribution network, obtain the active regulation nodes and reactive regulation nodes that have the most significant impact on the voltage level of the distribution network, and calculate the sensitivity of the voltage to the node injection power;

[0006] Step 2: Taking the maximum output of distributed photovoltaic power in the distribution network, minimizing the active network loss in the steady-state operation of the distribution network and minimizing the node voltage deviation from the rated value as the goal, establish the objective function and determine the constraint conditions;

[0007] Step 3: Solve the distributed photovoltaic automatic power generation control optimization model through the improved immune bimodal particle swarm optimization algorithm, and optimize the active and reactive output of the photovoltaic power generation system in the distribution network;

[0008] Step 4: By obtaining the optimal control that satisfies the constraints, the adjusted output of the distributed photovoltaic power instruction, the parallel capacitor switching number instruction and the SVC reactive output instruction is obtained. If the optimal control result that satisfies the constraints cannot be obtained, the node with the second largest sensitivity is added as the adjustment node, and return to step 1 for optimization again.

[0009] In step 1, the steady-state operation point of the distribution network is locally linearized to obtain the active and reactive regulation nodes that have the most significant impact on the voltage level of the distribution network. For a power network with N nodes, let n = N-1, take the balance node as the reference node, and expand the photovoltaic node into the power flow equation to obtain 2n polar coordinate form Newton method power flow correction equations:

[0010]

[0011]

[0012] Where: is the Jacobian matrix; Δθ and ΔV are the node voltage phase angle and the assignment correction vector respectively; V is the n-dimensional node voltage amplitude diagonal matrix; B and G are the real and imaginary parts of the node admittance matrix respectively. The calculation method of the voltage sensitivity to the node injection power is as follows:

[0013] ΔV=[(B+Q)(GP) -1 (BQ)+(G+P)] -1 ΔP-[(GP)(B+Q) -1 (G+P)+(BQ)] -1 ΔQ.

[0014] In step 2, the expression of the objective function is:

[0015]

[0016] Where: DG n is the photovoltaic access node in the distribution network; P i,DG,pre is the predicted value of the photovoltaic active output connected to node i; P i,DG is the distributed photovoltaic active output value at node i; Q i,DG is the reactive power output value of distributed photovoltaic on node i; v j is the voltage amplitude at node j, is the rated value of the voltage amplitude at node j; α is the active weight coefficient; β is the reactive weight coefficient.

[0017] In step 2, the constraints include power balance constraints, node voltage constraints, gateway switching power constraints, group capacitor constraints, SVC capacity constraints, and distributed photovoltaic module operation constraints;

[0018] The power balance constraint condition is: when the distribution network is operating normally, the active and reactive injection power of each node and the node voltage satisfy the power balance equation:

[0019]

[0020] Where: P i With Q i are the active and reactive injection of node i respectively; is the voltage vector of node i; is the corresponding element of the system admittance matrix;

[0021] The node voltage constraint: To ensure the safe operation of the distribution network, the voltage amplitude of all nodes is kept within the set range:

[0022]

[0023] Where: v i is the voltage amplitude of node i; and are the maximum and minimum values ​​of the node voltage amplitude respectively;

[0024] The gateway exchange power constraint: In order to suppress the impact of power fluctuations in the photovoltaic park on the upper power grid, the exchange power of the root node of the park distribution network must meet the following requirements:

[0025]

[0026] Where: P1 and Q1 are the active and reactive power flowing into the distribution network from the root node respectively; P1 max With P1 min are the upper and lower limits of the active power exchanged at the gateway set by the dispatching center. and They are the upper and lower limits of the reactive power exchanged at the gateway set by the dispatch center;

[0027] The grouped capacitor constraint: The grouped capacitor is an important reactive power compensation device in the distribution network. Its switching state is a discrete decision variable. The discrete variable is continuous in the form of binary coding supplementary constraints. For node i, the discrete gear capacitor group switching is converted into:

[0028]

[0029]

[0030]

[0031]

[0032] Where: Q i,com is the operating capacity of the capacitor bank at node i; is the compensation amount of each gear of the capacitor bank at node i; is the current compensation capacity; is a number between 0 and 1, through Control it to 0, 1 variables, and then change t i Control is an integer. For example, if the capacitor position is extended, the number of binary code bits needs to be increased.

