Wind power-containing power grid dispatching strategy optimization method and device

Abstract

The invention provides a wind power-containing power grid dispatching strategy optimization method and device, and belongs to the technical field of power dispatching, and the method comprises the steps: constructing a multi-time-scale collaborative optimization model, the model optimization target is that the sum of the operation cost and the control cost is the lowest, and the control cost comprises the reconstruction cost, the SVG adjustment cost, the fan wind curtailment cost and the load interruption cost; reconstructing and optimizing to obtain a power distribution network switch strategy; according to the power distribution network switching strategy, adjusting the reactive power output of a static var generator, the wind curtailment power of a fan and the active power of an interruptible load to obtain a power grid dispatching strategy; performing power grid dispatching by using the power grid dispatching strategy; checking the thermal stability of the power distribution network line by using the thermal stability margin increment index; and outputting the power grid dispatching strategy passing the thermal stability check. The method can solve the problem of unnecessary wind curtailment caused by judging the thermal stability of the power grid system through the static current-carrying capacity.

Description

Technical Field

[0001] The invention relates to a method and a device for optimizing a dispatching strategy of a power grid containing wind power, and belongs to the technical field of relay protection. Background Art

[0002] Distributed wind power is widely used due to its advantages such as clean energy and local consumption, but its intermittent and volatile nature poses a huge challenge to the safe operation of the distribution network. Under the premise of ensuring the safety and stability of the distribution network, in order to improve the flexibility of the distribution network and reduce the operating cost of the distribution network, the distribution network is usually optimized and dispatched by means of network reconstruction, distributed power output regulation, system reactive power optimization, and coordinated optimization of multiple methods. However, in the optimization process, a fixed maximum allowable current carrying capacity (hereinafter referred to as current carrying capacity), namely static current carrying capacity (STR), is often used to judge the thermal stability of the system.

[0003] For specific line models, STR sets the current carrying capacity based on strict environmental parameters and the maximum allowable temperature of the line. Its thermal stability analysis results tend to be conservative, especially for wind power distribution networks. The conservative estimate of the thermal stability of wind power transmission channels limits the transmission capacity, which is easy to cause unnecessary wind abandonment and is not conducive to wind power consumption. Therefore, the problem of unnecessary wind abandonment caused by STR in the operation of distribution networks needs to be solved urgently. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a method and device for optimizing a grid dispatching strategy containing wind power, so as to solve the problem of unnecessary wind abandonment caused by judging the thermal stability of the grid system by static current carrying capacity.

[0005] In order to solve the above technical problems, the present invention is implemented by adopting the following technical solutions:

[0006] In a first aspect, the present invention provides a method for optimizing a wind power grid dispatching strategy, comprising:

[0007] S1: construct a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of operating cost and control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost;

[0008] S2: Collect the grid topology and operating condition information including wind power, reconstruct and optimize the grid topology including wind power, and obtain the switching strategy of the distribution network after reconstruction and optimization;

[0009] S3: In the distribution network, according to the distribution network switch strategy obtained in step S2, the reactive output of the static VAR generator, the wind turbine abandoned wind power and the interruptible load active power are adjusted to obtain an optimized grid dispatching strategy including wind power;

[0010] S4: Performing grid dispatching using the grid dispatching strategy including wind power obtained in step S3;

[0011] S5: Use the thermal stability margin increment index to check the thermal stability of the distribution network line;

[0012] S6: If the thermal stability check of the distribution network line is qualified, output the grid dispatching strategy including wind power after the thermal stability check in step S5.

[0013] In the aforementioned method for optimizing the dispatching strategy of a power grid including wind power, the multiple time scales include the total period, the reconstruction period, the dispatching period and the verification period using multiple time scales;

[0014] A total period is divided into several reconstruction periods;

[0015] A reconstruction period is divided into several scheduling periods;

[0016] A scheduling period is divided into several verification periods.

[0017] In the above-mentioned method for optimizing the dispatching strategy of a power grid including wind power, in step S1, in a multi-time scale collaborative optimization model, the objective function is established according to the optimization goal of minimizing the operation and control costs as follows:

[0018] min(C CT +C OP );

[0019] In the formula, C CT Indicates control cost; C OP Indicates the operating cost;

[0020] Control costs CT , the switch action vector x in each period t , static VAR generator SVG reactive power output vector Q SVG,t 、Wind turbine abandoned wind AWP power vector P AWT,t and the interruptible load IL output vector P IL,t The cost and composition of each period include: the total period T r , reconstruction period T d and scheduling period T v , C CT The calculation formula is:

[0021]

[0022] In the formula, the L×1-dimensional switch action vector x t The cost and the reconstruction cost are SW,t =γ SW ·d(x t -xt-1 ), γ SW is the cost of the switching action per unit number, and the function d is used to take the vector (x t -x t-1 ) the total number of non-zero elements; N SVG ×1-dimensional static VAR generator SVG reactive output vector Q SVG,t The cost and SVG adjustment cost are SVG adjustment cost C SVG,t =γ SVG Q SVG,t , γ SVG 1×N WT Dimensional SVG unit capacity adjustment cost vector; N WT ×1-dimensional wind turbine abandoned wind AWP power vector P AWT,t The cost and the penalty for wind abandonment, wind abandonment penalty C AWT,t =γ AWT ·P AWT,t , γ AWT 1×N WT Dimensional unit wind curtailment penalty cost vector; N IL ×1-dimensional interruptible load IL active output vector P IL,t The cost and load interruption cost are load interruption cost C IL,t =γ IL ·P IL,t , γ IL 1×N IL Dimensional unit load interruption cost vector;

[0023] Operating cost OP The cost of purchasing electricity from the wind power station area to the main grid in each period is C G,t 、Wind power operation cost C WT,t and distribution network loss cost C LS,t Composition, the calculation formula is:

[0024]

[0025] In the formula, the cost of purchasing electricity from the main grid in the wind power station area is C G,t =γ G ·p G,t , γ G is the unit capacity electricity purchase cost, p G,t The access point exchanges active power with the main grid; the wind power operation cost C WT,t =γ WT ·P WT,t , γ WT 1×N WT The power generation cost vector of the wind turbine per unit capacity, P WT,t N WT ×1-dimensional wind turbine active output vector; distribution network loss cost C LS,t=γ LS ·P LS,t , γ LS is the unit line loss cost, P LS,t is the bus loss of the power distribution network.

[0026] The above-mentioned method for optimizing the dispatching strategy of a power grid including wind power comprises, in step S2:

[0027] S21: Obtain the grid topology and operating condition information including wind power, close all the connection and section switches of the distribution network so that the distribution lines form several ring networks;

[0028] S22: determining clustering scenarios of each reconstruction period according to the source-load curve, and in each generated scenario, using the improved golden jackal algorithm to obtain the distribution network switching strategy of each ring network after reconstruction optimization in the reconstruction period;

[0029] The source-load curve includes a source curve and a load curve, the source curve is a wind turbine output curve, and the load curve is a load curve;

[0030] Each reconstruction period corresponds to a type of scenario. The improved golden jackal algorithm includes a population mutation mechanism and a neighbor search mechanism. The reconstruction distribution network switch strategy is to reconstruct and change the opening and closing mode of the distribution network connection and section switches.

[0031] In the above-mentioned method for optimizing the dispatching strategy of a power grid including wind power, in step S22, the process of solving the population variation mechanism includes:

[0032] S221: The traditional golden jackal algorithm generates an initial population, and the tent mapping is used to improve the uniformity of the initial population to generate an evolved population;

[0033] The initial population is composed of multiple individuals, each of which is represented by a switch action vector. The process of generating an evolutionary population by Tent mapping includes:

[0034] Perform tent mapping on each individual in the initial population to generate individuals in the evolutionary population. The individuals x in the evolutionary population 0,n Satisfy the following formula:

[0035]

[0036] In the formula, x min is the switch action vector of the L×1-dimensional ring network number lower limit, and its elements correspond to the number lower limits of the switches in each ring network. max is the switch action vector of the upper limit of the number of the L×1-dimensional ring network, and each element thereof corresponds to the upper limit of the number of switches in each ring network. is the Hadamard product, which is used to multiply the elements of each group of corresponding positions in two matrices of the same order to obtain a new matrix, ω nis an L×1 dimensional random vector, any element ω n,i ∈[0,1], Tent() is the Tent mapping function, and the calculation formula is as follows:

[0037]

[0038] S222: Generate a mutant population by increasing the randomness of the evolving population through adaptive t-distribution mutation of the current evolving population;

[0039] The process of generating a mutant population by adaptive t-distribution mutation includes:

[0040] Individual x of the evolutionary population at the current kth iteration k,n After the adaptive t distribution mutation, the corresponding individuals of the mutant population are generated The calculation is as follows:

[0041]

[0042] Where K is the total number of iterations, and the t distribution function Trnd(k) takes k as the parameter degree of freedom;

[0043] S223: Calculate the individual fitness of the evolution population in step S221 and the mutation population in step S222, and compare the fitness of each group of corresponding individuals in the two populations;

[0044] Among them, the switch action vector represents the distribution network switching strategy, and the calculation of individual fitness includes:

[0045] Applying the distribution network switch strategy to the distribution network and calculating the corresponding distribution network bus loss;

[0046] The individual fitness is measured by using the calculated total loss of the power distribution network, and the individual fitness is negatively correlated with the corresponding total loss of the power distribution network;

[0047] S224: retain the individuals with high fitness in each group of corresponding individuals in step S22 to form a preferred population.

