Multi-candidate-position SNOP planning method and system based on improved JAYA algorithm

Abstract

The invention discloses a multi-candidate-position SNOP planning method and system based on an improved JAYA algorithm, and the method comprises the steps: obtaining a plurality of historical data samples of the operation of a power distribution network, and carrying out the clustering of the plurality of historical data samples, and generating a typical daily source load scene; establishing a power distribution network operation risk assessment index, and constructing an SNOP configuration double-layer optimization model; and solving the SNOP configuration double-layer optimization model by adopting an improved JAYA algorithm based on a typical daily source load scene and a power distribution network operation risk assessment index to obtain an SNOP optimal configuration result. According to the method, the optimal planning configuration result of the SNOP can be output, so that the operation safety and economy of the power distribution network can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network planning, and in particular to a multi-candidate location SNOP planning method and system based on an improved JAYA algorithm. Background Art

[0002] As an advanced power electronic device, the flexible soft switch (SNOP) is used in active distribution systems to replace traditional open-circuit switches. The device can provide flexible control of active and reactive power and voltage regulation in the operating state, thereby enhancing the controllability of the distribution system and further improving the economy, flexibility and reliability of the power grid. Current research generally focuses on the economic issues of SNOP configuration, but rarely conducts comprehensive considerations from the perspective of the comprehensive operation risk of the distribution network. It is particularly important to quantify the impact of introducing SNOP on the operation risk of the distribution network and incorporate the role of SNOP in improving the operation safety of the distribution network into its optimization model. In-depth research in this area will not only help to comprehensively evaluate the potential value of SNOP, but also provide theoretical support for achieving safer and more reliable power system operation. Therefore, it is necessary to provide a SNOP planning scheme that takes into account the operation risk of the distribution network. Summary of the invention

[0003] The present invention aims to solve the technical problems in the related art at least to a certain extent. To this end, the first object of the present invention is to provide a multi-candidate location SNOP planning method based on an improved JAYA algorithm, which can output the SNOP optimal planning configuration result, thereby facilitating the improvement of the safety and economy of distribution network operation.

[0004] The second object of the present invention is to provide a multi-candidate position SNOP planning system based on the improved JAYA algorithm.

[0005] To achieve the above object, the present invention is implemented through the following technical solutions:

[0006] A multi-candidate position SNOP planning method based on an improved JAYA algorithm, comprising:

[0007] Obtain multiple historical data samples of distribution network operation, cluster the multiple historical data samples, and generate typical daily source-load scenarios;

[0008] Establish distribution network operation risk assessment indicators and construct a two-layer optimization model for SNOP configuration;

[0009] Based on the typical daily source-load scenario and distribution network operation risk assessment indicators, the improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model to obtain the SNOP optimal configuration result.

[0010] Preferably, clustering multiple historical data samples includes:

[0011] Determine characteristic indicators of multiple dimensions in each historical data sample;

[0012] Standardize the characteristic indicators of multiple dimensions in each historical data sample;

[0013] Calculate the weights of characteristic indicators of multiple dimensions in each historical data sample after standardization;

[0014] Based on the characteristic index values ​​of multiple dimensions in each historical data sample after standardization and their corresponding weights, the comprehensive index distance between each historical data sample is calculated, so as to achieve clustering of multiple historical data samples based on the comprehensive index distance.

[0015] Preferably, the distribution network operation risk assessment indicators include a voltage over-limit indicator, a line power over-limit indicator and a power factor over-limit risk indicator.

[0016] Preferably, the SNOP configuration double-layer optimization model includes an upper layer model and a lower layer model.

[0017] Preferably, the objective function of the upper model is a function that aims to minimize the sum of the annualized investment cost of SNOP and the annual comprehensive operating cost, wherein the annualized investment cost of SNOP includes the SNOP site selection cost, installation cost and residual value recovery income.

[0018] Preferably, the objective function of the lower model is a function with the goal of minimizing the annual comprehensive operating cost, wherein the annual comprehensive operating cost includes SNOP equipment operation and maintenance cost, loss cost, abandoned light cost and load shedding cost.

[0019] Preferably, the constraints in the SNOP configuration two-layer optimization model include SNOP operation constraints, photovoltaic output constraints, line flow constraints and distribution network operation risk assessment index constraints.

