Mobile energy storage planning method considering distribution network reliability and flexibility under extreme weather conditions
By simulating the spatiotemporal characteristics of extreme weather and building an evaluation index system, the mobile energy storage configuration of the distribution network is optimized, and the problem of improving the reliability and elasticity of the distribution network in extreme weather is solved, and more efficient calculation and optimization results are achieved.
Patent Information
- Application Number
- CN202210647326.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-09
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2042-06-09
AI Technical Summary
The existing technology has fewer solutions for improving the reliability and elasticity of the distribution network in extreme weather, especially in multiple faults, and lacks scientific evaluation indicators and optimized configuration methods, resulting in the safe operation of the distribution network in extreme situations. threaten.
By simulating the spatiotemporal characteristics of extreme weather, a failure rate model and evaluation index system were established, and the extreme weather scenarios were reduced by using Monte Carlo method and K-means clustering algorithm, and combining Latin hypercube sampling and COBL-PSO algorithm to construct the entire disaster period The mobile energy storage optimization planning model is used to optimize the mobile energy storage configuration with the pre-disaster loss and post-disaster recovery rate as the goal.
It improves the computing efficiency and scientific planning of the distribution network in extreme weather, enhances the reliability and elastic recovery capabilities of power supply, solves the limitations of single period optimization, and achieves more accurate evaluation and optimization results.
Smart Images

Figure CN115062833B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of distribution network energy storage system planning, and in particular to a mobile energy storage planning method that takes into account the reliability and elasticity of the distribution network under extreme weather conditions. Background Art
[0002] Affected by factors such as global climate change, the frequency of extreme weather events is increasing year by year, and the safe and stable operation of distribution networks is further threatened and challenged. Flexible dispatch resources are urgently needed to enhance the reliability of distribution network power supply. Academic research and engineering applications are paying more and more attention to the impact of low-probability and high-loss events on power systems, and the concept of "resilience" is introduced to evaluate the power supply recovery capability of power grids under extreme events, especially the support and recovery capability of key loads in distribution networks. At present, there are few solutions to improve the reliability and resilience of distribution networks based on the uncertainty of the spatiotemporal impact of natural disasters such as typhoons on distribution networks. They mainly focus on emergency repairs and reinforcement, and lack research on the utilization of flexible resources in distribution networks.
[0003] As a new type of energy storage resource, mobile energy storage has the characteristics of spatiotemporal flexibility and rapid response. It can effectively improve the intermittent fluctuation characteristics of distributed power sources by cooperating with the power supply system. Therefore, it is of great significance to study the optimal configuration and operation of mobile energy storage to improve the reliability and flexibility of the distribution network, which is of great significance to the safe operation of the power grid. Most of the existing studies are based on subjective prediction of faults or single-period optimization. The operation planning of mobile energy storage has limitations, and in extreme cases, multiple faults often occur in the distribution network. Considering only the N-1 safety criterion of the distribution network is not enough to ensure the safe operation of the distribution network. It is urgent to combine the topological structure and operation status of the distribution network and propose scientific evaluation indicators to evaluate the rationality of the optimal configuration scheme of mobile energy storage. Summary of the invention
[0004] In order to overcome the shortcomings of the above-mentioned existing methods, the present invention proposes a mobile energy storage planning method that considers the reliability and flexibility of the distribution network under extreme weather conditions. It effectively quantifies the impact of faults by simulating the spatiotemporal characteristics of extreme weather, uses an improved algorithm to determine the optimal planning scheme for mobile energy storage, thereby improving calculation efficiency, and can use a reliability and flexibility evaluation index system to evaluate the scientific nature of the planning scheme.
[0005] To achieve the above object, the technical solution adopted by the present invention is: a mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions, which includes:
[0006] Step 1: Establish a distribution component failure rate model to model the uncertainty scenario of failure disasters in the distribution network;
[0007] Step 2: The extreme weather scenarios are extracted by Monte Carlo method and reduced by K-means clustering algorithm. Considering the uncertainty of load and distributed power output, Latin hypercube sampling is used to extract load, distributed power output and distribution network fault scenarios;
[0008] Step 3: Select evaluation indicators from multiple perspectives, calculate the objective weight of each indicator, and construct a distribution network reliability and flexibility evaluation indicator system;
[0009] Step 4: Taking the pre-disaster loss and post-disaster recovery rate as optimization targets, a mobile energy storage optimization planning model considering the entire disaster period is established;
[0010] Step 5: Solve the mobile energy storage optimization planning model.