[0033] The SVC capacity constraint: Due to the capacity limitation of the SVC, its reactive output needs to be kept within a preset range:

[0034]

[0035] In the formula, and are the minimum and maximum compensation amounts of the operational capacity of the capacitor bank at node i respectively;

[0036] The distributed photovoltaic module operation constraints:

[0037]

[0038] In the formula, S i,DG is the photovoltaic module capacity of node i, P i,DG is the distributed photovoltaic active output value at node i, Q i,DG is the reactive power output value of distributed photovoltaic on node i, is the predicted value of the photovoltaic active power output connected to node i.

[0039] In step 3, the distributed photovoltaic automatic power generation control optimization model is solved by using an improved immune bimodal particle swarm optimization algorithm. The improved immune bimodal particle swarm optimization algorithm is improved by adding a local reinforcement learning operator to assist some optimal particles to jump out of the local extreme point. The reinforcement learning operator calculation formula is:

[0040]

[0041] Where: G kd is the enhanced particle; G best is the global optimal individual of the current generation; m is the limit constant; t is the number of generations; T is the total number of generations executed by the algorithm; X r1d With X r2d The positions are randomly selected and are not the same.

[0042] In step 4, the method for obtaining the output adjustment of the distributed photovoltaic power command, the parallel capacitor switching number command and the SVC reactive output command is:

[0043] According to the improved immune dual-modal particle swarm optimization algorithm, the distribution network topology, line parameters and module real-time input are pre-input into the algorithm. The module real-time input is the voltage and power vector of each node in the distribution network. If the current node voltage does not exceed the limit, wait for the next cycle. If there is an over-limit, solve the node voltage to the photovoltaic node injection active and reactive power sensitivity matrix, the node with the largest voltage over-limit node sensitivity is the adjustment node, and the remaining nodes work in the maximum power point tracking state. The active and reactive power of the adjustment node are used as the adjustment variables for comprehensive optimization. The improved particle swarm optimization algorithm is used for comprehensive optimization. If there is an optimal result, the power instruction of the distributed photovoltaic, the parallel capacitor switching number instruction and the adjusted output of the SVC reactive output instruction are obtained. If there is no optimal result, the node with the second largest sensitivity is added as the adjustment node and returned for optimization again.

[0044] In step 4, after obtaining the adjusted output of the distributed photovoltaic power command, the parallel capacitor switching number command and the SVC reactive output command, if the voltage exceeds the limit, the voltage deviation coefficient feedback of the objective function is increased, and the voltage deviation coefficient value is gradually increased with the voltage just not exceeding the limit as the boundary to correct the power flow calculation deviation caused by uncertain factors.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] By deriving the sensitivity matrix of distribution network voltage to node injection power, a distributed photovoltaic active-reactive coordinated optimization method based on sensitivity analysis is proposed. The active and reactive regulation nodes with the most significant impact on the distribution network voltage level are obtained by sensitivity analysis, and a distributed photovoltaic automatic power generation control optimization model is established. The model is solved by an improved immune bimodal particle swarm optimization algorithm, and the active and reactive output of the photovoltaic power generation system in the distribution network is optimized and controlled online.

[0047] Each adjustment process of the method of the present invention avoids the adjustment of all photovoltaic modules. At the same time, the solution variables within an optimization cycle will not increase with the increase of photovoltaic access points in the distribution network. When the distributed photovoltaic penetration rate in the distribution network gradually increases, the model will not be difficult to solve. It is suitable for large-scale distributed photovoltaic access and has great economic significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to make the contents of the present invention more clearly understood, the present invention is further described in detail below based on specific embodiments in conjunction with the accompanying drawings.

[0049] Figure 1 It is a flow chart of the control steps;

[0050] Figure 2 To improve the particle swarm algorithm flow chart;

[0051] Figure 3 Flowchart of the comprehensive optimization method. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0053] like Figure 1 As shown, the distributed photovoltaic voltage regulation method based on reactive voltage sensitivity described in the present invention, the steps of the control strategy are:

[0054] Step 1: Perform local linearization processing near the steady-state operation point of the distribution network, obtain the active regulation nodes and reactive regulation nodes that have the most significant impact on the voltage level of the distribution network, and calculate the sensitivity of the voltage to the node injection power.

[0055] Step 2: With the goal of maximizing the distributed photovoltaic output in the distribution network, minimizing the active network loss in the steady-state operation of the distribution network, and minimizing the node voltage deviation from the rated value, establish the objective function and determine the relevant constraints.