[0020] Preferably, based on the typical daily source-load scenario and the distribution network operation risk assessment index, an improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model, including:

[0021] Based on the typical daily source-load scenario and the distribution network operation risk assessment index, the improved JAYA algorithm is used to solve the lower model to obtain the minimum annual comprehensive operation cost, and the minimum annual comprehensive operation cost is substituted into the upper model;

[0022] The improved JAYA algorithm is used to solve the upper model again to obtain the optimal configuration result of SNOP, wherein the configuration result includes the location and capacity of SNOP.

[0023] Preferably, the improved JAYA algorithm is an algorithm that comprehensively improves the JAYA algorithm by executing perturbations, optimal individual variable crossover, and pinhole imaging reverse learning strategies in three dimensions.

[0024] To achieve the above object, the second aspect of the present invention provides a multi-candidate position SNOP planning system based on an improved JAYA algorithm, comprising:

[0025] A clustering module is used to obtain multiple historical data samples of the distribution network operation, and cluster the multiple historical data samples to generate a typical daily source-load scenario;

[0026] A construction module is used to establish distribution network operation risk assessment indicators and build a two-layer optimization model for SNOP configuration;

[0027] The solution module is used to solve the SNOP configuration double-layer optimization model based on the typical daily source-load scenario and the distribution network operation risk assessment index by using the improved JAYA algorithm to obtain the SNOP optimal configuration result.

[0028] The present invention has at least the following technical effects:

[0029] The present invention provides a multi-candidate location SNOP planning method and system based on an improved JAYA algorithm. The SNOP planning method specifically first obtains multiple historical data samples of the operation of the distribution network, then clusters the multiple historical data samples to generate a typical daily source-load scenario, then establishes a distribution network operation risk assessment index, and constructs a SNOP configuration double-layer optimization model. After that, based on the typical daily source-load scenario and the distribution network operation risk assessment index, the improved JAYA algorithm is used to solve the lower model in the double-layer optimization model to obtain the minimum annual comprehensive operating cost, and the minimum annual comprehensive operating cost is substituted into the upper model, and then the improved JAYA algorithm is used to solve the upper model, and the SNOP optimal configuration site selection and configuration capacity can be obtained. Therefore, the present invention can output the SNOP optimal planning configuration result through the above-mentioned SNOP planning strategy, so as to facilitate the improvement of the safety and economy of the distribution network operation.

[0030] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 The flowchart of the multi-candidate position SNOP planning method based on the improved JAYA algorithm according to an embodiment of the present invention.

[0032] Figure 2 This is a flow chart of a two-layer optimization model for solving SNOP configuration using the improved JAYA algorithm according to an embodiment of the present invention.

[0033] Figure 3 Schematic diagram of the simulation verification model.

[0034] Figure 4 is the maximum operation risk index r of different preset distribution network lines R,max Schematic diagram of the impact on the economics of SNOP planning.

[0035] Figure 5 Schematic diagram of the iterative process of solving the two-layer model for each algorithm.

[0036] Figure 6 4 is a structural block diagram of a multi-candidate position SNOP planning system based on an improved JAYA algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The present embodiment is described in detail below, and examples of the embodiment are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0038] The following describes the multi-candidate position SNOP planning method and system based on the improved JAYA algorithm of this embodiment with reference to the accompanying drawings.

[0039] Figure 1 FIG. 1 is a flow chart of a multi-candidate position SNOP planning method based on an improved JAYA algorithm according to an embodiment of the present invention. Figure 1 As shown, the multi-candidate position SNOP planning method based on the improved JAYA algorithm includes:

[0040] Step S1: Acquire multiple historical data samples of the distribution network operation, and cluster the multiple historical data samples to generate a typical daily source-load scenario.

[0041] In this embodiment, clustering of multiple historical data samples includes: determining characteristic indicators of multiple dimensions in each historical data sample; standardizing the characteristic indicators of multiple dimensions in each historical data sample; calculating the weights of the characteristic indicators of multiple dimensions in each historical data sample after standardization; and calculating the comprehensive indicator distance between each historical data sample based on the characteristic indicator values ​​of multiple dimensions in each historical data sample after standardization and their corresponding weights, so as to achieve clustering of multiple historical data samples based on the comprehensive indicator distance.