[0011] The present invention firstly selects typhoon disasters as the research object, simulates the randomness of typhoon weather through the Monte Carlo method, establishes a spatiotemporal fault model to quantify the impact of extreme weather on the distribution network, and uses the K-means algorithm to reduce extreme weather scenarios. Considering the uncertainty of load and distributed power output, Latin hypercube sampling is used to extract load, distributed power output and distribution network fault scenarios; secondly, from the two dimensions of power supply reliability and elastic resilience, evaluation indicators are selected from multiple angles, and the objective weights of each indicator are calculated to construct a distribution network reliability and elasticity evaluation index system for evaluating the effectiveness of the mobile energy storage optimization planning scheme; then, taking the minimum load shedding amount and the maximum key load recovery rate as the objective function, a mobile energy storage planning model considering all disaster periods is established; finally, the COBL-PSO algorithm is used to solve the mobile energy storage optimization planning model, and the optimal operation of mobile energy storage and load switching output is achieved, so as to enhance the typhoon disaster absorption and resistance capacity of the distribution network.
[0012] Furthermore, the step 1 comprises:
[0013] Step 1.1: Select extreme weather represented by typhoons and establish a wind speed and direction distribution model
[0014] The Weibull distribution is used to describe the uncertainty of wind speed at a single moment:
[0015] F(v M ,θ)=p(θ)f(v M ) (1)
[0016] In formula (1), p(θ) is the wind direction frequency function, θ is the angle between the wind direction and the line, and f(v M ) is the Weibull distribution, v M is the maximum wind speed near the center of the typhoon;
[0017] Step 1.2: Establish a typhoon wind field model
[0018] Using the Jelesnianski circularly symmetric typhoon wind field model:
[0019]
[0020] In formula (2), v r is the wind speed at a distance r from the center, the wind direction is the counterclockwise tangent direction of the simulated circle, R is the maximum wind speed radius, which obeys the log-normal distribution, and r is the distance rate from the node to the center of the radius;
[0021] Step 1.3: Considering the influence of wind direction on component failure rate, establish a component failure rate model including wind speed and wind direction
[0022] The failure rate of the line when the wind speed is v is p1(v):
[0023]
[0024] In formula (3), λ θ,1 is the influence coefficient of wind direction on line fault, v crit,1 and v abs,1 is the critical wind speed of the line, η norm,1 is the normal operation failure probability of the line, μ1 is the sensitivity coefficient of the line to wind speed;
[0025] When the wind speed is v, the failure rate of the pole is p p (v):
[0026]
[0027] In formula (4), λ θ,p is the influence coefficient of wind direction on tower failure, v crit,p and v abs,p is the critical wind speed of the pole, η norm,p is the normal operation failure probability of the line, μ p is the sensitivity coefficient of the pole to wind speed;
[0028] Step 1.4: Equivalent the component failure rate model to a series model of the line failure rate model and the pole failure rate model
[0029]
[0030] In formula (5), p h (v) is the failure rate of the hth branch, m h and h are the number of lines and poles of the h-th branch respectively.
[0031] Furthermore, the step 2 comprises:
[0032] Step 2.1: Sampling of extreme weather scenarios
[0033] K extreme weather scenarios are extracted by Monte Carlo method, and the duration T of extreme weather is described by normal distribution. d :
[0034] T d ~N(μ,σ 2 ) (6)
[0035] In formula (6), μ and σ are the expectation and standard deviation of extreme weather duration, respectively;
[0036] Step 2.2: Scene reduction
[0037] Step 2.2.1, determine K extreme weather scenarios as the number of clusters, and randomly extract k initial cluster centers;
[0038] Step 2.2.2: Use K-means to iteratively cluster until the clustering converges and obtain k extreme weather scenarios:
[0039] Cluster centers at the cth iteration for:
[0040]
[0041] In formula (7), is the kth value of the cth iteration c Cluster centers, h e For the cth iteration The e-th node in For the cth iteration The number of nodes in ;
[0042] The sum of the distances from each node to its corresponding cluster center in the cth iteration is f c :
[0043]
[0044] In formula (8), is the distance from the e-th node to its corresponding cluster center in the c-th iteration;
[0045] Step 2.3: Create a fault scenario
[0046] Latin hypercube sampling is used to extract load, distributed generation output and distribution network fault scenarios;
[0047] Normal distribution is used to describe the load and distributed generation output:
[0048]
[0049] In formula (9), P tDG and P t L are the actual values of the distributed generation output and load at time t, and are the predicted values of distributed generation output and load at time t respectively.
[0050] Furthermore, the step 3 comprises:
[0051] Step 3.1: Select power supply reliability indicators
[0052] The reliability of power supply is characterized from three aspects: time, power shortage and reliability rate;
[0053] Step 3.2: Select elastic resilience index
[0054] The elastic recovery force is characterized from three aspects: time, speed and recovery rate;
[0055] Step 3.3: Calculate the objective weight of the indicator
[0056] The actual data of indicators in different evaluation scenarios are used to calculate the indicator entropy value and the objective weight of each indicator.