[0056] Step 3: Solve the distributed photovoltaic automatic power generation control optimization model through the improved immune bimodal particle swarm optimization algorithm, and optimize the active and reactive output of the photovoltaic power generation system in the distribution network.

[0057] Step 4: If the optimal control result that satisfies the constraints is obtained, the distributed photovoltaic power instruction, the number of parallel capacitor switching and the SVC reactive output instruction are adjusted. If there is no optimal result, the node with the second largest sensitivity is added as the adjustment node and returned for optimization again.

[0058] In step 1, the active and reactive regulation nodes that have the most significant impact on the voltage level of the distribution network are obtained by local linearization near the steady-state operation point of the distribution network. For a power network with N nodes, let n = N-1, take the balance node as the reference node, and expand the photovoltaic node into the power flow equation, and 2n polar coordinate form Newton method power flow correction equations can be obtained.

[0059]

[0060]

[0061] Where: is the Jacobian matrix; Δθ and ΔV are the node voltage phase angle and the assignment correction vector respectively; V is the n-dimensional node voltage amplitude diagonal matrix; B and G are the real and imaginary parts of the node admittance matrix respectively. Bcosθ is a simplified way of writing the matrix, which has the same structure as B, and its elements are the corresponding elements in B. ij and cosθ ij The product of P and Q is n-order diagonal matrix, and its diagonal elements are P i / V i 2 and Q i / V i 2 ; ΔP and ΔQ are the active and reactive power injection disturbances of the node.

[0062] The calculated sensitivity of voltage to node injection power is:

[0063] ΔV=[(B+Q)(GP) -1 (BQ)+(G+P)] -1 ΔP-[(GP)(B+Q) -1 (G+P)+(BQ)] -1 ΔQ

[0064] Furthermore, in step 2, the objective function is established with the goal of maximizing the distributed photovoltaic output in the distribution network, minimizing the active network loss in the steady-state operation of the distribution network, and minimizing the node voltage deviation from the rated value, and the relevant constraints are determined as follows:

[0065] The distributed photovoltaic output in the distribution network is maximized, the active network loss in the steady-state operation of the distribution network is minimized, and the node voltage deviation from the rated value is minimized. The active-reactive output value of the active-reactive regulation node is used as the solution variable to establish a distributed photovoltaic active-reactive coordination optimization model. Its objective function is:

[0066]

[0067] Where: DG i is the photovoltaic access node in the distribution network; P i,DG,pre The predicted value of the photovoltaic active output connected to node i; P i,DG is the distributed photovoltaic active output value at node i; Q i,DG is the reactive power output value of distributed photovoltaic on node i; v j is the voltage amplitude at node j, is the rated value of the voltage amplitude at node j (all per unit value); α is the active power weight coefficient; β is the reactive power weight coefficient.

[0068] In order to make the optimization results meet the system operation requirements, the following constraints are added to the optimization algorithm, including power balance constraints, node voltage constraints, gateway switching power constraints, grouped capacitor constraints, SVC capacity constraints and distributed photovoltaic module operation constraints:

[0069] Power balance constraint: When the distribution network is operating normally, the active and reactive injected power of each node and the node voltage satisfy the power balance equation.

[0070]

[0071] Where: P i With Q i are the active and reactive injection of node i respectively; is the voltage vector of node i; Y ij is the corresponding element of the system admittance matrix.

[0072] Node voltage constraint: To ensure the safe operation of the distribution network, the voltage amplitude of all nodes must be kept within a certain range.

[0073]

[0074] Where: v i is the voltage amplitude of node i; and are the maximum and minimum values ​​of the node voltage amplitude respectively.

[0075] Gateway exchange power constraint: In order to suppress the impact of power fluctuations in the photovoltaic park on the upper power grid, it is necessary to consider the exchange power of the root node of the park distribution network.

[0076]

[0077] Where: P1 and Q1 are the active and reactive power flowing into the distribution network from the root node respectively; P1 min With P1 max are the lower and upper bounds of the gateway switching power set by the dispatch center, and They are the upper and lower limits of the reactive power exchanged at the gateway set by the dispatching center.