[0042] Due to the presence of multiple dimensions and multiple time scales in historical data samples, direct clustering methods may not be able to fully capture the inherent characteristics of these features, which may lead to complex calculations and potential misclassification. In order to solve this problem and improve the efficiency and quality of clustering, it is necessary to extract relevant feature indicators before clustering. In this study, a set of characteristic indicators with five dimensions that are conventionally used to describe the power fluctuation curve were selected. These indicators that describe the characteristics of electric load and photovoltaic timing output include daily load / power supply power v1, daily peak-to-valley difference rate v2, daily maximum load / power supply power time v3, maximum load / power supply power moment v4 and minimum load / power supply power moment v5. Among them, taking a type such as electric load or photovoltaic timing output, i.e., power supply power, as an example, the calculation of each characteristic indicator is as follows:

[0043]

[0044] v4=T max (4)

[0045] v5=T min (5)

[0046] Where P ave is the daily average power; P max With P min is the maximum and minimum power of the day; t is time, T is the operation cycle; P t is the power at time t; T max and T min and are the moments when the maximum and minimum power occur respectively.

[0047] In order to ensure that multiple characteristic indicators can be compared, formula (6) is used to standardize each characteristic indicator.

[0048]

[0049] In the formula, v m,n is the mth characteristic index value of the nth historical data sample, including two categories: electric load and photovoltaic output, so each historical data sample contains 10 characteristic indicators; v' m,n is the index v m,n The standardized value, v n Represents the nth historical data sample.

[0050] The weight of the characteristic index plays a vital role in accurately describing the curve and capturing the significance of different indicators. This paper uses the literature entropy weight method to calculate each characteristic index. The weight calculation method is shown in formulas (7)-(9).

[0051]

[0052] In the formula, pm,n is the proportion of the mth characteristic index of the nth historical data sample; N is the total number of historical data samples; E m is the information entropy of the mth characteristic index; W m is the weight of the mth feature index.

[0053] Based on the calculated characteristic index information and weight information, by calculating the distance between each standardized index value of two historical data samples, the comprehensive index distance between the two historical data samples can be obtained as shown in formula (10).

[0054]

[0055] Where n1 and n2 are two historical data samples; D n1,n2 is the comprehensive index distance between two historical data samples; W m,n1 is the weight of the mth characteristic index of the n1th historical data sample; v' m,n1 、v' m,n2 are the standardized values ​​of the corresponding indicators. The comprehensive indicator distance is used to characterize the similarity between two historical data samples. Based on the comprehensive indicator distance, the k-means (clustering algorithm) algorithm is used to cluster the historical data samples to generate a typical daily source-load scenario.

[0056] Step S2: Establish distribution network operation risk assessment indicators and construct a two-layer optimization model for SNOP configuration.

[0057] The distribution network operation risk assessment index is used to characterize the risk of the distribution network during normal operation. Common risk indicators include node voltage over-limit, line power over-limit, and power factor over-limit risk indicators. Each risk indicator is scaled according to the maximum acceptable risk of the distribution network, so that each risk indicator becomes a standardized result distributed in the interval [0,1] when the distribution network is operating safely, and the larger the indicator value, the greater the risk. Set the distribution network voltage over-limit threshold to 10%, the line power over-limit threshold to 85%, and the power factor over-limit risk threshold to 9%.

[0058] (1) Voltage over-limit indicator

[0059]

[0060] In the formula, r V is the voltage over-limit indicator; π s is the probability of scene s occurring; S is the number of uncertain scenes; j represents the jth node; j n is the total number of nodes; is the voltage of node j at time t in scenario s; and V are the upper and lower limits of the node voltage.

[0061] (2) Line power over-limit indicator

[0062]

[0063] In the formula, r P is the line power limit indicator; L represents the Lth line, L n Indicates the total number of lines; P ij,max is the upper limit of active power of line ij; is the active power of line ij.

[0064] (3) Power factor over-limit risk indicator

[0065]

[0066] In the formula, r Z It is the power factor over-limit risk indicator; is the system power factor deviation; It is the maximum system power factor deviation that can be accepted during normal system operation.

[0067] After standardizing the three risk indicators, since the three risk indicators are equally important, the comprehensive operation risk indicator of the planned distribution network can be expressed as shown in formula (14).

[0068]

[0069] In the formula, r R It is the comprehensive operation risk indicator of the distribution network under planning.

[0070] In one embodiment of the present invention, the SNOP configuration two-layer optimization model includes an upper model and a lower model. The objective function of the upper model is a function that minimizes the sum of the annualized investment cost of SNOP and the annual comprehensive operating cost, wherein the annualized investment cost of SNOP includes the SNOP site selection cost, installation cost, and residual value recovery income. The objective function of the lower model is a function that minimizes the annual comprehensive operating cost, wherein the annual comprehensive operating cost includes the SNOP equipment operation and maintenance cost, loss cost, abandoned light cost, and load shedding cost.