[0057] Furthermore, the step 3.1 includes:
[0058] Step 3.1.1: Establish the fault average power outage time indicator T SAIDI
[0059]
[0060] In formula (10), t i is the time of the ith fault outage in the distribution network, and N is the number of fault outages;
[0061] Step 3.1.2: Establish the average power outage and power shortage index E AENS
[0062]
[0063] In formula (11), D i is the power shortage caused by the ith power outage;
[0064] Step 3.1.2: Establish the average power supply reliability index R ASAI
[0065]
[0066] In formula (12), t s The statistical period.
[0067] Furthermore, the step 3.2 includes:
[0068] Step 3.2.1. Establish load recovery time index T RCL
[0069]
[0070] In formula (13), is the scheduling time for the mobile energy storage to move to the i-th node, is the fault repair time on the i-th node;
[0071] Step 3.2.2: Establish load recovery speed index V RSL
[0072]
[0073] In formula (14), T r is the fault recovery time, n1 is the total number of system load nodes, P r,i,j is the load power restored by the i-th node in the j-th time period;
[0074] Step 3.2.3: Establish the power shortage rate index R ESR
[0075]
[0076] In formula (15), ω i is the load weight value, P i L is the total load of the ith node, P ln,i,t is the load of the normal operation of the i-th node at time t, and Δt represents the length of the time period.
[0077] Furthermore, the step 3.3 includes:
[0078] Step 3.3.1: Construct the evaluation index matrix:
[0079] Assuming there are a scenarios to be evaluated and b evaluation indicators, the evaluation indicator matrix Z is quantitatively constructed:
[0080]
[0081] In formula (16), the element z ij It represents the value of the jth indicator in the i-th evaluation scenario.
[0082] Step 3.3.2, calculate the index proportion:
[0083] For the jth indicator, calculate the proportion p of the i-th evaluation scenario under this indicator ij ;
[0084]
[0085] Step 3.3.3, calculate the index entropy value:
[0086] Calculate the entropy value e of the jth indicator according to the entropy formula j :
[0087]
[0088] Step 3.3.4, calculate the index entropy weight:
[0089] Calculate the entropy weight ω of the jth indicator based on the indicator entropy value j :
[0090]
[0091] Furthermore, the step 4 comprises:
[0092] Step 4.1: Establish the objective function
[0093] Establish a mobile energy storage optimization planning scheme that takes into account all disaster periods, with the pre-disaster loss and post-disaster recovery rate as the optimization targets;
[0094] Step 4.1.1: Minimizing load shedding is the objective function
[0095]
[0096] In formula (20), C cut is the load shedding amount, a i,t is a 0-1 variable, indicating whether the load at the distribution network node i is removed at time t. i,t =0 means the load at the distribution network node i is removed at time t, a i,t =1 means that the load at the distribution network node i is powered normally at time t. is the load of distribution network node i at time t, T d To use normal distribution to describe the duration of extreme weather, n1 is the total number of system load nodes, ω i is the load weight value;
[0097] Step 4.1.2: Take the maximum critical load recovery rate as the objective function:
[0098]
[0099] In formula (21), H key is the critical load recovery rate, is the critical load lost in the ith node, P key,i is the critical load of the ith node, nkey Indicates the number of critical load nodes;
[0100] Step 4.2: Establish planning and operation constraints for distribution network and mobile energy storage
[0101] Step 4.2.1. Power flow constraints
[0102]
[0103] In formula (22), P i,t and Q i,t are the active power and reactive power of the ith node at time t, U i,t and U j,t are the voltage values of the i-th node and the j-th node, G ij and B ij are the conductance and susceptance between the i-th node and the j-th node, θ ij The voltage phase angle between the i-th node and the j-th node; N1 represents the number of distribution network nodes;
[0104] Step 4.2.2: Voltage Constraint
[0105]
[0106] In formula (23), and are the upper and lower limits of the voltage of the i-th node respectively;
[0107] Step 4.2.3: Line transmission power constraints
[0108]
[0109] In formula (24), S ij is the transmission power between lines ij, is the maximum transmission capacity between lines ij, x ij,t is a 0-1 variable, indicating the operation status of line ij at time t, x ij,t =0 indicates line failure, x ij,t =1 means the line is operating normally at time t;
[0110] Step 4.2.4: Power balance constraints
[0111]
[0112] In formula (25), i, j, s are all distribution network nodes, δ(i) and ε(i) are the parent node and child node sets of the i-th node, respectively, and P is,t and Q is,t are the active power and reactive power flowing through line is at time t, is the active power injected by the distributed generation at the i-th node at time t, P ji,t and Q ji,t are the active power and reactive power flowing through line ji at time t, and are respectively the active load and reactive load of the ith node under normal conditions at time t, P ls,i,t and Q ls,i,t are respectively the active load and reactive load lost by the ith node at time t, and are the charging and discharging power of the mobile energy storage at the i-th node at time t respectively;
[0113] Step 4.2.5: Mobile energy storage operation constraints
[0114]
[0115] In formula (26), is the remaining power of the mobile energy storage at the i-th node at time t, γ1 and γ2 are the discharge and charging efficiencies of the mobile energy storage, respectively. and are the minimum and maximum capacities of mobile energy storage at the i-th node respectively; and is a 0-1 variable, indicating the charging and discharging state of the mobile energy storage at time t. If the mobile energy storage at the i-th node is in the discharging state at time t, then and are the maximum charging and discharging powers of the mobile energy storage at the i-th node respectively.