[0048] In the above-mentioned method for optimizing the dispatching strategy of a power grid including wind power, in step S22, the solution process of the nearest neighbor search mechanism includes:

[0049] S225: Compare the individual fitness of the optimal population generated in step S224, and select the individual with the highest fitness as the male jackal x m,k,n , select the individual with the second highest fitness as the female jackal x f,k,n ;

[0050] S226: updating the positions of individuals in the population according to the prey escape energy, and forming the updated individuals into a prey population, including the following steps:

[0051] Calculate the prey escape energy. The calculation formula of the prey escape energy E is as follows:

[0052]

[0053] Where E0 is the initial avoidance energy randomly generated by the traditional golden jackal algorithm, and E0∈(0,1);

[0054] Update the positions of individuals in the population according to the calculated prey escape energy:

[0055] Male Jackal m,k,n Used to generate individuals in the optimal population in step S224 Corresponding male jackal prey x mp,k,n ;

[0056] Female jackal fp,k,n Used to generate individuals in the optimal population in step S224 Corresponding female jackal prey x fp,k,n ;

[0057] According to the generated male jackal prey x mp,k,n and female jackal prey x fp,k,n Calculate individual The individual x corresponding to the prey population k+1,n ;

[0058] The calculation process of the position of individuals in the prey escape energy update population is as follows:

[0059] When E ≥ 1, the male jackal's prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing:

[0060]

[0061] When E<1, the male jackal's prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing:

[0062]

[0063] In the formula, the Levy distribution random vector rl and the individual in the optimal population are vectors of the same order, The individuals of the optimal population generated in step S224 after the kth iteration;

[0064] Collect updated individuals x k+1,n Form prey populations;

[0065] S227: Use the nearest neighbor search mechanism to search for the individual x generated in step S226 k+1,n Randomly add or subtract one from any element in to get individual x k+1,n The nearest neighbor solution for an element in ;

[0066] Among them, individual x k+1,n is an L×1-dimensional switch action vector, where L is the number of rings and each element corresponds to the number of the switch in each ring;

[0067] S228: Loop through step S227 to obtain x k+1,n The nearest neighbor scheme for each element in ;

[0068] S229: x obtained in step S226 is k+1,n and the distribution network switch strategies corresponding to the neighboring solutions obtained in step S228 are respectively applied to the distribution network to calculate the corresponding individual fitness;

[0069] S2210: Compare x in step S229 k+1,n And the individual fitness of the nearest neighbor solution obtained, retaining each x k+1,n The distribution network switching strategy with the highest fitness among its neighboring solutions is used as the switching action vector corresponding to this strategy as the new x k+1,n ;

[0070] The solution process of improving the traditional golden jackal algorithm in step S22 also includes:

[0071] S2211: Collect the new x in step S2210 k+1,n Combine into new evolutionary populations;

[0072] S2212: Determine the new x in step S2210 k+1,n Whether the corresponding number of iterations k reaches the upper limit,

[0073] If the number of iterations does not reach the upper limit, the number of iterations k is increased by 1 and the new evolution population generated in step S2211 is used as the current evolution population to execute step S222;

[0074] If the number of iterations reaches the upper limit, the new x in step S2210 is output. k+1,n The distribution network switching strategy with the highest corresponding fitness is used as the distribution network switching strategy solved in step S22.

[0075] In the above-mentioned method for optimizing the dispatching strategy of a power grid including wind power, in step S5, the calculation formula of the thermal stability margin increment is:

[0076] λ=ΔI TR -ΔI DN ;

[0077] In the formula, ΔITR is the dynamic current carrying capacity change, ΔI DN is the maximum current change of the distribution network;

[0078] The calculation formula for dynamic current carrying capacity change is:

[0079]

[0080] Where sgn() is a function that returns the sign of the variable, ΔV w is the wind speed change, indicating the current wind speed V w,t The initial wind speed V w,0 The difference between the dynamic current carrying capacity change rate Among them, the convection heat dissipation coefficient k c are preset values ​​related to wind angle, line diameter, air density, air thermal conductivity, air viscosity, etc. c,0 is the initial temperature of the fan conductor, T a is the ambient temperature, R(T c,0 ) is the initial temperature of the fan conductor at the fan conductor T c,0 Resistance value under, mC p is the conductor heat capacity of the fan, t represents the period from the initial moment to the current moment;

[0081] The calculation formula for the maximum current change of the distribution network is:

[0082] ΔI DN =k DN |ΔV w,cal | 0.5 ;

[0083] In the formula, k DN is the rate of change of distribution network current, ΔV w,cal is the wind speed change; distribution network current change rate P r is the rated power of the fan, V r is the rated speed of the fan, V i is the fan cut-in speed, R + is the equivalent resistance of the fan port, R - is the equivalent resistance of the substation port, which is the power supply access port of the power distribution main grid; wind speed change min() is an operation that returns the minimum value, and max() is an operation that returns the maximum value.

[0084] In the above-mentioned method for optimizing the dispatching strategy of a power grid including wind power, if the thermal stability check of the distribution network line fails in step S5, the method returns to step S3, readjusts the reactive output of the static VAR generator, the wind turbine abandoned power and the active power of the interruptible load, and obtains a new dispatching strategy for a power grid including wind power.

[0085] The aforementioned method for optimizing the dispatching strategy of a power grid containing wind power, which performs rolling optimization and solves the multi-time scale collaborative optimization model to obtain an optimized dispatching strategy of a power grid containing wind power, comprises:

[0086] In the previous total period, the distribution network switching strategy of each reconstruction period is obtained according to step S2, and the distribution network switching strategy of each reconstruction period obtained is correspondingly applied to the current total period;

[0087] At the beginning of each dispatching period in each reconstruction period of the current total period, a grid dispatching strategy containing wind power is obtained according to step S3, and the obtained grid dispatching strategy containing wind power is applied to the current dispatching period;

[0088] In the current total period, the verification period is used as the optimization period of the safety verification module, and the thermal stability margin increment is cyclically calculated according to step S5 and the thermal stability of the distribution network line is verified;

[0089] In the current total period, the distribution network switching strategy of each reconstruction period is obtained according to step S2, and the distribution network switching strategy of each reconstruction period obtained is correspondingly applied to the next total period.

[0090] In a second aspect, the present invention provides a wind power grid dispatch strategy optimization device, comprising a model building module, a switch strategy module, a dispatch strategy module, a strategy execution module, a thermal stability module and a strategy output module;

[0091] The model building module is used to build a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of the operating cost and the control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost;

[0092] The switch strategy module is used to collect the topology and operating condition information of the power grid containing wind power, and reconstruct and optimize the power grid topology containing wind power to obtain the switch strategy of the distribution network after reconstruction and optimization;

[0093] The dispatching strategy module is used to adjust the reactive output of the static VAR generator, the abandoned wind power of the wind turbine and the active power of the interruptible load as a control means in the distribution network according to the distribution network switching strategy of the switching strategy module, so as to obtain an optimized grid dispatching strategy containing wind power;

[0094] The strategy execution module is used to perform grid dispatching using the grid dispatching strategy containing wind power obtained in step S3;

[0095] The thermal stability module is used to check the thermal stability of the distribution network line using the thermal stability margin increment index;

[0096] The strategy output module is used to output the wind power-containing power grid dispatching strategy after the thermal stability check in the thermal stability module if the thermal stability check of the distribution network line is qualified.

[0097] Beneficial effects:

[0098] The method for optimizing the dispatching strategy of a power grid containing wind power of the present invention adopts the thermal stability margin increment as a constraint condition for determining the thermal stability of the line, and can more accurately determine the thermal stability of the distribution system. While ensuring the safe operation of the distribution network, it reduces the situation of unnecessary wind abandonment caused by overly conservative estimation of the line current carrying capacity, thereby improving the economy of the power grid operation.

[0099] The present invention designs a method for quickly calculating the thermal stability margin increment, including: quickly evaluating the change in dynamic current carrying capacity of the line and the change in the maximum current of the distribution network according to the change in wind speed, and quickly calculating the thermal stability margin increment according to the evaluation result.

[0100] The multi-time scale collaborative optimization model designed by the present invention is aimed at the variable characteristics of the control variables: distribution network switches, reactive output of static VAR generators, wind turbine wind power abandonment and interruptible load active power. The distribution network switches have duality and are discrete variables; the reactive output of static VAR generators, wind turbine wind power abandonment and interruptible load active power are continuous variables. The optimization problems of discrete variables and continuous variables need to be optimized and solved using different algorithms. The present invention classifies the control variables, and the distribution network switches are classified as the control variables of the distribution network topology. The reactive output of static VAR generators, wind turbine wind power abandonment and interruptible load active power are classified as the control variables of the dispatching strategy. A method for evaluating the reconstruction optimization effect is designed to solve the distribution network switch strategy, and a method for evaluating the dispatching optimization effect is designed to solve the dispatching strategy. At the same time, a safety verification effect evaluation method based on the thermal stability margin increment is designed to verify the thermal stability of the distribution network lines. The dispatching strategy is implemented based on the distribution network topology, and the verification of the thermal stability of the distribution network lines is implemented based on the dispatching strategy. The time required for the verification of the thermal stability of the lines is shorter than the time required for the optimization solution. Therefore, the model of the present invention adopts multiple time scales to collaboratively perform rolling optimization solution on the model, continuously provide optimized power grid dispatching strategies for the distribution network, and improve the stability and safety of the distribution network system.

[0101] In view of the dependence of the dispatching strategy on the distribution network topology, the present invention adopts an improved golden jackal algorithm. On the basis of the traditional golden jackal algorithm with convergence advantages, population mutation and nearest neighbor search mechanism are introduced to further improve the global convergence performance and improve the solution efficiency.

[0102] The population mutation mechanism first uses Tent mapping to improve the uniformity of the initial population. Then, it increases the randomness of the evolving population through adaptive t-distribution mutation, making the population space characteristics uniform and randomly distributed, laying a good foundation for the subsequent evolutionary process, allowing the algorithm to explore the solution space more widely in the future, thereby avoiding converging to the local optimal solution too early.

[0103] The population update method mainly depends on the position of male and female jackals. In order to avoid falling into the local optimum, the nearest neighbor search mechanism is introduced in combination with the variable characteristics of the distribution network switch to improve the performance and efficiency of the algorithm and enhance the application scope and effect of the algorithm. The traditional golden jackal algorithm is a global search strategy, and the nearest neighbor search is a local search strategy. The combination of global search and local search can achieve the goal of obtaining the global optimal distribution network switch strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0104] Figure 1 It is a framework diagram of the method for optimizing the grid dispatching strategy including wind power in Embodiment 1 of the present invention;

[0105] Figure 2 is a flow chart for solving the multi-time scale collaborative optimization model in Example 1 of the present invention;

[0106] Figure 3 This is a topological diagram of the IEEE-33 node power distribution system used for testing the effect of Example 2 of the present invention. DETAILED DESCRIPTION

[0107] It should be noted that: the technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0108] In this article, the character " / " generally indicates that the previous and next associated objects are in an "or" relationship.