[0071] Specifically, the SNOP configuration two-layer optimization model consists of two layers, namely the planning layer and the operation layer. The upper model optimizes the equipment site selection and capacity configuration with the goal of minimizing the annualized investment cost of SNOP and the annual comprehensive operating cost; the lower model optimizes the operation scheduling strategy based on the typical daily source-load scenario and the set maximum allowable operating risk index of the distribution network with the goal of minimizing the annual comprehensive operating cost. The interactive relationship between the two-layer models is that the lower model optimizes the typical daily scheduling strategy of the distribution network based on the SNOP planning results, and feeds the annual comprehensive operating cost back to the upper model. The upper model solves the SNOP site selection and configuration capacity, and then passes it to the lower model for cyclic solution to obtain the final optimal SNOP configuration result.

[0072] The objective function of the upper model is to minimize the sum of the annualized investment cost of SNOP and the annual comprehensive operating cost, as shown in formula (15).

[0073] minf up =C INV +C OPE (15)

[0074] In the formula, f up is the objective function of the upper model; C INV is the annualized investment cost of SNOP; C OPE is the annual comprehensive operating cost of the distribution network. Among them, the annualized investment cost of SNOP includes the SNOP site selection cost, installation cost and residual value recovery income, as shown in formulas (16) to (20).

[0075] C INV =γ(C site +C cap -C rec ) (16)

[0076] In the formula, γ is the return on investment, and the calculation formula is:

[0077]

[0078] In the formula, C site , C cap With C rec are SNOP site selection cost, installation cost and residual value recovery income respectively; T N is the life cycle; a is the discount rate; c is the site is the location cost at the location h; h is a Boolean variable indicating whether a SNOP device is installed at the location h. s is the total number of candidate positions; c cap K is the installation cost per unit capacity of SNOP; s The number of SNOP installations; For the kth s The installed capacity of SNOPs; c rec is the recycling factor of SNOP equipment; For the kth s The remaining recycling value of SNOP.

[0079] Among them, the annual comprehensive operating cost in the upper model is calculated by the lower model.

[0080] The objective function of the lower model is shown in formula (21), which includes SNOP equipment operation and maintenance cost, loss cost, abandoned light cost and load shedding cost.

[0081] minf down =C ma +C loss +C PV,ab +C L,cut (twenty one)

[0082] In the formula, f down is the objective function of the lower model; C ma , C loss , C PV,ab , C L,cut are the SNOP equipment operation and maintenance cost, loss cost, abandoned light cost and load shedding cost, respectively, and are calculated by equations (22) to (25).

[0083]

[0084] Where, d s is the number of days with typical day scene s in a year; are the interaction powers of nodes i1 and i2 at time t in scenario s; are the SNOP converter loss coefficients corresponding to nodes i1 and i2 respectively. It can be approximately considered that the SNOP converter loss is linearly related to the transmission power; and are the power of abandoned light and load shedding at time t under scenario s; c ma 、c loss 、c PV,ab 、c L,cut They are the unit operation and maintenance cost, loss cost, abandoned light cost and load shedding cost of SNOP equipment.

[0085] Furthermore, the constraints in the SNOP configuration two-layer optimization model include SNOP operation constraints, PV output constraints, line flow constraints, and distribution network operation risk assessment index constraints.

[0086] (1) SNOP operation constraints

[0087] Under normal operation of the distribution network, SNOP can control the active power flow between the connected feeders and supply or absorb reactive power at its interface terminals. In this paper, back-to-back voltage source converters (B2BVSCs) are applied to the optimization model of SNOP site selection and scale under steady-state conditions, and the SNOP control mode selects the PQ-VdcQ (active-reactive) control mode. When the loss coefficient of high power transmission between two converters is considered, the constraint of active power transmission between two nodes can be described by equation (26).

[0088]

[0089] The SNOP operation of nodes i1 and i2 satisfies the capacity constraint and can be expressed as shown in equations (27) and (28).

[0090]

[0091] In the formula, are the interactive reactive power of nodes i1 and i2 at time t in scenario s, Install SNOP capacity between nodes i1 and i2.

[0092] (2) Photovoltaic output constraints

[0093] Distributed photovoltaic inverters can provide a certain degree of reactive power support. The operating constraints of distributed photovoltaics are:

[0094]

[0095] In the formula, and are the active power and reactive power of the photovoltaic power plant at node j at time t in scenario s, respectively; is the maximum active power output of the photovoltaic power plant at node j; and are the upper and lower limits of the PV reactive power of node j, respectively; is the power factor angle of the photovoltaic power plant at node j at time t in scenario s.