[0116] Furthermore, in step 5, the center of gravity reverse learning-particle swarm algorithm is used to solve the mobile energy storage optimization planning model.
[0117] Furthermore, the specific content of step 5 is as follows:
[0118] Step 5.1: Determine the population size S, maximum number of iterations Q, learning factors c1, c2, and particle flight speed [-v m , v m ], determine the optimization planning model parameters and set the planning cycle;
[0119] Step 5.2, input the initial random particles: the output capacity value of the mobile energy storage in each fault scenario, and record the position of each particle;
[0120] Step 5.3: Calculate the j-th velocity v of particle i at the t-th iteration ij (t) and position x ij (t) performing iterative updates;
[0121] v ij (t) = wv ij (t-1)+c1r1[p best,ij (t-1)-x ij (t-1)]+c2r2[g best,j (t-1)-x ij (t-1)](27)
[0122] x ij (t) = x ij (t-1)+v ij (t) (28)
[0123] In formulas (27) and (28), w is the inertia weight coefficient, c1 and c2 are learning coefficients, r1 and r2 are random numbers uniformly distributed between [0,1], and p best,ij (t) and g best,j (t) are the best positions experienced by particle i and all particles before the tth iteration respectively;
[0124] If the j-th velocity v of particle i at the t-th iteration ij (t)≥v m , then let v ij (t) = v m ; if v ij (t) <v m , then let v ij (t) = -v m , continue to iterate;
[0125] Step 5.4: Calculate the center of the point set G
[0126]
[0127] In formula (29), x i is the i-th particle in the search space, and m is the total number of particles in the search space;
[0128] Step 5.5: Calculate the reverse solution of the current particle
[0129]
[0130] Step 5.6: Use the objective function (20) and (21) in the optimization planning model as the fitness function in the particle swarm algorithm, calculate and compare the fitness value f best (x ij ), Update the particles selectively, retain the optimal fitness value and the corresponding optimal position of the individual particles, and then determine the global optimal position;
[0131] Step 5.7, determine whether the current mobile energy storage output value is within the limit range, or whether the current number of iterations exceeds the maximum number of iterations. If so, output the current mobile energy storage output value and the calculation ends; otherwise, return to step 5.3 and continue the calculation.
[0132] Compared with the prior art, the present invention has the following beneficial effects:
[0133] In view of the subjectivity of predicting extreme faults, the present invention fits the typhoon wind speed distribution, combines the distribution line parameters with the typhoon travel parameters, and uses the spatiotemporal characteristics of the line failure rate during the disaster to reflect the impact of the typhoon disaster on the distribution network caused by line failures. At the same time, the intensity of extreme weather, the duration of the fault, the importance of the load, and the randomness of the load and distributed power output in the distribution network are taken into account, and multiple scenarios of the whole process under the disaster are generated as the basis for subsequent optimization.
[0134] The present invention selects evaluation indicators from the two dimensions of power supply reliability and elastic resilience, which solves the limitation of optimizing only a single time period under extreme disasters, and comprehensively determines the correlation between evaluation indicators through correlation analysis based on models and data, calculates the objective weight of each indicator, and constructs a more accurate and intuitive distribution network reliability and elasticity evaluation indicator system, which can well evaluate the rationality of the mobile energy storage optimization planning scheme.
[0135] The present invention combines the center of gravity reverse learning method with the particle swarm optimization algorithm to solve the mobile energy storage optimization planning model considering all disaster periods. Compared with the traditional optimization algorithm, it has faster convergence speed and better results. BRIEF DESCRIPTION OF THE DRAWINGS
[0136] Figure 1 This is a flow chart of the mobile energy storage planning method of the present invention. DETAILED DESCRIPTION
[0137] The present invention is further described below in conjunction with the embodiments and the accompanying drawings, but the protection scope of the present invention is not limited to the following embodiments. Any modifications and changes made to the present invention within the spirit of the present invention and the protection scope of the claims fall within the protection scope of the present invention.