[0078] Grouped capacitor constraints: Grouped capacitors are important reactive power compensation equipment in distribution networks. Their switching states are essentially discrete decision variables. We propose to use binary coding supplementary constraints to make discrete variables continuous. For node i, the discrete-gear capacitor group switching is converted into

[0079]

[0080]

[0081]

[0082]

[0083] Where: Q i,com is the operating capacity of the capacitor bank at node i; is the compensation amount of each gear of the capacitor bank at node i; is the current compensation capacity; is a number between 0 and 1, Control it to 0, 1 variables, and then change t i Control is an integer. If the capacitor range is extended, just increase the number of binary code bits.

[0084] SVC capacity constraint: Due to the capacity limitation of SVC, its reactive output needs to be kept within a certain range.

[0085]

[0086] In the formula, and They are the minimum and maximum compensation amounts of the operating capacity of the capacitor bank at node i respectively.

[0087] Distributed photovoltaic module operation constraints:

[0088]

[0089] In the formula, S i,DG is the photovoltaic module capacity of node i.

[0090] In step 3, the distributed photovoltaic automatic generation control optimization model is solved by the improved immune bimodal particle swarm optimization algorithm to optimize the active and reactive output of the photovoltaic power generation system in the distribution network. Figure 2 As shown in the figure, based on the characteristics of rapidity and accuracy in solving optimization problems, the immune bimodal particle swarm optimization algorithm is used to solve the distributed photovoltaic automatic power generation control optimization model through the improved immune bimodal particle swarm optimization algorithm. The improved immune bimodal particle swarm optimization algorithm is improved by adding a local reinforcement learning operator to assist some optimal particles to jump out of local extreme points. The calculation formula of the reinforcement operator is:

[0091]

[0092] Where: G kd is the enhanced particle; G best is the global optimal individual of the current generation; m is the limit constant; t is the number of generations; T is the total number of generations executed by the algorithm; X r1d With X r2d The positions are randomly selected and are not the same.

[0093] In step 4, the comprehensive optimization model of active and reactive power is as follows: Figure 3 As shown, according to step 3, the improved immune dual-mode particle swarm optimization algorithm is used to pre-input the distribution network topology, line parameters and module real-time input into the algorithm. The module real-time input is the voltage and power vector of each node in the distribution network. If the current node voltage does not exceed the limit, wait for the next cycle. If it exceeds the limit, solve the node voltage to inject active and reactive power sensitivity matrix to the photovoltaic node. The node with the greatest sensitivity to the voltage exceeding the limit is the regulating node, and the remaining nodes work in the maximum power point tracking state. Next, the active and reactive power of the regulating node are used as the regulating variables for comprehensive optimization, and the improved particle swarm optimization algorithm is used for comprehensive optimization. If there is an optimal result, the power command of the distributed photovoltaic, the number of parallel capacitor switching commands and the SVC reactive output command are obtained. If there is no optimal result, the node with the second largest sensitivity is added as the regulating node and returned for optimization again.

[0094] After the photovoltaic node of the distribution network receives the regulation instruction to change the output power, if the voltage still exceeds the limit, the voltage deviation coefficient feedback of the objective function is increased to correct the power flow calculation deviation caused by uncertain factors.