[0096] (3) Line flow constraints based on preset operation risk indicators

[0097] Using Boolean variables To represent the switching state of line ji, the line power flow constraint can be expressed as shown in equations (32) to (37).

[0098]

[0099] Equations (32) and (33) are node power balance constraints. and are respectively the active power and reactive power of line ji; r ji With x ji is the resistance and reactance of line ji; is the current of line ji at time t. A Boolean variable representing the switch status of line ik; and is the active power and reactive power of line ji. Equation (34) is the Ohm's law constraint for the voltage drop between two nodes between lines. is the voltage of nodes i and j at time t. The line current can be obtained by the constraint of equation (35). Equations (36) and (37) represent the active and reactive power injection balance of node i, that is, power injection minus load consumption. Among them, are the active power and reactive power of the photovoltaic power of node i at time t in scenario s respectively; are the interactive power and interactive reactive power of node i at time t in scenario s, respectively; are the active and reactive power of the load of node i at time t in scenario s, respectively; are the active and reactive injection balance powers of node i at time t in scenario s, respectively.

[0100] Line flow constraints also include operation risk constraints. This embodiment is based on the method of presetting risk indicators, and sets the maximum line operation risk indicator r in the planning stage. R,max , then the line voltage over-limit risk, current over-limit risk and power factor over-limit risk constraints can be expressed as shown in formula (38).

[0101]

[0102] Among them, r Z System power factor deviation in the indicator It is mainly determined by the load and the photovoltaic inverter power factor, as shown in formula (39).

[0103]

[0104] In the formula, is the power factor angle, is the maximum power factor angle; is the apparent power of the photovoltaic power of node j at time t in scenario s; is the apparent power of the load of node j at time t in scenario s.

[0105] Step S3: Based on the typical daily source-load scenario and the distribution network operation risk assessment indicators, the improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model to obtain the SNOP optimal configuration result.

[0106] In this embodiment, the improved JAYA algorithm is an algorithm that comprehensively improves the JAYA algorithm by executing disturbance, optimal individual variable crossover and pinhole imaging reverse learning strategy.

[0107] The planning problem of SNOP site selection and capacity determination belongs to a large-scale mixed integer nonlinear programming problem, which is difficult to solve directly. In view of the SNOP configuration double-layer optimization model proposed in this embodiment, an improved JAYA algorithm is proposed.

[0108] The JAYA algorithm is a population-based optimization algorithm. Its main feature is that it has no additional parameters, which makes it different from traditional evolutionary algorithms. For evolutionary algorithms, parameter adjustment is a tedious task. During the evolutionary process, the JAYA algorithm optimizes the optimization process by approaching the optimal solution and the worst solution.

[0109] f(x) is the objective function, x ita is the decision variable value of the it-th iteration, i.e., the individual value. and The optimal objective function value and the corresponding optimal decision variable value found for the current iteration ita. For the minimization optimization problem, is smaller than all other candidate solutions. On the contrary, and is the worst objective function value and the corresponding worst decision variable value found in the current iteration ita, so, is greater than all other candidate solutions.

[0110] As with other heuristic algorithms, the first step is to initialize the population, as shown in Equation (40).

[0111] x 0,e =B l +r0(B u -B l ) e=1,2,3...,n p (40)

[0112] In the formula, e is the population size, n is p is the total number of populations; r0 is a random number; B u , B l is the initialization population coefficient; x 0,e is the individual value of the 0th iteration in population e.

[0113] Calculate the objective function value of individuals in the population, and then use the JAYA algorithm to evolve the population as shown in formula (41).

[0114]

[0115] In the formula, x'ita,e is the individual value of the population after the it-th iteration in the population e; x ita,e is the individual value after the ita-th iteration in population e; r1 and r2 are two random numbers uniformly distributed between [0,1]. If the objective function f(x' ita,e ) is better than f(x ita,e ), then x' ita,e Replace x ita,e , otherwise, maintain the original population.

[0116] Until the ita+1th iteration fitness And the fitness of the it-th iteration The value difference is less than the error threshold ε th .