[0138] This embodiment provides a mobile energy storage planning method that considers the reliability and flexibility of the distribution network under extreme weather conditions, and the steps include: establishing load, distributed power output and distribution network failure scenarios; constructing a distribution network reliability and flexibility evaluation index system to evaluate the effectiveness of the mobile energy storage optimization planning scheme; taking the minimum load shedding and the maximum critical load recovery rate as the objective function, establishing a mobile energy storage planning model that considers all disaster periods; and using the COBL-PSO algorithm to solve the mobile energy storage optimization planning model. Specifically, Figure 1As shown, proceed as follows:
[0139] Step 1: In response to severe disturbances or faults such as natural disasters on the distribution network side, a component failure rate model is established to model the uncertainty scenarios of failure disasters in the distribution network.
[0140] Step 1.1, select extreme weather represented by typhoons and establish a wind speed and direction distribution model:
[0141] The Weibull distribution is used to describe the uncertainty of wind speed at a single moment:
[0142] F(v M ,θ)=p(θ)f(v M ) (1)
[0143] In formula (1), p(θ) is the wind direction frequency function, θ is the angle between the wind direction and the line, and f(v M ) is the Weibull distribution, v M is the maximum wind speed near the center of the typhoon;
[0144] Step 1.2: Establish a typhoon wind field model:
[0145] There are differences in wind speed within the typhoon wind field, and the Jelesnianski circularly symmetric typhoon wind field model is used:
[0146]
[0147] In formula (2): v r is the wind speed at a distance r from the center, the wind direction is the counterclockwise tangent direction of the simulated circle, R is the maximum wind speed radius, which obeys the log-normal distribution, and r is the distance rate from the node to the center of the radius;
[0148] Step 1.3: Considering the influence of wind direction on component failure rate, a component failure rate model including wind speed and wind direction is established:
[0149] The failure rate of the line when the wind speed is v is p1(v):
[0150]
[0151] In formula (3): θ,1 is the influence coefficient of wind direction on line fault, v crit,1 and v abs,1 is the critical wind speed of the line, η norm,1 is the normal operation failure probability of the line, μ1 is the sensitivity coefficient of the line to wind speed;
[0152] When the wind speed is v, the failure rate of the pole is p p (v):
[0153]
[0154] In formula (4): θ,p is the influence coefficient of wind direction on tower failure, v crit,p and v abs,p is the critical wind speed of the pole, η norm,p is the normal operation failure probability of the line, μ p is the sensitivity coefficient of the pole to wind speed;
[0155] Step 1.4: Equivalent the component failure rate model to a series model of the line failure rate model and the pole failure rate model
[0156]
[0157] In formula (5): p h (v) is the failure rate of the hth branch, m h and h are the number of lines and poles of the h-th branch respectively;
[0158] Step 2: Extreme weather scenarios are extracted through the Monte Carlo method, and the K-means clustering algorithm is used to reduce the extreme weather scenarios. Considering the uncertainty of load and distributed power output, Latin hypercube sampling is used to extract load, distributed power output and distribution network fault scenarios.
[0159] Step 2.1: Sampling of extreme weather scenarios:
[0160] The randomness of extreme weather is mainly reflected in the intensity and duration of extreme weather. K extreme weather scenarios are extracted through the Monte Carlo method:
[0161] Use normal distribution to describe the duration of extreme weather T d :
[0162] T d ~N(μ,σ 2 ) (6)
[0163] In formula (6), μ and σ are the expectation and standard deviation of extreme weather duration, respectively;
[0164] Step 2.2: Scene reduction
[0165] Step 2.2.1, determine K extreme weather scenarios as the number of clusters, and randomly extract k initial cluster centers;
[0166] Step 2.2.2: Use K-means to iteratively cluster until the clustering converges and obtain k extreme weather scenarios:
[0167] Cluster centers at the cth iteration for:
[0168]
[0169] In formula (7), is the kth value of the cth iteration c Cluster centers, h e For the cth iteration The e-th node in For the cth iteration The number of nodes in ;
[0170] The sum of the distances from each node to its corresponding cluster center in the cth iteration is f c :
[0171]
[0172] In formula (8), is the distance from the e-th node to its corresponding cluster center in the c-th iteration;
[0173] Step 2.3: Create a fault scenario
[0174] Use Latin hypercube sampling to extract load, distributed generation output and fault scenarios;
[0175] Normal distribution is used to describe the load and distributed generation output:
[0176]
[0177] In formula (9), P t DG and P t L are the actual values of the distributed generation output and load at time t, and are the predicted values of distributed generation output and load at time t respectively.
[0178] Step 3: Select evaluation indicators from multiple perspectives, calculate the objective weights of each indicator, and construct a distribution network reliability and resilience evaluation indicator system
[0179] Step 3.1: Select power supply reliability indicators
[0180] The power supply reliability index is used to reflect the direct impact of power outages on users, and characterizes the power supply reliability from three aspects: time, power shortage and reliability rate.