[0109] Example 1

[0110] This embodiment provides a method for optimizing a wind power grid dispatching strategy, comprising the following steps:

[0111] S1: construct a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of operating cost and control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost;

[0112] S2: Collect the grid topology and operating condition information including wind power, reconstruct and optimize the grid topology including wind power, and obtain the switching strategy of the distribution network after reconstruction and optimization;

[0113] S3: In the distribution network, according to the distribution network switch strategy obtained in step S2, the reactive output of the static VAR generator, the wind turbine abandoned wind power and the interruptible load active power are adjusted to obtain an optimized grid dispatching strategy including wind power;

[0114] S4: Performing grid dispatching using the grid dispatching strategy including wind power obtained in step S3;

[0115] S5: Use the thermal stability margin increment index to check the thermal stability of the distribution network line;

[0116] S6: If the thermal stability check of the distribution network line is qualified, output the grid dispatching strategy including wind power after the thermal stability check in step S5.

[0117] The above is a form of the method for optimizing the dispatching strategy of a power grid containing wind power according to this embodiment, which is applied to optimizing the dispatching strategy of a power grid containing wind power.

[0118] In order to better illustrate the optimization method of the grid dispatch strategy containing wind power in this embodiment, the following is combined with Figure 1 The framework diagram of the method for optimizing the grid dispatching strategy containing wind power is shown in the figure. Starting from the construction idea of ​​the method, the invention idea of ​​the method is sorted out. The specific contents are as follows:

[0119] In order to avoid the problem of unnecessary wind abandonment in the wind power distribution system caused by static current carrying capacity, the present invention considers using dynamic current carrying capacity instead of static current carrying capacity STR as a measurement indicator for the wind power distribution system; at the same time, in order to avoid the high purchase, installation and maintenance costs of some equipment in the dynamic current carrying capacity DTR system, and the difficulty in achieving full coverage of the system, the present invention uses the thermal stability margin increment calculated according to the dynamic current carrying capacity to replace the original static current carrying capacity to judge the thermal stability of the system.

[0120] 1. Derivation of the thermal stability margin increment calculation method

[0121] 1.1 Calculation of current carrying capacity of overhead lines

[0122] According to IEEE 738 standard, ignoring corona loss, etc., the current carrying capacity of overhead lines is I TR Satisfies the steady-state heat balance equation:

[0123]

[0124] In the formula, q c is the convective heat dissipation, q r represents the amount of heat dissipated by radiation, q s Indicates the amount of heat absorbed by the conductor due to sunlight, R(T c ) represents the conductor temperature T c The resistance value when the convection heat dissipation is qc It can be calculated according to three models: high wind speed, low wind speed and zero wind speed. Only in the high wind speed scenario does the wind speed have a greater impact on the current carrying capacity. In addition, the present invention only considers the thermal stability margin increment calculation of the wind power distribution network. Therefore, only the high wind speed case is considered. The convective heat dissipation q c The calculation formula is as follows:

[0125]

[0126] Among them, the convection heat dissipation coefficient k c V is the preset value related to wind angle, line diameter, air density, air thermal conductivity, air viscosity, etc. w is the wind speed, T a is the ambient temperature;

[0127] Radiative heat dissipation q r The calculation formula is as follows:

[0128] q r =k r ((T c +273) 4 -(T a +273) 4 ) (3)

[0129] Among them, the radiation heat dissipation coefficient k r It is related to the line diameter, the surface radiation coefficient of the conductor, the age of the overhead line, etc.; the conductor's sunlight absorption heat q s The calculation formula is as follows:

[0130] q s =k s (4)

[0131] Among them, the solar heat absorption coefficient k s It is related to the solar absorptivity of the conductor surface, solar irradiance and line diameter; R(T c ) is calculated as follows:

[0132] R(T c ) = k j (T c -T a )+R a (5)

[0133] Among them, the thermal sensitivity coefficient k j is a constant.

[0134] 1.2 Effect of wind speed on current carrying capacity

[0135] In strong wind weather, as the wind speed changes, Δq r <<Δq c , radiant heat dissipation q rCan be approximately constant; and the conductor solar absorption q s Because it is highly independent of wind speed, the changes affected by wind speed can be ignored. Based on the above assumptions, this paper only considers V w With T c right The impact is as follows:

[0136]

[0137] According to formula (1), the partial derivatives are

[0138]

[0139] ΔV w Indicates the current wind speed V w，t The initial wind speed V w，0 The difference can be set according to weather forecast data, ΔT c Indicates the current time T c，t With the initial time T c，0 The difference, ΔT c The non-steady-state heat balance equation established by the heat exchange process is calculated as follows:

[0140]

[0141] Among them, mC p is the heat capacity of the conductor. According to the first-order Taylor expansion: The initial state (V w,0 ,T c,0 ) into formula (1), the current state (V w,t ,T c,t ) into formula (7), considering that heat exchange is the cause of ΔT c The main reasons are get:

[0142]

[0143] Substituting formula (8) into formula (6), we get:

[0144]

[0145] Take the dynamic current carrying capacity change rate In summary, the dynamic current carrying capacity change ΔI is derived TR The formula is as follows:

[0146]

[0147] Among them, sgn() is a sign function that returns the positive or negative sign of the variable.

[0148] In the period t, the wind speed change and the line current carrying capacity change are ΔV w With ΔI TR , and ΔV w With ΔI TR There is a strong positive correlation. For low or zero wind speed scenarios, considering that the dynamic current carrying capacity DTR is minimally affected by wind speed, the static current carrying capacity STR is still used to conservatively estimate the current carrying capacity. Calculation formula (10) is applied to high wind speed conditions. At high wind speeds, the present invention uses formula (10) to calculate the change in dynamic current carrying capacity.

[0149] 1.3 Effect of wind speed on current in wind-related distribution networks

[0150] Wind speed not only affects the dynamic current carrying capacity DTR, but also affects the output power P of the wind turbine. WT Affects network flow distribution, P WT With V w The following formula can be approximately satisfied:

[0151]

[0152] Where V i 、V r With V o are the fan cut-in, rated and cut-out speeds respectively; P r is the rated power of the fan. Then P WT = 0. The present invention only focuses on V w ∈[V i ,V o ), in order to ensure the reliable operation of the distribution network, it is necessary to pay attention to the lines where the wind speed changes cause the flow to increase. WT The piecewise linear property of is introduced to calculate the wind speed change ΔV w,cal The calculation of is as follows:

[0153]

[0154] In order to ensure the safe operation of the distribution network, it is necessary to check the maximum current increase caused by wind speed changes. ΔV w >0, the fan output increases significantly, ΔV w <0, the power supply of the distribution main grid increases significantly. If the change in network loss is ignored, the change in both does not exceed ΔV w,cal The change in fan output caused by the corresponding load power change ΔP d sum:

[0155]

[0156] Remember R + , R -They are the equivalent resistance of the fan port and the equivalent resistance of the substation port, respectively. The line model used for the fan outlets is the same, so the equivalent resistance of each fan port is the same, and the substation port is the power supply access port of the distribution main network.

[0157] The square of the maximum current change in the distribution network is:

[0158]

[0159] Substituting equation (13) into equation (14), for the distribution network containing wind power, ΔV w,cal The adjacent bus load change ΔP caused by d Much smaller than the fan rated power P r , that is, ΔP d =P r , record the distribution network current change rate Then the maximum current change of the distribution network ΔI DN for:

[0160]

[0161] Therefore, regardless of whether the wind speed increases or decreases, it will cause the current in some lines to increase, causing safety hazards.

[0162] 1.4 TSM incremental index of thermal stability margin of wind power distribution network

[0163] The distribution network operation is subject to thermal stability constraints, that is, the line i current I TR Should not exceed its current carrying capacity I TR，i The thermal stability margin TSM of line i is defined as follows:

[0164] η I，i =I TR，i -I i (16)

[0165] If η is known I，i The initial value of can also be determined based on its increment:

[0166] Δη I，i =ΔI TR，i -ΔI i (17)

[0167] Where: ΔI TR，i and ΔI i are the change in the dynamic current carrying capacity DTR of line i and the change in the distribution network current relative to the initial moment. I，i Minimum value min{Δη I，i} describes the system thermal stability margin TSM increment, substituting equations (10) and (15) into equation (17), we get the thermal stability margin TSM increment index λ:

[0168]

[0169] Therefore, if we do not consider the effect of environmental factors such as solar irradiance on k TR The influence of λ is only related to V w It is related that when the value of λ is larger, the thermal stability of the distribution network is stronger.

[0170] The verification of the thermal stability of the distribution network lines is measured by the thermal stability margin increment index. The calculation of the thermal stability margin increment includes: calculating the change in the dynamic current carrying capacity of the line and the change in the maximum current of the distribution network according to the change in wind speed, and calculating the thermal stability margin increment based on the calculation results.

[0171] This paper judges the thermal stability of the system by comparing λ and its threshold Λ. When λ ≥ Λ, the system meets the thermal stability constraint; otherwise, it is necessary to adjust the dispatch strategy to reduce the current to a safe range. λ takes into account the impact of real-time wind speed and solves the problem of unnecessary wind curtailment caused by overly conservative estimation of traditional static current carrying capacity STR. 2. Constructing a framework for optimizing the dispatch strategy of a power grid with wind power

[0172] like Figure 1 As shown, in order to ensure the reliability and economy of the wind power distribution network operation, the present invention constructs the following based on the thermal stability margin TSM increment index λ: Figure 1 The framework of the power grid dispatch strategy optimization method including wind power shown is used to obtain the optimal power grid dispatch strategy.