[0117] For formula (41), this cluster update method has relatively poor population diversity, which makes the algorithm easily fall into local optimality. For meta-heuristic algorithms, it is necessary to promote and achieve a balance between exploration and development during the population evolution process. Since the search direction is controlled by the best and worst solutions, the exploration and development of the JAYA algorithm are relatively poor. There are various improvements, such as the introduction of chaos methods, adaptive methods, and algorithms based on elite methods. However, the current improvement methods are all single strategies, and it is more effective to improve the heuristic algorithm from multiple angles. Therefore, this embodiment adopts three strategies to improve the JAYA algorithm, namely, execution perturbation, optimal individual variable crossover, and pinhole imaging reverse learning strategy.

[0118] (1) Execution disturbance

[0119] The JAYA algorithm is similar to other heuristic algorithms in that the solution process is affected by the initial population, which makes the algorithm easily approach a suboptimal solution in the middle and late stages. To prevent the algorithm from falling into a local optimum, this embodiment uses an execution perturbation method to improve the JAYA algorithm.

[0120] In the later iterations of the JAYA algorithm, the population is randomly generated. If the perturbed fitness value is better than the current fitness, it is updated and replaced. By introducing a probabilistic perturbation strategy in the mid- and late-stage iterations, the local convergence problem can be effectively reduced, as shown in formula (42).

[0121]

[0122] In the formula, ita max is the maximum number of iterations; x ita+1,e is the individual value after the ita+1th iteration in population e; is the individual value after the ita+1th iteration in the random population ex that performs the perturbation; f(x ita+1,e ), are the objective function values ​​corresponding to the individual values ​​respectively.

[0123] (2) Optimal individual variable crossover

[0124] When the JAYA algorithm is used to solve practical engineering problems, the positions of individuals in the population are often multidimensional. For example, the JAYA algorithm is used to solve the SNOP site selection and volume setting model in this article, which contains two-dimensional variables, namely site selection and volume setting variables. When the JAYA algorithm solves such problems, it is affected by the mutual influence between the two-dimensional variables (not linear). The population update strategy of formula (41) will make the richness of its population individuals very small, and the rapid global search capability will be weakened. For this reason, this embodiment adopts an optimal individual variable crossover method to cooperate with the population update process of formula (41). After executing the population update of formula (41), the best individual in the population is selected, the vector position in the individual is randomly exchanged, and compared with the fitness of the worst solution. If the fitness is better, the worst solution is replaced to generate a new population. This method can make the variables of individuals in the population have multiple dimensional richness without changing the population structure on a large scale, which is conducive to rapid global search, as shown in formula (43).

[0125]

[0126] In the formula, The new individual is obtained by processing the optimal individual through the variable crossover method; for The corresponding objective function value.

[0127] (3) Pinhole imaging reverse learning strategy

[0128] The above strategy of using the best individual variable crossover to replace the worst solution can improve the richness of multi-dimensional variables of the population. In addition, in view of the problem that the JAYA algorithm is prone to fall into local optimality when solving multi-dimensional complex planning problems, this embodiment introduces a pinhole imaging reverse learning strategy based on optical laws.

[0129] Pinhole imaging reverse learning is a general reverse learning strategy. The adjustment factor is adjusted by changing the distance between the wall and the plate with pinholes, thereby changing the optimal solution to obtain a better individual position, which can be expressed by formula (44).

[0130]

[0131] Where μ is the adjustment factor. A new candidate solution is generated by implementing the pinhole imaging reverse learning method for the position of the global optimal solution. Subsequently, the fitness of this new solution is evaluated according to the objective function. If the new solution performs better than the existing solution, it is included as a new preferred solution; otherwise, the original solution is retained. This method helps to enhance the overall performance of the algorithm and promote the efficiency of the optimization process.

[0132] In this embodiment, based on typical daily source-load scenarios and distribution network operation risk assessment indicators, an improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model, including: based on typical daily source-load scenarios and distribution network operation risk assessment indicators, an improved JAYA algorithm is used to solve the lower model to obtain the minimum annual comprehensive operating cost, and the minimum annual comprehensive operating cost is substituted into the upper model; the improved JAYA algorithm is used to solve the upper model again to obtain the SNOP optimal configuration result, wherein the configuration result includes the location and capacity of the SNOP. The process of solving the SNOP configuration two-layer optimization model with the improved JAYA algorithm is as follows: Figure 2 shown.

[0133] In this embodiment, the improved JAYA algorithm for solving the SNOP configuration double-layer optimization model has a faster iterative convergence speed and better stability than the standard JAYA algorithm.