[0181] Step 3.1.1: Establish the fault average power outage time indicator T SAIDI :
[0182]
[0183] In formula (10), t i is the time of the ith fault outage in the distribution network, and N is the number of fault outages;
[0184] Step 3.1.2: Establish the average power outage and power shortage index E AENS :
[0185]
[0186] In formula (11), D i is the power shortage caused by the ith power outage;
[0187] Step 3.1.2: Establish the average power supply reliability index R ASAI :
[0188]
[0189] In formula (12), t s The statistical period.
[0190] Step 3.2: Select elastic resilience index
[0191] The system's overall recovery capacity evaluation index is used to reflect the overall recovery capacity from a disaster to a normal operating state, and characterizes the elastic recovery capacity from three aspects: time, speed and recovery rate.
[0192] Step 3.2.1. Establish load recovery time index T RCL :
[0193]
[0194] In formula (13), is the scheduling time for the mobile energy storage to move to the i-th node, is the fault repair time on the i-th node;
[0195] Step 3.2.2: Establish load recovery speed index V RSL :
[0196]
[0197] In formula (14), T r is the fault recovery time, n1 is the total number of system load nodes, P r,i,j is the load power restored by the i-th node in the j-th time period;
[0198] Step 3.2.3: Establish the power shortage rate index R ESR :
[0199]
[0200] In formula (15), ω i is the load weight value, P i L is the total load of the ith node, P ln,i,t is the load of the normal operation of the i-th node at time t; Δt represents the length of the time period.
[0201] Step 3.3: Calculate the objective weight of the indicator
[0202] The actual data of indicators in different evaluation scenarios are used to calculate the indicator entropy value and the objective weight of each indicator.
[0203] Step 3.3.1: Construct the evaluation index matrix:
[0204] Assuming there are a scenarios to be evaluated and b evaluation indicators, the evaluation indicator matrix Z is quantitatively constructed:
[0205]
[0206] In formula (16), the element z ij It represents the value of the jth indicator in the i-th evaluation scenario.
[0207] Step 3.3.2, calculate the index proportion:
[0208] For the jth indicator, calculate the proportion p of the i-th evaluation scenario under this indicator ij ;
[0209]
[0210] Step 3.3.3, calculate the index entropy value:
[0211] Calculate the entropy value e of the jth indicator according to the entropy formula j :
[0212]
[0213] Step 3.3.4, calculate the index entropy weight:
[0214] Calculate the entropy weight ω of the jth indicator based on the indicator entropy value j :
[0215]
[0216] Step 4: Taking the pre-disaster loss amount and post-disaster recovery rate as the optimization targets, a mobile energy storage optimization planning model considering the entire disaster period is established.
[0217] Step 4.1, establish the objective function:
[0218] Establish a mobile energy storage planning and operation plan that takes into account all disaster periods, with the pre-disaster loss amount and post-disaster recovery rate as the two optimization targets.
[0219] Step 4.1.1: Minimizing load shedding is the objective function
[0220]
[0221] In formula (20), C cut is the load shedding amount, a i,t is a 0-1 variable, indicating whether the load at the distribution network node i is removed at time t. i,t =0 means the load at the distribution network node i is removed at time t, a i,t =1 means that the load at the distribution network node i is powered normally at time t. is the load of distribution network node i at time t.
[0222] Step 4.1.2: Take the maximum critical load recovery rate as the objective function:
[0223]
[0224] In formula (21), H key is the critical load recovery rate, is the critical load lost in the ith node, P key,i is the critical load of the ith node, n key Indicates the number of critical load nodes.
[0225] Step 4.2: Establish planning and operation constraints for distribution network and mobile energy storage
[0226] Step 4.2.1. Power flow constraints
[0227]
[0228] In formula (22), P i,t and Q i,t are the active power and reactive power of the ith node at time t, U i,t and U j,t and the voltage values of the i-th node and the j-th node, G ij and B ij are the conductance and susceptance between the i-th node and the j-th node, θ ij The voltage phase angle between the i-th node and the j-th node; N1 represents the number of distribution network nodes.
[0229] Step 4.2.2: Voltage Constraint
[0230]
[0231] In formula (23), and are the upper and lower limits of the voltage of the i-th node respectively;
[0232] Step 4.2.3: Line transmission power constraints
[0233]
[0234] In formula (24), S ij The transmission power between lines ij, is the maximum transmission capacity between lines ij, x ij,t is a 0-1 variable, indicating the operation status of line ij at time t, x ij,t =0 indicates line failure, x ij,t =1 indicates that the line operates normally at time t.