[0173] The framework provides an inventive idea for a method for optimizing a grid dispatch strategy including wind power, which includes three steps:

[0174] Step 1: Obtain the thermal stability margin increment index of the wind power distribution network;

[0175] Step 2: Construct a multi-time scale collaborative optimization model based on the indicators in step 1;

[0176] Step 3: Solve the model in step 2.

[0177] The aforementioned first part: thermal stability margin increment calculation method has described in detail the derivation of the step one indicator. This embodiment first uses calculation formula (10) to evaluate the dynamic current carrying capacity change, uses calculation formula (15) to evaluate the maximum current change of the distribution network, and uses calculation formula (18) to calculate the thermal stability margin increment required by the present invention.

[0178] In the second step, the thermal stability margin increment is used as the thermal stability constraint condition to establish a multi-time scale collaborative optimization model, which includes three parts: reconstruction optimization, scheduling optimization and safety verification.

[0179] In step three, the three components of the model in step two are analyzed and solved using the improved golden jackal algorithm, interior point method and numerical comparison method.

[0180] 3. Multi-timescale collaborative optimization model

[0181] 3.1 Model Overview

[0182] The model of the present invention includes a reconstruction optimization part, a dispatching optimization part and a safety verification part. The control variable of the reconstruction optimization part is the distribution network switch, the control variables of the dispatching optimization part are the reactive output of the static VAR generator, the wind turbine abandoned wind power and the interruptible load active power, and the verification content of the safety verification part is the thermal stability margin increment.

[0183] In view of the different scheduling contents of each part of the model, the present invention has designed a variety of time divisions and multiple time scales, including the total period, reconstruction period, scheduling period and verification period. Each period adopts a different time scale:

[0184] The time scale of the total period is 24h;

[0185] A total period is divided into several reconstruction periods, and the time scale of the reconstruction period is several hours; a reconstruction period is divided into several scheduling periods, and the time scale of the scheduling period is 1 hour; the time scale of the verification period is 15 minutes.

[0186] Multi-timescale collaborative optimization includes:

[0187] The reconstruction optimization part updates and solves the problem according to the divided reconstruction periods as optimization cycles;

[0188] The scheduling optimization part updates and solves the optimization cycle according to the divided scheduling periods;

[0189] The safety verification part updates the thermal stability of the verification circuit according to the verification period as the optimized cycle.

[0190] 3.2 Establishing the objective function according to the optimization goal

[0191] The objective function is established based on the optimization goal of minimizing operation and control costs as follows:

[0192] min(C CT +C OP )(19)

[0193] In the formula, C CT Indicates control cost; C OP Indicates the operating cost;

[0194] 3.2.1 Control Costs C CT

[0195] Control costsCT , the switch action vector x in each period t , static VAR generator SVG reactive power output vector Q SVG,t 、Wind turbine abandoned wind AWP power vector P AWT,t and the interruptible load IL output vector P IL,t The cost and composition of each period include: the total period T r , reconstruction period T d and scheduling period T v , C CT The calculation formula is:

[0196]

[0197] In the formula, the L×1-dimensional switch action vector x t The cost and the reconstruction cost are SW,t =γ SW ·d(x t -x t-1 ), γ SW is the cost of the switching action per unit number, and the function d is used to take the vector (x t -x t-1 ) the total number of non-zero elements; N SVG ×1-dimensional static VAR generator SVG reactive output vector Q SVG,t The cost and SVG adjustment cost are SVG adjustment cost C SVG,t =γ SVG Q SVG,t , γ SVG 1×N WT Dimensional SVG unit capacity adjustment cost vector; N WT ×1-dimensional wind turbine abandoned wind AWP power vector P AWT,t The cost and the penalty for wind abandonment, wind abandonment penalty C AWT,t =γ AWT ·P AWT,t , γ AWT 1×N WT Dimensional unit wind curtailment penalty cost vector; N IL ×1-dimensional interruptible load IL active output vector P IL,t The cost and load interruption cost are load interruption cost C IL,t =γ IL ·P IL,t , γ IL 1×N IL dimensional unit load interruption cost vector.

[0198] 3.2.2 Operating Cost C OP

[0199] Operating cost OPThe cost of purchasing electricity from the wind power station area to the main grid in each period is C G,t 、Wind power operation cost C WT,t and distribution network loss cost C LS,t Composition, the calculation formula is:

[0200]

[0201] Where, the cost of purchasing electricity from the main grid in the wind power station area is C G,t =γ G ·p G,t , γ G is the unit capacity electricity purchase cost, p G,t The access point exchanges active power with the main grid; the wind power operation cost C WT,t =γ WT ·P WT,t , γ WT 1×N WT The power generation cost vector of the wind turbine per unit capacity, P WT,t N WT ×1-dimensional wind turbine active output vector; distribution network loss cost C LS,t =γ LS ·P LS,t , γ LS is the unit line loss cost, P LS,t The total loss of the power distribution network refers to the energy loss when the current is transmitted in the power distribution network. The total loss of the power distribution network in the present invention is the active power loss;

[0202] 3.3 Constraints

[0203] The constraints of the multi-time scale collaborative optimization model include power flow equality constraints, control variable controllability constraints and state variable safety constraints.

[0204] 3.3.1 Power flow equality constraints

[0205] Define the N×1-dimensional node active load vector as P D,t , each element corresponds to the active load of each node; the N×1-dimensional active injection power vector is f P , each element corresponds to the active injection power of each node; the N×1-dimensional reactive load vector is Q D,t , each element corresponds to the reactive load of each node; the N×1-dimensional reactive injection power vector is f Q , each element corresponds to the reactive injection power of each node.

[0206] p G,t The access point exchanges active power with the main network; P WT,t N WT ×1-dimensional wind turbine active output vector, each element corresponds to the active output of each wind turbine; PAWT,t N WT ×1-dimensional wind turbine abandoned wind AWP power vector; q G,t Exchange reactive power between the access point and the main network; Q WT,t N WT ×1-dimensional wind turbine reactive power vector, each element corresponds to the reactive power output of each wind turbine; Q SVG,t N SVG ×1-dimensional static VAR generator SVG reactive power output vector, the total number of wind turbines in this embodiment is the same as the total number of static VAR generators SVG.

[0207] The association matrix consists of 0s and 1s, and only the numbers in the rows where 1s are located represent the access point nodes of the corresponding components.

[0208] A value of 1 indicates that the node with the corresponding number in this row is the access point of the corresponding component; C G is the N×1-dimensional node-main network access point association matrix; C WT is N×N WT Dimensional node-wind turbine access point association matrix; C IL is N×N IL Node-interruptible load phase correlation matrix; C SVG is N×N SVG Dimensional node-SVG access point association matrix. The power flow equality constraint is as follows:

[0209]

[0210] Where, the active injection power vector f P Including the active injection power of N nodes, the reactive injection power vector f Q Includes the reactive injection power of N nodes, where i and j are positive integers within the value range of N.

[0211] The injected active power f of node j P,j The calculation formula is:

[0212] f P,j =u j ∑ j∈i u i (g ij cosθ ij +b ij sinθ ij );

[0213] The injected reactive power f of node j Q,j The calculation formula is:

[0214] f Q,j =u j ∑ j∈i u i (g ij sinθij -b ij cosθ ij );

[0215] Among them, u j and u i is the voltage between node j and node i; θ ij is the phase angle difference between nodes i and j; g ij is the conductance between nodes i and j; b ij Susceptance between nodes i and j.

[0216] 3.3.2 Controllability constraints of control variables

[0217] I) Switching action constraints

[0218] The reconstruction of the partial distribution network of the present invention should satisfy that after applying the reconstructed distribution network switching strategy, the topology is still radial and the number of switching actions in the total time period is less than the upper limit. The constraint function of the switching action in the control variable is expressed as follows:

[0219]

[0220] In the formula, h(x t ) is the radial topological constraint function, h(x t )=0 means the execution switch action vector x t After the switching strategy is implemented, the distribution network satisfies the radial topology constraint; the function d is used to take the vector (x t -x t-1 ), TSC is the total number of non-zero elements in the r The upper limit of the number of switch actions allowed within a cycle.

[0221] II) Fan Constraints

[0222] The constraints on fan output in the control variables are as follows:

[0223]

[0224] Where, 0 WT N WT ×1-dimensional all-zero vector; P AWT,max,t N WT ×1-dimensional wind turbine abandoned wind power upper limit vector, each element corresponds to the wind turbine abandoned wind power upper limit value; P WT,max,t N WT ×1-dimensional wind turbine active output upper limit vector, each element of which corresponds to the active output upper limit value of each wind turbine; S WT,max N WT ×1-dimensional wind turbine capacity upper limit vector, each element corresponds to the upper limit value of each wind turbine capacity; P WT,t N WT×1-dimensional wind turbine active output vector; Q WT,t N WT ×1-dimensional wind turbine reactive power vector; β lag N WT dimensional lag reduction factor diagonal matrix, the ath diagonal element of which is β lag,a , a is N WT A positive integer within the value range, the minimum lagging power factor of the ath fan is λ lag,a , then β lead N WT dimensional forward reduction factor diagonal matrix, whose ath diagonal element is β lead,a , the minimum leading power factor of the ath fan is λ lead,a , then

[0225] III) Static VAR generator constraints

[0226] The constraints on the reactive power output of the static VAR generator SVG in the control variables are as follows:

[0227] Q SVG,min ≤Q SVG,t ≤Q SVG,max (25)

[0228] In the formula, Q SVG,max N SVG ×1-dimensional static VAR generator SVG reactive output upper limit vector, each element of which corresponds to the reactive output upper limit value of each static VAR generator SVG; Q SVG,min N SVG ×1-dimensional reactive power lower limit vector of the static VAR generator SVG, each element of which corresponds to the reactive power lower limit value of each static VAR generator SVG.

[0229] IV) Load Constraint

[0230] The constraints of interruptible load in the control variables are as follows:

[0231] P IL,min ≤P IL,t ≤P IL,max (26)

[0232] Where P IL,max N IL ×1-dimensional interruptible load active output upper limit vector, each element of which corresponds to the active output upper limit value of each interruptible load; P IL,min N IL ×1-dimensional interruptible load active output lower limit vector, each element of which corresponds to the lower limit value of each interruptible load active output.