[0134] This embodiment uses an improved 33-node example combined with the above model for simulation verification. Figure 3 As shown in the figure, the voltage level is 12.66KV, and distributed photovoltaic power sources are added at nodes 7, 11, 15, 22, 25, 28, and 32, including 37 branches, five of which are branches with tie switches. By presetting the maximum allowable distribution network operation risk index, SNOP equipment is configured to replace traditional tie line switches to improve the economic cost of distribution network operation. The photovoltaic risk coefficient and load shedding risk coefficient are set to 0.2 and 0.42 yuan / (kW.h) respectively.

[0135] (1) Analysis of SNOP planning results

[0136] Distribution network operation risk assessment index R Normal operation is between [0,1]. To ensure the rationality of SNOP planning, the maximum operation risk index r of the line in the planning stage is set. R,max =0.9, which is used to leave a certain risk margin for distribution network operation.

[0137] (2) SNOP configuration results

[0138] This embodiment can effectively solve the SNOP planning model for multiple candidate locations based on the solution algorithm of the improved JAYA. From the economic perspective, the number of SNOP devices installed is closely related to the site selection cost. The site selection cost is an important factor affecting the number, location and capacity of SNOP installations. When the site selection cost is small, SNOP can be installed in multiple locations to achieve flexible interconnection of multiple lines; and as the site selection cost increases, SNOP is only installed in some locations or even in a single location to achieve the optimal economic cost of overall planning and operation. Therefore, the site selection cost is an important consideration for the SNOP planning problem in multiple candidate locations.

[0139] (3) Operational risk indicators and SNOP investment economics analysis

[0140] Figure 4 The maximum operation risk index r of different preset distribution network lines is compared. R,max Impact on the economics of SNOP planning. Figure 4 It can be seen that as r R,max With the increase of SNOP, the investment cost of SNOP and the operation cost of distribution network will increase. The distribution network can reduce the operation risk of distribution network and improve the operation safety of distribution network by increasing the configuration capacity of SNOP.

[0141] (4) Improved JAYA algorithm verification

[0142] In order to analyze the effectiveness of the improved JAYA algorithm in solving the SNOP configuration two-level optimization model, the standard JAYA algorithm and the genetic algorithm were used to solve the two-level model for comparison. Figure 5 As shown. It can be seen that compared with the other two algorithms, the improved JAYA algorithm has faster optimization ability in the early stage of iteration. If the iteration convergence threshold ξ ita ≤0.1, the JAYA algorithm and the genetic algorithm converged after 85 and 62 iterations respectively, in contrast, the improved JAYA algorithm only took 43 iterations to achieve convergence.

[0143] In order to analyze the accuracy and stability of the algorithm, the iterative calculation of each algorithm was repeated 20 times to obtain a large number of objective function values. The statistical results of the maximum value, minimum value, average value and standard deviation are shown in Table 1. It can be seen that the average values ​​of the objective functions obtained by the three algorithms are not much different, and the improved JAYA algorithm has a slightly optimal average value, which proves the accuracy of the improved JAYA algorithm. In addition, from the standard deviation results in Table 1, the improved JAYA algorithm has a significant reduction compared with the two algorithms, which means better solution stability. The main reason for this improvement is that the improved JAYA algorithm improves the algorithm performance by executing interference, optimal individual variable crossover and pinhole imaging reverse learning strategies in each iteration.

[0144] Table 1 Comparison of fitness values ​​of various algorithms

[0145]

[0146] Figure 6 FIG. 1 is a structural block diagram of a multi-candidate position SNOP planning system based on an improved JAYA algorithm according to an embodiment of the present invention. Figure 6 As shown, the multi-candidate position SNOP planning system 100 based on the improved JAYA algorithm includes a clustering module 10, a construction module 20 and a solution module 30 connected to the clustering module 10 and the construction module 20 respectively.

[0147] Among them, the clustering module 10 is used to obtain multiple historical data samples of the distribution network operation, and cluster the multiple historical data samples to generate a typical daily source-load scenario; the construction module 20 is used to establish the distribution network operation risk assessment index, and construct a SNOP configuration two-layer optimization model; the solution module 30 is used to solve the SNOP configuration two-layer optimization model based on the typical daily source-load scenario and the distribution network operation risk assessment index, and use the improved JAYA algorithm to obtain the SNOP optimal configuration result.

[0148] It should be noted that the specific implementation of the multi-candidate position SNOP planning system based on the improved JAYA algorithm in the embodiment of the present invention can refer to the specific implementation of the multi-candidate position SNOP planning method based on the improved JAYA algorithm mentioned above. To avoid redundancy, it will not be repeated here.