[0095] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A distributed photovoltaic voltage regulation method based on reactive voltage sensitivity, characterized in that: The steps include: Step 1: Perform local linearization on the steady-state operation point of the distribution network, obtain the active regulation nodes and reactive regulation nodes that have the most significant impact on the voltage level of the distribution network, and calculate the sensitivity of the voltage to the node injection power; Step 2: Taking the maximum output of distributed photovoltaic power in the distribution network, minimizing the active network loss in the steady-state operation of the distribution network and minimizing the node voltage deviation from the rated value as the goal, establish the objective function and determine the constraint conditions; The expression of the objective function is: Where: DG n P is the photovoltaic access node in the distribution network, and the calculation is performed from node 1 to node n; i,DG,pre is the predicted value of the photovoltaic active output connected to node i; P i,DG is the distributed photovoltaic active output value at node i; Q i,DG is the reactive power output value of distributed photovoltaic at node i; v j is the voltage amplitude at node j, is the voltage amplitude rating of node j; α is the active weight coefficient; β is the reactive weight coefficient; The constraints include power balance constraints, node voltage constraints, gateway switching power constraints, group capacitor constraints, SVC capacity constraints and distributed photovoltaic module operation constraints; The power balance constraint condition is: when the distribution network is operating normally, the active and reactive injection power of each node and the node voltage satisfy the power balance equation: Where: P i With Q i are the active and reactive injection of node i respectively; is the voltage vector of node i; is the corresponding element of the system admittance matrix, is the magnitude of the voltage vector at node i; The node voltage constraint: keep the voltage amplitude of all nodes within the set range: Where: v i is the voltage amplitude of node i; and are the minimum and maximum values ​​of the node voltage amplitude respectively; The gateway exchange power constraint: In order to suppress the impact of power fluctuations in the photovoltaic park on the upper power grid, the exchange power of the root node of the park distribution network must meet the following requirements: Where: P1 and Q1 are the active and reactive power flowing into the distribution network from the root node respectively; P1 max With P1 min are the upper and lower limits of the active power exchanged at the gateway set by the dispatching center. and They are the upper and lower limits of the reactive power exchanged at the gateway set by the dispatch center; The grouped capacitor constraint: The switching state of the grouped capacitor is a discrete decision variable. The discrete variable is continuous in the form of binary coded supplementary constraints. For node i, the discrete gear capacitor group switching is converted into: Where: Q i,com is the operational capacity of the capacitor bank at node i; is the compensation amount of each gear of the capacitor bank at node i; is the current compensation capacity; is a number between 0 and 1, through Control it to 0, 1 variables, and set t i Controlled as an integer, when the capacitor position is extended, the number of bits of binary code is increased; The SVC capacity constraint: Due to the capacity limitation of the SVC, its reactive output needs to be kept within a preset range: In the formula, and are the minimum and maximum compensation amounts of the operational capacity of the capacitor bank at node i respectively; The distributed photovoltaic module operation constraints: In the formula, S i,DG is the photovoltaic module capacity of node i, P i,DG is the distributed photovoltaic active output value at node i, Q i,DG is the reactive power output value of distributed photovoltaic on node i, is the predicted value of the photovoltaic active output connected to node i; Step 3: Solve the distributed photovoltaic automatic power generation control optimization model through the improved immune bimodal particle swarm optimization algorithm, and optimize the active and reactive output of the photovoltaic power generation system in the distribution network; the improved immune bimodal particle swarm optimization algorithm is improved by adding a local reinforcement learning operator to assist some optimal particles to jump out of the local extreme point, and the reinforcement learning operator calculation formula is: Where: G kd is the enhanced particle; G best is the global optimal individual of the current generation; m is the limit constant; t is the number of generations; T is the total number of generations executed by the algorithm; X r1d With X r2d The positions are randomly selected and are not the same; Step 4: by obtaining the optimal control satisfying the constraint conditions, the distributed photovoltaic power instruction, the parallel capacitor switching number instruction and the adjusted output of the SVC reactive output instruction are obtained. If the optimal control result satisfying the constraint conditions cannot be obtained, the node with the second largest sensitivity is added as the regulating node, and the process returns to step 1 for optimization again; The method for obtaining the power instruction of distributed photovoltaic, the number of parallel capacitor switching instructions and the adjusted output of SVC reactive output instruction is as follows: According to the improved immune bimodal particle swarm optimization algorithm, the distribution network topology, line parameters and module real-time input are pre-input into the algorithm. The module real-time input is the voltage and power vector of each node in the distribution network. If the current voltage of each node does not exceed the limit, wait for the next cycle. If there is an over-limit, solve the node voltage to the photovoltaic node injection active and reactive power sensitivity matrix. The node with the largest voltage over-limit node sensitivity is the adjustment node, and the remaining nodes work in the maximum power point tracking state. The active and reactive power of the adjustment node are used as the adjustment variables for comprehensive optimization. The improved immune bimodal particle swarm optimization algorithm is used for comprehensive optimization. If there is an optimal result, the power instruction of the distributed photovoltaic, the parallel capacitor switching number instruction and the SVC reactive output instruction are obtained. If there is no optimal result, the node with the second largest sensitivity is added as the adjustment node and return to step 1 for optimization again; After obtaining the adjusted output of the distributed photovoltaic power command, the parallel capacitor switching number command and the SVC reactive output command, if the voltage exceeds the limit, the voltage deviation coefficient of the objective function is increased and gradually increased until the voltage does not exceed the limit, thereby correcting the power flow calculation deviation caused by uncertain factors.

Images