[0233] 3.3.3 State variable security constraints

[0234] The safety constraints of power distribution state variables include voltage safety and thermal stability safety. The safety constraints of state variables are as follows:

[0235]

[0236] Where U min is an N×1-dimensional node voltage lower limit vector, each element of which corresponds to the lower limit value of each node voltage; U max is an N×1-dimensional node voltage upper limit vector, each element of which corresponds to the upper limit value of each node voltage; U t represents the N×1-dimensional node voltage vector in time period t, and each element corresponds to the voltage of each node in time period t; λt is the thermal stability margin increment of the line at time t, and Λ is the thermal stability margin increment threshold, which is a manually set value.

[0237] 4.1 Solution Process Overview

[0238] The solution process of the multi-time scale collaborative optimization model is as follows: Figure 2 As shown, it includes initialization, reconstruction optimization solution, scheduling optimization solution and thermal stability verification.

[0239] The solution process is summarized as follows:

[0240] The grid topology and operating condition information including wind power is obtained, and all the interconnection and section switches of the distribution network are closed to form several ring networks. After the clustering scenario is determined according to the source-load curve, the model is optimized on a rolling basis with multiple hours as the reconstruction period, 1 hour as the scheduling period, and 15 minutes as the verification period. The improved golden jackal algorithm is used to improve the solution efficiency of the reconstruction optimization part, and the interior point method is used to improve the solution reliability of the scheduling optimization part, so as to obtain the optimal grid scheduling strategy.

[0241] Among them, operating condition information refers to a series of data and indicators that reflect the actual operating conditions of the distribution network, mainly including node information, branch information and generator information of the distribution network.

[0242] 4.2 Source-load curve determines clustering scenario

[0243] The "source" in the source-load curve refers to the output curve of new energy, and in this invention, it refers to the output curve of wind turbines. The "load" in the source-load curve refers to the curve of load (i.e., power demand), and in this invention, it refers to the load curve. This process aims to simplify the complex source-load curve data into limited, representative typical scenarios through clustering technology, so as to facilitate the subsequent planning, operation, prediction and management of power systems.

[0244] The present invention predicts the source-load curve of the next day (next 24 hours), determines the number of clustering periods based on the net load curve formed by superimposing the load and the fan output by the elbow method, and obtains the clustering scenario by clustering using the k-means clustering algorithm according to the determined number of clustering periods. To simplify the analysis, the simulation does not consider the prediction deviation of the fan and the load, and ignores the difference in wind speed and fan output in the distribution network area.

[0245] The clustering scenarios of each reconstruction period are determined according to the source-load curve. In each generated scenario, the improved golden jackal algorithm is used to reconstruct the distribution network switching strategy in each ring network to solve the distribution network switching strategy after reconstruction and optimization in each reconstruction period of the reconstruction optimization part.

[0246] For each reconstruction period corresponding to a type of scenario, the improved golden jackal algorithm of the present invention adopts a population mutation mechanism and a nearest neighbor search mechanism to improve the traditional golden jackal algorithm, and the distribution network switching strategy is to reconstruct and change the opening and closing mode of the distribution network connection and section switches.

[0247] The improved golden jackal algorithm of the present invention can improve the global convergence performance.

[0248] The basic process of the traditional golden jackal algorithm usually includes the following steps:

[0249] Population initialization: Randomly generate a certain number of individuals as the initial population, and determine the upper and lower boundaries of the search space.

[0250] Fitness evaluation: According to the fitness function (or objective function) of the problem, the fitness value of each individual is calculated to evaluate its quality.

[0251] The improved golden jackal algorithm of the present invention uses the distribution network bus loss corresponding to the distribution network switch strategy as the fitness value to evaluate the fitness.

[0252] Individual selection and update: Selection is made based on the fitness value of the individual, and individuals with higher fitness are usually selected as male jackals (optimal solution) and female jackals (suboptimal solution). Then, the positions of other individuals are updated according to the hunting behavior of the golden jackal to generate new and potentially better solutions.

[0253] Iteration and convergence: Repeat the above steps until the number of iterations reaches the upper limit, the algorithm stops iterating and outputs the optimal solution.

[0254] 4.3 Population variation mechanism

[0255] In order to improve the global search capability of the algorithm, the population mutation mechanism first uses Tent mapping to improve the uniformity of the initial population, and then increases the randomness of the evolving population through adaptive t-distribution mutation, so that the population space characteristics take into account both uniformity and randomness.

[0256] 4.3.1Tent Mapping

[0257] The traditional golden jackal algorithm generates an initial population, and the tent mapping is used to improve the uniformity of the initial population to generate an evolved population.

[0258] The initial population is composed of a plurality of individuals, and the individuals are represented as switch action vectors. The generation of the evolutionary population by Tent mapping includes:

[0259] Perform tent mapping on each individual in the initial population to generate an individual of the evolutionary population, where the individual x of the evolutionary population 0,n Satisfy the following formula:

[0260]

[0261] In the formula, x max is the switch action vector of the upper limit of the L×1-dimensional ring network number, and each element thereof corresponds to the action state of the switch of the upper limit of the number in each ring network; min is the switch action vector of the lower limit of the number of the L×1-dimensional ring network, and each element thereof corresponds to the action state of the switch of the lower limit number in each ring network; is the Hadamard product, which is used to multiply the elements of each group of corresponding positions in two matrices of the same order (that is, the number of rows and columns of the two matrices are the same) to obtain a new matrix; ω n is an L×1 dimensional random vector, any element ω n,i ∈[0,1]; Tent() is the Tent mapping function, and the calculation formula is as follows:

[0262]

[0263] 4.3.2 Adaptive t distribution variation

[0264] The current evolutionary population is modified by adaptive t distribution to increase the randomness of the evolutionary population and generate a mutant population.

[0265] The adaptive t-distribution mutation to generate a mutant population includes:

[0266] Individual x of the evolutionary population at the current kth iteration k,n After the adaptive t distribution mutation, the corresponding individuals of the mutant population are generated The calculation is as follows:

[0267]

[0268] Where K is the total number of iterations, the t-distribution function Trnd(k) takes k as the parameter degree of freedom, and the coefficient of variation (1-k / (K-1)) takes into account the influence of k on the t-distribution, so that the adaptive t-distribution variation has good global search ability in the initial stage and converges to the optimal solution with the iteration process.

[0269] In the early stage of iteration, the adaptive t-distribution mutation is similar to the Cauchy mutation, which has good global search ability; as the k value increases, the adaptive t-distribution mutation is similar to the Gaussian mutation, which has good local search ability. Therefore, the adaptive t-distribution mutation combines the advantages of Cauchy mutation and Gaussian mutation to avoid falling into partial optimality and converge to the global optimality.

[0270] After obtaining the evolutionary population generated by Tent mapping and the mutant population generated by adaptive t distribution mutation, the fitness of the individuals in the two populations is calculated. Since the individuals in the evolutionary population correspond to the individuals in the mutant population, there is a corresponding relationship between the individuals in the two populations. The individuals in the corresponding positions of the two populations are grouped in pairs. The fitness of each group of corresponding individuals in the two populations is compared, and the individuals with high fitness in each group of corresponding individuals are retained to form the optimal population.

[0271] Among them, the switch action vector represents the distribution network switching strategy, and the calculation of individual fitness includes:

[0272] After applying the distribution network switch strategy to the distribution network, the corresponding distribution network bus loss is calculated, and the calculated distribution network bus loss is used to measure the individual fitness, and the individual fitness is negatively correlated with the corresponding distribution network bus loss.

[0273] The individuals of the population of the present invention are represented by switch action vectors, which represent the switching strategy of the distribution network. The distribution network switching strategy is applied to the distribution network to calculate the corresponding distribution network bus loss. The distribution network bus loss can measure the fitness of the individuals. The individuals with high distribution network bus loss have high fitness, and the individuals with low distribution network bus loss have high fitness. Therefore, the individual fitness is negatively correlated with the corresponding distribution network bus loss.

[0274] 4.4 Neighbor Search Mechanism

[0275] The traditional golden jackal optimization algorithm simulates the hunting characteristics of golden jackals, where male jackals lead and female jackals follow.

[0276] Compare the individual fitness of the generated optimal population and select the individual with the highest fitness as the male jackal x m,k,n , select the individual with the second highest fitness as the female jackal x f,k,n ;

[0277] The positions of individuals in the population are updated according to the prey escape energy, and the updated individuals are organized into a prey population, including the following:

[0278] Calculate the prey escape energy. The calculation formula of the prey escape energy E is as follows:

[0279]

[0280] Where E0 is the initial avoidance energy randomly generated by the traditional golden jackal algorithm, and E0∈(0,1).

[0281] Updating the positions of individuals in the population according to the calculated prey escape energy includes:

[0282] Male Jackal m,k,n Used to generate individuals in the optimal population in step S224 Corresponding male jackal prey x mp,k,n ;

[0283] Female jackal fp,k,n Used to generate individuals in the optimal population in step S224 Corresponding female jackal prey x fp,k,n ;

[0284] According to the generated male jackal prey x mp,k,n and female jackal prey x fp,k,n Calculate individual The individual x corresponding to the prey population k+1,n ;

[0285] The calculation process of the position of individuals in the prey escape energy update population is as follows:

[0286] When E ≥ 1, the prey escape energy is high, the golden jackal is in the stage of exploring prey, and the male jackal prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing:

[0287]

[0288] When E<1, the prey's escape energy decreases, and the golden jackal is in the stage of besieging the prey. mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing:

[0289]

[0290] In the formula, the Levy distribution random vector rl and the individual in the optimal population Vectors of the same order, is the individual of the optimal population after the kth iteration; collect the updated individuals x k+1,n Form a prey population.

[0291] According to the calculation formulas (32) and (33), each individual in the optimal population Can be updated to an individual x k+1,n , all individuals x after update k+1,n Able to form prey groups.