[0235] Step 4.2.4: Power balance constraints
[0236]
[0237] In formula (25), i, j, s are all distribution network nodes, δ(i) and ε(i) are the parent node and child node sets of the i-th node, respectively, and P is,t and Q is,t are the active power and reactive power flowing through line is at time t, is the active power injected by the distributed generation at the i-th node at time t, P ji,t and Q ji,t are respectively the active load and reactive load of the ith node under normal conditions at time t, and are respectively the active load and reactive load of the ith node under normal conditions at time t, P ls,i,t and Q ls,i,t are respectively the active load and reactive load lost by the ith node at time t, and are respectively the charging and discharging power of the mobile energy storage at the i-th node at time t.
[0238] Step 4.2.5: Mobile energy storage operation constraints
[0239]
[0240] In formula (26), is the remaining power of the mobile energy storage at the i-th node at time t, γ1 and γ2 are the discharge and charging efficiencies of the mobile energy storage, respectively. and are the minimum and maximum capacities of mobile energy storage at the i-th node, and is a 0-1 variable, indicating the charging and discharging state of the mobile energy storage at time t. If the mobile energy storage at the i-th node is in the discharging state at time t, then and are the maximum charging and discharging powers of the mobile energy storage at the i-th node respectively.
[0241] Step 5: Solve the mobile energy storage optimization planning model
[0242] Combining the advantages of particle swarm optimization, such as simplicity, flexibility, good adaptability and strong optimization ability of center of gravity reverse learning, the center of gravity reverse learning-particle swarm optimization algorithm (COBL-PSO) is used to solve the mobile energy storage optimization planning model.
[0243] Step 5.1: Determine the population size S, maximum number of iterations Q, learning factors c1, c2, and particle flight speed [-v m , v m ], determine the optimization planning model parameters and set the planning cycle.
[0244] Step 5.2, input the initial random particles: the output capacity value of the mobile energy storage in each fault scenario, and record the position of each particle;
[0245] Step 5.3: Calculate the j-th velocity and position x of particle i at the t-th iteration ij (t) performing iterative updates;
[0246] v ij (t) = wv ij (t-1)+c1r1[p best,ij (t-1)-x ij (t-1)]+c2r2[g best,j (t-1)-x ij (t-1)](27)
[0247] x ij (t) = x ij (t-1)+v ij (t) (28)
[0248] In formulas (27) and (28), w is the inertia weight coefficient, c1 and c2 are learning coefficients, r1 and r2 are random numbers uniformly distributed between [0, 1], and p best,ij (t) and g best,j (t) are the best positions experienced by particle i and all particles before the tth iteration respectively;
[0249] If the j-th velocity v of particle i at the t-th iteration ij (t)≥v m , then let vij v(t) = v m ; If v ij v(t) < vm, then let v ij v(t) = -v m , and continue the iteration;
[0250] Step 5.4, Calculate the center G of the point set
[0251]
[0252] In Equation (29), x i is the i-th particle in the search space, and m is the total number of particles in the search space;
[0253] Step 5.5, Calculate the reverse solution of the current particle
[0254]
[0255] Step 5.6, Use the objective functions in Equations (20) and (21) in the optimization planning model as the fitness function in the particle swarm optimization algorithm, calculate and compare the fitness values f best (x ij ), Select and update the particles optimally, retain the optimal fitness value and the corresponding individual optimal position of the particle, and then determine the global optimal position;
[0256] Step 5.7, Judge whether the current obtained range of the mobile energy storage output value or the current iteration number exceeds the maximum iteration number. If so, output the current output value of the mobile energy storage, and the operation ends; otherwise, return to Step 5.3 and continue the operation.
[0257] In summary, the present invention proposes a mobile energy storage planning method considering the reliability and resilience of the distribution network under extreme weather for the fault uncertainty of the distribution network under extreme weather. By simulating the spatio-temporal characteristics of extreme weather, the fault impact is effectively quantified, thereby generating the whole process and multi-scenarios under disasters; Evaluation indexes are selected from two dimensions of power supply reliability and resilience recovery, and the objective weights of each index are calculated to construct a more accurate and intuitive evaluation index system for the reliability and resilience of the distribution network, solving the limitation of only optimizing for a single time period under extreme disasters, and being able to well evaluate the scientificity of the mobile energy storage planning; A mobile energy storage planning model considering the whole disaster period is constructed and solved by the COBL-PSO algorithm, with a faster convergence speed and improved calculation efficiency.