[0149] In summary, the present invention provides a multi-candidate location SNOP planning method and system based on the improved JAYA algorithm. The SNOP planning method specifically first obtains multiple historical data samples of the distribution network operation, then clusters the multiple historical data samples to generate a typical daily source-load scenario, then establishes a distribution network operation risk assessment index, and constructs a SNOP configuration double-layer optimization model. After that, based on the typical daily source-load scenario and the distribution network operation risk assessment index, the improved JAYA algorithm is used to solve the lower model in the double-layer optimization model to obtain the minimum annual comprehensive operating cost, and the minimum annual comprehensive operating cost is substituted into the upper model, and then the improved JAYA algorithm is used to solve the upper model, and the SNOP optimal configuration site selection and configuration capacity can be obtained. Therefore, the present invention can output the SNOP optimal planning configuration result through the above-mentioned SNOP planning strategy, so as to facilitate the improvement of the safety and economy of the distribution network operation.

[0150] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0151] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A multi-candidate position SNOP planning method based on an improved JAYA algorithm, characterized in that: include: Obtain multiple historical data samples of distribution network operation, cluster the multiple historical data samples, and generate typical daily source-load scenarios; Establish distribution network operation risk assessment indicators and construct a two-layer optimization model for SNOP configuration; Based on the typical daily source-load scenario and distribution network operation risk assessment indicators, the improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model to obtain the SNOP optimal configuration result.

2. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 1, characterized in that: Clustering of multiple historical data samples, including: Determine characteristic indicators of multiple dimensions in each historical data sample; Standardize the characteristic indicators of multiple dimensions in each historical data sample; Calculate the weights of characteristic indicators of multiple dimensions in each historical data sample after standardization; Based on the characteristic index values ​​of multiple dimensions in each historical data sample after standardization and their corresponding weights, the comprehensive index distance between each historical data sample is calculated, so as to achieve clustering of multiple historical data samples based on the comprehensive index distance.

3. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 1, characterized in that: The distribution network operation risk assessment indicators include voltage over-limit indicators, line power over-limit indicators and power factor over-limit risk indicators.

4. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 1, characterized in that: The SNOP configuration double-layer optimization model includes an upper layer model and a lower layer model.

5. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 4, characterized in that: The objective function of the upper model is a function that aims to minimize the sum of the annualized investment cost of SNOP and the annual comprehensive operating cost, wherein the annualized investment cost of SNOP includes the SNOP site selection cost, installation cost and residual value recovery income.

6. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 5, characterized in that: The objective function of the lower model is a function that aims to minimize the annual comprehensive operating cost, wherein the annual comprehensive operating cost includes SNOP equipment operation and maintenance cost, loss cost, abandoned light cost and load shedding cost.

7. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 6, characterized in that: The constraints in the SNOP configuration two-layer optimization model include SNOP operation constraints, photovoltaic output constraints, line flow constraints and distribution network operation risk assessment index constraints.

8. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 7, characterized in that: Based on the typical daily source-load scenario and distribution network operation risk assessment indicators, the improved JAYA algorithm is used to solve the SNOP configuration two-layer optimization model, including: Based on the typical daily source-load scenario and the distribution network operation risk assessment index, the improved JAYA algorithm is used to solve the lower model to obtain the minimum annual comprehensive operation cost, and the minimum annual comprehensive operation cost is substituted into the upper model; The improved JAYA algorithm is used to solve the upper model again to obtain the optimal configuration result of SNOP, wherein the configuration result includes the location and capacity of SNOP.

9. The multi-candidate position SNOP planning method based on the improved JAYA algorithm as claimed in claim 1, characterized in that: The improved JAYA algorithm is an algorithm that comprehensively improves the JAYA algorithm by executing disturbances, optimal individual variable crossover, and pinhole imaging reverse learning strategy.

10. A multi-candidate position SNOP planning system based on the improved JAYA algorithm, characterized in that: include: A clustering module is used to obtain multiple historical data samples of the distribution network operation, and cluster the multiple historical data samples to generate a typical daily source-load scenario; A construction module is used to establish distribution network operation risk assessment indicators and build a two-layer optimization model for SNOP configuration; The solution module is used to solve the SNOP configuration double-layer optimization model based on the typical daily source-load scenario and the distribution network operation risk assessment index by using the improved JAYA algorithm to obtain the SNOP optimal configuration result.

Images