[0292] In order to avoid the limitation of male and female jackals that makes the algorithm fall into local optimum, the improved golden jackal algorithm combines the characteristics of reconstruction optimization part, introduces the nearest neighbor search mechanism, and k+1,n Randomly add or subtract one from the elements in to get x k+1,n The nearest neighbor scheme for an element in .

[0293] Among them, x k+1,n is an L×1 dimensional switch action vector, L is the number of ring networks, each element corresponds to the number of the switch in each ring network, and the distribution network switch strategy adjustment content represented by the random addition or subtraction of any element includes: closing the switch corresponding to the element, and randomly opening an adjacent switch on either side of the left or right side of the loop where the switch is located.

[0294] Traverse x k+1,n To obtain x k+1,n After the nearest neighbor scheme of each element in, x k+1,n The distribution network switching strategies corresponding to the obtained neighboring solutions are applied to the distribution network respectively, and the corresponding individual fitness is calculated and compared. k+1,n The distribution network switching strategy with the highest fitness among its neighboring solutions is used as the switching action vector corresponding to this strategy as the new x k+1,n .

[0295] Each individual x in the prey population k+1,n Contains L elements, traverses each individual x k+1,n L nearest neighbor solutions can be obtained, because x k+1,n is an L×1-dimensional switch action vector, then its nearest neighbor solution is also an L×1-dimensional switch action vector. The distribution network switching strategies corresponding to these L+1 switch action vectors are applied to the distribution network respectively, and the values ​​of the L+1 distribution network bus losses can be calculated accordingly. Based on these L+1 values, the individual x k+1,n The fitness of its nearest neighbor solutions is determined, the distribution network switching strategy with the highest fitness is retained, and the switching action vector with the highest fitness is used as the new individual x of the prey population. k+1,n .

[0296] Each individual x in the prey population k+1,n New individuals can be obtained by the above methods, and these new individuals are collected and combined into a new evolutionary population.

[0297] At this time, it is necessary to determine the new individual x of the prey population k+1,n Whether the corresponding number of iterations k reaches the upper limit of the total number of iterations K,

[0298] If the number of iterations does not reach the upper limit, the number of iterations k is increased by 1 and the newly generated evolving population is used as the current evolving population to start the next iteration, and adaptive t-distribution mutation is performed on the individuals of the current evolving population.

[0299] If the number of iterations reaches the upper limit, the new individual x of the prey population is output k+1,n The distribution network switching strategy with the highest fitness in the solution is used as the distribution network switching strategy for a reconstruction period.

[0300] The above Sections 4.3 to 4.4 are the process of solving the distribution network switching strategy for a reconstruction period. From Section 4.2, it can be seen that the total period is divided into several reconstruction periods. The distribution network switching strategy for other reconstruction periods can be solved according to the solution process from Sections 4.3 to 4.4.

[0301] 4.5 Scheduling Optimization Solution and Thermal Stability Verification

[0302] 4.5.1 Scheduling Optimization Solution

[0303] The distribution network switching strategy solved in Section 4.4 is applied to the distribution network. After the distribution network topology is updated, in each reconstruction period, the scheduling period is used as the optimization cycle, and the reactive output of the static VAR generator, the wind turbine abandoned wind power and the active power of the interruptible load are adjusted as the control means to solve the scheduling optimization part, and the optimized grid scheduling strategy containing wind power is updated every 1 hour.

[0304] The process of solving the scheduling optimization part includes: using the Matlab Optimization Toolbox solver, relaxing the variables with the interior point method, and obtaining the optimized power grid scheduling strategy containing wind power. The present invention completes the process of solving the scheduling optimization part quickly within a few seconds each time, so the power grid scheduling strategy solved in real time at the beginning of each hour and applied, and the power grid scheduling strategy solved again when the thermal stability check of the distribution network line fails, will not cause delays in updating the power grid scheduling strategy.

[0305] The interior point method can efficiently solve large-scale optimization problems. As the scale of the problem increases, the computational complexity will not increase dramatically. It can avoid boundary search, and the optimization is always carried out within the feasible domain, avoiding the complexity and instability that may be encountered on the boundary. It can also maintain good numerical stability and can find a relatively accurate optimal solution or approximate optimal solution under various constraints.

[0306] 4.5.2 Thermal stability check

[0307] The grid dispatching strategy solved in Section 4.5.1 is applied to the distribution network. In each dispatching period, the verification period is used as the optimization period, and the thermal stability of the distribution network lines is verified using the safety verification part.

[0308] The process of checking the thermal stability of distribution network lines includes:

[0309] The thermal stability margin increment λ is updated every 15 minutes, and a numerical comparison is performed. By comparing the thermal stability margin increment λ with the setting threshold, it can quickly determine whether the thermal stability of the line under the control of the power grid dispatching strategy solved in Section 4.5.1 has passed the verification. If the calculated thermal stability margin increment is within the setting threshold range, the thermal stability verification is passed; if the calculated thermal stability margin increment is not within the setting threshold range, the thermal stability verification is not passed, and it is necessary to adjust the control means in Section 4.5.1 to solve the new power grid dispatching strategy.

[0310] 4.5 Rolling Optimization and Solution

[0311] The method of the present invention can perform rolling optimization and solve the multi-time scale collaborative optimization model to obtain an optimized wind power grid dispatching strategy, including:

[0312] Solving the reconstruction optimization part in the previous total period, obtaining the distribution network switching strategy of each reconstruction period, and applying the obtained distribution network switching strategy of each reconstruction period to the current total period accordingly;

[0313] At the beginning of each dispatching period in each reconstruction period of the current total period, solving the dispatching optimization part, obtaining the grid dispatching strategy containing wind power, and applying the obtained grid dispatching strategy containing wind power to the current dispatching period;

[0314] In the current total period, the verification period is used as the optimization cycle of the safety verification part, and the thermal stability margin increment is cyclically calculated and the thermal stability of the distribution network lines is verified;

[0315] The reconstruction optimization part is solved in the current total period, the distribution network switching strategy of each reconstruction period is obtained, and the distribution network switching strategy of each reconstruction period obtained is correspondingly applied to the next total period.

[0316] In this embodiment, the method of the present invention solves the reconstruction optimization part 24 hours in advance, obtains the distribution network switch strategy of each reconstruction period, and when the time enters the current 24 hours, the reconstruction period divided in advance is used as the optimization cycle, and the distribution network open-loop strategy of the corresponding reconstruction period obtained is applied to the distribution network to update the distribution network topology;

[0317] In each reconstruction period of the current 24 hours, the dispatch optimization part is solved at the beginning of each hour to obtain the grid dispatch strategy containing wind power. When the time enters the current hour, the obtained grid dispatch strategy containing wind power is applied to the current hour.

[0318] In the current 24 hours, the thermal stability of the distribution network lines is cyclically calculated and verified with 15 minutes as the optimization cycle of the safety verification part;

[0319] If the thermal stability check of the distribution network line is passed, the grid dispatching strategy containing wind power in the current dispatching period will be used; if the thermal stability check of the distribution network line is not passed, the grid dispatching strategy containing wind power in the current dispatching period will be re-solved and applied, with 15 minutes as the optimization cycle of the safety check part, and the thermal stability of the distribution network line will be cyclically calculated and checked;

[0320] In the current 24 hours, the reconstruction optimization part is solved to obtain the distribution network switching strategy for each reconstruction period. When the time enters the next 24 hours, the reconstruction period divided by the current 24 hours is used as the optimization cycle, and the distribution network switching strategy for the corresponding reconstruction period is applied to the distribution network to update the distribution network topology.

[0321] Example 2

[0322] Based on the same inventive concept as Example 1, this example introduces a test method for testing the optimization effect of the method of the present invention.

[0323] refer to Figure 3 The IEEE-33 node power distribution system topology diagram used for testing the effect is shown in the present invention. Figure 3 The distribution system model shown tests the optimization effect of the method of the present invention, wherein nodes 34-38 are wind turbine access nodes introduced in the present invention, WT represents wind turbine, nodes 18 / 22 / 25 / 33 are interruptible load access points in the present invention, and IL represents interruptible load.

[0324] In the initialization phase, all disconnectors and tie switches on the distribution network are closed to form a Figure 3 There are 5 ring networks shown, so the switch action vector of this embodiment is 5×1 dimensional.

[0325] In the previous 24 hours, according to the predicted source-load curve of the current 24 hours, the elbow method is used to superimpose the load and the wind turbine output to form a net load curve to determine the number of clustering periods as 5. According to the determined number of clustering periods, the k-means clustering algorithm is used to cluster and obtain the clustering scenario. To simplify the analysis, the simulation does not consider the prediction deviation of the wind turbine and the load, and ignores the difference in wind speed and wind turbine output in the distribution network area.

[0326] The total time period T used in the present invention r is 24h, the total time period T r Divided into 5 reconstruction periods T d : 4, 4, 10, 3, 3; each reconstruction period T d Corresponding to a type of scene.

[0327] In the first reconstruction period from 0 to 4 hours, the traditional golden hairpin algorithm is used to generate the initial population, which contains 30 individuals, and each individual has a 5×1 dimensional switch action vector;

[0328] Each individual is tent mapped separately, and the obtained individuals form an evolutionary population, which includes 30 individuals;

[0329] Each individual in the current evolutionary population is subjected to adaptive t-distribution mutation, and the obtained individuals constitute the mutation population, which includes 30 individuals;

[0330] The 30 individuals in the evolutionary population correspond to the 30 individuals in the mutation population. They form a group of two to form 30 groups of corresponding individuals. The fitness of these 30 groups of individuals is calculated respectively. The individuals with high fitness in each group are retained to form a preferential population. The preferential population includes 30 individuals.

[0331] Calculate the escape energy E, and choose the calculation method corresponding to exploring prey or besieging prey according to the calculation result of E. When E≥1, the prey escape energy is high, the golden jackal is in the stage of exploring prey, and the male jackal prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update the individuals of the optimal population according to calculation formula (32); when E<1, the prey escape energy decreases, the golden jackal is in the stage of besieging the prey, and the male jackal prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update the individuals of the optimal population according to calculation formula (33). Update and combine the individuals of the optimal population into a prey population, which includes 30 individuals;

[0332] Perform a nearest neighbor search on any element of any individual in the prey population to obtain a nearest neighbor solution;

[0333] The nearest neighbor search is performed to obtain the nearest neighbor solution for each element of the individual. The nearest neighbor solution is a 5×1 dimensional switch action vector.

[0334] Furthermore, we obtain the nearest neighbor solutions for each individual in the prey population. One individual has 5 nearest neighbor solutions.

[0335] Calculate the fitness of each individual in the prey population and the corresponding neighbor solution;

[0336] The individual with the highest fitness among the six nearest neighbor solutions of each prey population is retained as the new individual; the one with the highest fitness among all individuals and nearest neighbor solutions is the optimal distribution network switching strategy at this time;

[0337] 30 new individuals form a new evolutionary population, and the number of iterations is the initial value 0;

[0338] The number of iterations is increased by 1, and the new evolutionary population is used as the current evolutionary population. The iterative optimization is performed until the total number of iterations reaches 50. The distribution network switching strategy with the highest fitness among all individuals and neighboring schemes at the 50th iteration is output, which is the distribution network switching strategy in the first reconstruction period from 0 to 4 hours.

[0339] Using the same solution process, the distribution network switching strategy in the second reconstruction period from 4 to 8 o'clock, the distribution network switching strategy in the third reconstruction period from 8 to 18 o'clock, the distribution network switching strategy in the fourth reconstruction period from 18 to 21 o'clock, and the distribution network switching strategy in the fifth reconstruction period from 21 to 24 o'clock are output.

[0340] At 0:00 on the same day, the reactive output of the static VAR generator, the wind turbine abandoned power and the active power of the interruptible load are adjusted as control means. The Matlab Optimization Toolbox solver is used to relax the variables with the interior point method to solve the scheduling optimization part and obtain the optimized grid scheduling strategy containing wind power. The obtained grid scheduling strategy containing wind power is applied during the scheduling period from 0:00 to 1:00 on the same day.

[0341] At 0-4 o'clock on the same day, the distribution network switch strategy of the first reconstruction period obtained the previous day is applied to update the distribution network topology. The first reconstruction period is 4 hours. Under this topology:

[0342] From 0:00 to 1:00 on the same day, the power grid dispatch strategy obtained by solving the dispatch optimization part at 0:00 on the same day is applied;

[0343] At 1-2 o'clock on the same day, the power grid dispatching strategy obtained by solving the dispatching optimization part at 1 o'clock on the same day is applied;

[0344] At 2-3 am on the same day, the grid dispatch strategy obtained by solving the dispatch optimization part at 2 am on the same day is applied;

[0345] At 3-4 o'clock on the same day, the power grid dispatching strategy obtained by solving the dispatching optimization part at 3 o'clock on the same day is applied.

[0346] At 4 a.m. on the same day, the first reconstruction period ended, and the distribution network topology was updated according to the distribution network switching strategy of the second reconstruction period. The topology structure also lasted for one reconstruction period. Under this topology structure, the power grid dispatching strategy continued to be optimized every hour, and the second reconstruction period was 4 hours.

[0347] At 8 a.m. on the same day, the second reconstruction period ended, and the distribution network topology was updated according to the distribution network switching strategy of the third reconstruction period. The topology structure also lasted for one reconstruction period. Under this topology structure, the power grid dispatching strategy continued to be optimized every hour. The third reconstruction period was 10 hours.

[0348] At 18:00 on the same day, the third reconstruction period ended, and the distribution network topology was updated according to the distribution network switching strategy of the fourth reconstruction period. The topology structure also lasted for one reconstruction period. Under this topology structure, the power grid dispatching strategy continued to be optimized every hour. The fourth reconstruction period was 3 hours.

[0349] At 21:00 on the same day, the third reconstruction period ended, and the distribution network topology was updated according to the distribution network switching strategy of the fifth reconstruction period. The topology structure also lasted for one reconstruction period. Under this topology structure, the power grid dispatching strategy continued to be optimized every hour. The fifth reconstruction period was 3 hours.

[0350] At 24:00 on the same day, the fifth reconstruction period ended.

[0351] refer to Figure 2 As shown, from 0 to 24 hours of the day, M reconstruction periods are divided according to the predicted 24-hour source-load curve clustering scenario of the next day, and the reconstruction optimization part is solved cyclically to obtain the distribution network switching strategies of the M reconstruction periods, and the distribution network switching strategies of these M reconstruction periods are output. When the time enters, the distribution network switching strategies of these M reconstruction periods are applied to the corresponding time.

[0352] At 24:00 on the same day, after the fifth reconstruction period ends, it enters 0:00 on the next day. The distribution network switching strategy of the first reconstruction period among the M reconstruction periods calculated from 0:00 to 24:00 on the same day and the power grid dispatching strategy obtained by solving the dispatching optimization part at 0:00 on the next day are applied.

[0353] In the above manner, the multi-time scale collaborative optimization model of the present invention can be optimized and solved in a rolling manner to obtain an optimized grid dispatching strategy containing wind power.

[0354] In order to improve the safety of the distribution network in the rolling optimization, the present invention updates and calculates the thermal stability margin increment λ every 15nin, and adopts a numerical comparison method to compare the thermal stability margin increment λ and the setting threshold Λ. The setting threshold in this embodiment is preset to -4.5. When the calculated thermal stability margin increment is λ≥Λ, that is, λ≥-4.5, the thermal stability check of the distribution network line is passed.

[0355] Example 3

[0356] Based on the same inventive concept as Example 1, this embodiment introduces a wind power grid dispatch strategy optimization device, including a model building module, a switch strategy module, a dispatch strategy module, a strategy execution module, a thermal stability module and a strategy output module;

[0357] The model building module is used to build a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of the operating cost and the control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost;

[0358] The switch strategy module is used to collect the topology and operating condition information of the power grid containing wind power, and reconstruct and optimize the power grid topology containing wind power to obtain the switch strategy of the distribution network after reconstruction and optimization;

[0359] The dispatching strategy module is used to adjust the reactive output of the static VAR generator, the abandoned wind power of the wind turbine and the active power of the interruptible load as a control means in the distribution network according to the distribution network switching strategy of the switching strategy module, so as to obtain an optimized grid dispatching strategy containing wind power;

[0360] The strategy execution module is used to perform grid dispatching using the grid dispatching strategy containing wind power obtained in step S3;

[0361] The thermal stability module is used to check the thermal stability of the distribution network line using the thermal stability margin increment index;

[0362] The strategy output module is used to output the wind power-containing power grid dispatching strategy after the thermal stability check in the thermal stability module if the thermal stability check of the distribution network line is qualified.

[0363] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0364] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0365] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0366] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0367] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for optimizing a power grid dispatching strategy including wind power, characterized in that: include: S1: construct a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of operating cost and control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost; S2: Collect the grid topology and operating condition information including wind power, reconstruct and optimize the grid topology including wind power, and obtain the switching strategy of the distribution network after reconstruction and optimization; S3: In the distribution network, according to the distribution network switch strategy obtained in step S2, the reactive output of the static VAR generator, the wind turbine abandoned wind power and the interruptible load active power are adjusted to obtain an optimized grid dispatching strategy including wind power; S4: Performing grid dispatching using the grid dispatching strategy including wind power obtained in step S3; S5: Use the thermal stability margin increment index to check the thermal stability of the distribution network line; S6: If the thermal stability check of the distribution network line is qualified, output the grid dispatching strategy including wind power after the thermal stability check in step S5.

2. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 1, characterized in that: The multiple time scales include the total period, the reconstruction period, the scheduling period and the verification period using multiple time scales; A total period is divided into several reconstruction periods; A reconstruction period is divided into several scheduling periods; A scheduling period is divided into several verification periods.

3. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 2 is characterized in that the steps In S1, in the multi-time scale collaborative optimization model, the objective function is established according to the optimization goal of minimizing the operation and control costs as follows: min(C CT +C OP ); In the formula, C CT Indicates control cost; C OP Indicates the operating cost; Control costs CT , the switch action vector x in each period t , static VAR generator SVG reactive power output vector Q SVG,t 、Wind turbine abandoned wind AWP power vector P AWT,t and the interruptible load IL output vector P IL,t The cost and composition of each period include: the total period T r , reconstruction period T d and scheduling period T v , C CT The calculation formula is: In the formula, the L×1-dimensional switch action vector x t The cost and the reconstruction cost are SW,t =γ SW ·d(x t -x t-1 ), γ SW is the cost of the switching action per unit number, and the function d is used to take the vector (x t -x t-1 ) the total number of non-zero elements; N SVG ×1-dimensional static VAR generator SVG reactive output vector Q SVG,t The cost and SVG adjustment cost are SVG adjustment cost C SVG,t =γ SVG Q SVG,t , γ SVG 1×N WT Dimensional SVG unit capacity adjustment cost vector; N WT ×1-dimensional wind turbine abandoned wind AWP power vector P AWT,t The cost and the penalty for wind abandonment, wind abandonment penalty C AWT,t =γ AWT ·P AWT,t , γ AWT 1×N WT Dimensional unit wind curtailment penalty cost vector; N IL ×1-dimensional interruptible load IL active output vector P IL,t The cost and load interruption cost are load interruption cost C IL,t =γ IL ·P IL,t , γ IL 1×N IL Dimensional unit load interruption cost vector; Operating cost OP The cost of purchasing electricity from the wind power station area to the main grid in each period is C G,t 、Wind power operation cost C WT,t and distribution network loss cost C LS,t Composition, the calculation formula is: In the formula, the cost of purchasing electricity from the main grid in the wind power station area is C G,t =γ G ·p G,t , γ G is the unit capacity electricity purchase cost, p G,t The access point exchanges active power with the main grid; the wind power operation cost C WT,t =γ WT ·P WT,t , γ WT 1×N WT The power generation cost vector of the wind turbine per unit capacity, P WT,t N WT ×1-dimensional wind turbine active output vector; distribution network loss cost C LS,t =γ LS ·P LS,t , γ LS is the unit line loss cost, P LS,t is the bus loss of the power distribution network.

4. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 2 is characterized in that: In step S2, it includes: S21: Obtain the grid topology and operating condition information including wind power, close all the connection and section switches of the distribution network so that the distribution lines form several ring networks; S22: determining clustering scenarios of each reconstruction period according to the source-load curve, and in each generated scenario, using the improved golden jackal algorithm to obtain the distribution network switching strategy of each ring network after reconstruction optimization in the reconstruction period; The source-load curve includes a source curve and a load curve, the source curve is a wind turbine output curve, and the load curve is a load curve; Each reconstruction period corresponds to a type of scenario. The improved golden jackal algorithm includes a population mutation mechanism and a neighbor search mechanism. The reconstruction distribution network switch strategy is to reconstruct and change the opening and closing mode of the distribution network connection and section switches.

5. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 4 is characterized in that: In step S22, the process of solving the population variation mechanism includes: S221: The traditional golden jackal algorithm generates an initial population, and the tent mapping is used to improve the uniformity of the initial population to generate an evolved population; The initial population is composed of multiple individuals, each of which is represented by a switch action vector. The process of generating an evolutionary population by Tent mapping includes: Perform tent mapping on each individual in the initial population to generate individuals in the evolutionary population. The individuals x in the evolutionary population 0,n Satisfy the following formula: In the formula, x min is the switch action vector of the L×1-dimensional ring network number lower limit, and its elements correspond to the number lower limits of the switches in each ring network. max is the switch action vector of the upper limit of the number of the L×1-dimensional ring network, and each element thereof corresponds to the upper limit of the number of switches in each ring network. is the Hadamard product, which is used to multiply the elements of each group of corresponding positions in two matrices of the same order to obtain a new matrix, ω n is an L×1 dimensional random vector, any element ω n,i ∈[0,1], Tent() is the Tent mapping function, and the calculation formula is as follows: S222: Generate a mutant population by increasing the randomness of the evolving population through adaptive t-distribution mutation of the current evolving population; The process of generating a mutant population by adaptive t-distribution mutation includes: Individual x of the evolutionary population at the current kth iteration k,n After the adaptive t distribution mutation, the corresponding individuals of the mutant population are generated The calculation is as follows: Where K is the total number of iterations, and the t distribution function Trnd(k) takes k as the parameter degree of freedom; S223: Calculate the individual fitness of the evolution population in step S221 and the mutation population in step S222, and compare the fitness of each group of corresponding individuals in the two populations; Among them, the switch action vector represents the distribution network switching strategy, and the calculation of individual fitness includes: Applying the distribution network switch strategy to the distribution network and calculating the corresponding distribution network bus loss; The individual fitness is measured by using the calculated total loss of the power distribution network, and the individual fitness is negatively correlated with the corresponding total loss of the power distribution network; S224: retain the individuals with high fitness in each group of corresponding individuals in step S22 to form a preferred population.

6. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 5 is characterized in that: In step S22, the solution process of the nearest neighbor search mechanism includes: S225: Compare the individual fitness of the optimal population generated in step S224, and select the individual with the highest fitness as the male jackal x m,k,n , select the individual with the second highest fitness as the female jackal x f,k,n ; S226: updating the positions of individuals in the population according to the prey escape energy, and forming the updated individuals into a prey population, including the following steps: Calculate the prey escape energy. The calculation formula of the prey escape energy E is as follows: Where E0 is the initial avoidance energy randomly generated by the traditional golden jackal algorithm, and E0∈(0,1); Update the positions of individuals in the population according to the calculated prey escape energy: Male Jackal m,k,n Used to generate individuals in the optimal population in step S224 Corresponding male jackal prey x mp,k,n ; Female jackal fp,k,n Used to generate individuals in the optimal population in step S224 Corresponding female jackal prey x fp,k,n ; According to the generated male jackal prey x mp,k,n and female jackal prey x fp,k,n Calculate individual The individual x corresponding to the prey population k+1,n ; The calculation process of the position of individuals in the prey escape energy update population is as follows: When E ≥ 1, the male jackal's prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing: When E<1, the male jackal's prey x mp,k,n , female jackal prey x fp,k,n and individuals x of the prey population k+1,n Update by pressing: In the formula, the Levy distribution random vector rl and the individual in the optimal population are vectors of the same order, is the individual of the optimal population generated in step S224 after the kth iteration; Collect updated individuals x k+1,n Form prey populations; S227: Use the nearest neighbor search mechanism to search for the individual x generated in step S226 k+1,n Randomly add or subtract one from any element in to get individual x k+1,n The nearest neighbor solution for an element in ; Among them, individual x k+1,n is an L×1-dimensional switch action vector, where L is the number of rings and each element corresponds to the number of the switch in each ring; S228: Loop through step S227 to obtain x k+1,n The nearest neighbor scheme for each element in ; S229: x obtained in step S226 is k+1,n and the distribution network switch strategies corresponding to the neighboring solutions obtained in step S228 are respectively applied to the distribution network to calculate the corresponding individual fitness; S2210: Compare x in step S229 k+1,n And the individual fitness of the nearest neighbor solution obtained, retaining each x k+1,n The distribution network switching strategy with the highest fitness among its neighboring solutions is used as the switching action vector corresponding to this strategy as the new x k+1,n ; The solution process of improving the traditional golden jackal algorithm in step S22 also includes: S2211: Collect the new x in step S2210 k+1,n Combine into new evolutionary populations; S2212: Determine the new x in step S2210 k+1,n Whether the corresponding number of iterations k reaches the upper limit, If the number of iterations does not reach the upper limit, the number of iterations k is increased by 1 and the new evolution population generated in step S2211 is used as the current evolution population to execute step S222; If the number of iterations reaches the upper limit, the new x in step S2210 is output. k+1,n The distribution network switching strategy with the highest corresponding fitness is used as the distribution network switching strategy solved in step S22.

7. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 1, characterized in that: In step S5, the calculation formula of the thermal stability margin increment is: λ=ΔI TR -ΔI DN ; In the formula, ΔI TR is the dynamic current carrying capacity change, ΔI DN is the maximum current change of the distribution network; The calculation formula for dynamic current carrying capacity change is: Where sgn() is a function that returns the sign of the variable, ΔV w is the wind speed change, indicating the current wind speed V w,t The initial wind speed V w,0 The difference between the dynamic current carrying capacity change rate Among them, the convection heat dissipation coefficient k c are preset values ​​related to wind angle, line diameter, air density, air thermal conductivity, air viscosity, etc. c,0 is the initial temperature of the fan conductor, T a is the ambient temperature, R(T c,0 ) is the initial temperature of the fan conductor at the fan conductor T c,0 Resistance value under, mC p is the conductor heat capacity of the fan, t represents the period from the initial moment to the current moment; The calculation formula for the maximum current change of the distribution network is: ΔI DN =k DN |ΔV w,cal | 0.5 ; In the formula, k DN is the rate of change of distribution network current, ΔV w,cal is the wind speed change; distribution network current change rate P r is the rated power of the fan, V r is the rated speed of the fan, V i is the fan cut-in speed, R + is the equivalent resistance of the fan port, R - is the equivalent resistance of the substation port, which is the power supply access port of the power distribution main grid; wind speed change min() is an operation that returns the minimum value, and max() is an operation that returns the maximum value.

8. The method for optimizing the dispatching strategy of a power grid containing wind power according to claim 1, characterized in that: In step S5, if the thermal stability check of the distribution network line fails, the process returns to step S3 to readjust the reactive output of the static VAR generator, the wind turbine abandoned power and the interruptible load active power, and obtain a new grid dispatching strategy containing wind power.

9. The method for optimizing the dispatching strategy of a power grid containing wind power according to any one of claims 4 to 8, characterized in that: The multi-time scale collaborative optimization model is rolled out and solved to obtain an optimized grid dispatch strategy containing wind power, including: In the previous total period, the distribution network switching strategy of each reconstruction period is obtained according to step S2, and the distribution network switching strategy of each reconstruction period obtained is correspondingly applied to the current total period; At the beginning of each dispatching period in each reconstruction period of the current total period, a grid dispatching strategy containing wind power is obtained according to step S3, and the obtained grid dispatching strategy containing wind power is applied to the current dispatching period; In the current total period, the verification period is used as the optimization period of the safety verification module, and the thermal stability margin increment is cyclically calculated according to step S5 and the thermal stability of the distribution network line is verified; In the current total period, the distribution network switching strategy of each reconstruction period is obtained according to step S2, and the distribution network switching strategy of each reconstruction period obtained is correspondingly applied to the next total period.

10. A device for optimizing the dispatching strategy of a power grid including wind power, characterized in that: It includes a model building module, a switch strategy module, a scheduling strategy module, a strategy execution module, a thermal stability module and a strategy output module; The model building module is used to build a multi-time scale collaborative optimization model, wherein the optimization target of the multi-time scale collaborative optimization model is to minimize the sum of the operating cost and the control cost, wherein the control cost includes: reconstruction cost, SVG adjustment cost, wind turbine abandonment cost and load interruption cost; The switch strategy module is used to collect the topology and operating condition information of the power grid containing wind power, and reconstruct and optimize the power grid topology containing wind power to obtain the switch strategy of the distribution network after reconstruction and optimization; The dispatching strategy module is used to adjust the reactive output of the static VAR generator, the abandoned wind power of the wind turbine and the active power of the interruptible load as a control means in the distribution network according to the distribution network switching strategy of the switching strategy module, so as to obtain an optimized grid dispatching strategy containing wind power; The strategy execution module is used to perform grid dispatching using the grid dispatching strategy containing wind power obtained in step S3; The thermal stability module is used to check the thermal stability of the distribution network line using the thermal stability margin increment index; The strategy output module is used to output the wind power-containing power grid dispatching strategy after the thermal stability check in the thermal stability module if the thermal stability check of the distribution network line is qualified.

Images