Claims
1. A mobile energy storage planning method considering the reliability and flexibility of distribution networks under extreme weather conditions, characterized in that: include: Step 1: Establish a distribution component failure rate model to model the uncertainty scenario of failure disasters in the distribution network; Step 2: The extreme weather scenarios are extracted by Monte Carlo method and reduced by K-means clustering algorithm. Considering the uncertainty of load and distributed power output, Latin hypercube sampling is used to extract load, distributed power output and distribution network fault scenarios; Step 3: Select evaluation indicators from multiple perspectives, calculate the objective weight of each indicator, and construct a distribution network reliability and flexibility evaluation indicator system; Step 4: Taking the pre-disaster loss and post-disaster recovery rate as optimization targets, a mobile energy storage optimization planning model considering the entire disaster period is established; Step 5, solving the mobile energy storage optimization planning model; The step 4 comprises: Step 4.1: Establish the objective function Establish a mobile energy storage optimization planning scheme that takes into account all disaster periods, with the pre-disaster loss and post-disaster recovery rate as the optimization targets; Step 4.1.1: Minimizing load shedding is the objective function In formula (20), C cut is the load shedding amount, a i,t is a 0-1 variable, indicating whether the load at the distribution network node i is removed at time t. i,t =0 means the load at the distribution network node i is removed at time t, a i,t =1 means that the load at the distribution network node i is powered normally at time t. is the load of distribution network node i at time t, T d To describe the duration of extreme weather using normal distribution, n1 is the total number of system load nodes, ω i is the load weight value; Step 4.1.2: Take the maximum critical load recovery rate as the objective function: In formula (21), H key is the critical load recovery rate, is the critical load lost in the ith node, P key,i is the critical load of the ith node, n key Indicates the number of critical load nodes; Step 4.2: Establish planning and operation constraints for distribution network and mobile energy storage Step 4.2.
1. Power flow constraints In formula (22), P i,t and Q i,t are the active power and reactive power of the ith node at time t, U i,t and U j,t are the voltage values of the i-th node and the j-th node respectively, G ij and B ij are the conductance and susceptance between the i-th node and the j-th node, θ ij The voltage phase angle between the i-th node and the j-th node; N1 represents the number of distribution network nodes; Step 4.2.2: Voltage Constraint In formula (23), and are the upper and lower limits of the voltage of the i-th node respectively; Step 4.2.3: Line transmission power constraints In formula (24), S ij is the transmission power between lines ij, is the maximum transmission capacity between lines ij, x ij,t is a 0-1 variable, indicating the operation status of line ij at time t, x ij,t =0 indicates line failure, x ij,t =1 means the line is operating normally at time t; Step 4.2.4: Power balance constraints In formula (25), i, j, s are all distribution network nodes, δ(i) and ε(i) are the parent node and child node set of the i-th node, respectively, and P is,t and Q is,t are the active power and reactive power flowing through line is at time t, is the active power injected by the distributed generation at the i-th node at time t, P ji,t and Q ji,t are the active power and reactive power flowing through line ji at time t, and are respectively the active load and reactive load of the ith node under normal conditions at time t, P ls,i,t and Q ls,i,t are respectively the active load and reactive load lost by the ith node at time t, and are the charging and discharging power of the mobile energy storage at the i-th node at time t respectively; Step 4.2.5: Mobile energy storage operation constraints In formula (26), is the remaining power of the mobile energy storage at the i-th node at time t, γ1 and γ2 are the discharge and charging efficiencies of the mobile energy storage, respectively. and are the minimum and maximum capacities of mobile energy storage at the i-th node respectively; and is a 0-1 variable, indicating the charging and discharging state of the mobile energy storage at time t. If the mobile energy storage at the i-th node is in the discharging state at time t, then and are the maximum charging and discharging power of the mobile energy storage at the i-th node respectively; In step 5, the center of gravity reverse learning-particle swarm algorithm is used to solve the mobile energy storage optimization planning model.
2. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 1 is characterized in that: The step 1 comprises: Step 1.1: Select extreme weather represented by typhoons and establish a wind speed and direction distribution model The Weibull distribution is used to describe the uncertainty of wind speed at a single moment: F(v M ,θ)=p(θ)f(v M ) (1) In formula (1), p(θ) is the wind direction frequency function, θ is the angle between the wind direction and the line, and f(v M ) is the Weibull distribution, v M is the maximum wind speed near the center of the typhoon; Step 1.2: Establish a typhoon wind field model Using the Jelesnianski circularly symmetric typhoon wind field model: In formula (2), v r is the wind speed at a distance r from the center, the wind direction is the counterclockwise tangent direction of the simulated circle, R is the maximum wind speed radius, which obeys the log-normal distribution, and r is the distance rate from the node to the center of the radius; Step 1.3: Considering the influence of wind direction on component failure rate, a component failure rate model including wind speed and wind direction is established. When the wind speed is v, the failure rate of the line is p1(v): In formula (3), λ θ,1 is the influence coefficient of wind direction on line fault, v crit,1 and v abs,1 is the critical wind speed of the line, η norm,1 is the normal operation failure probability of the line, μ1 is the sensitivity coefficient of the line to wind speed; When the wind speed is v, the failure rate of the pole is p p (v): In formula (4), λ θ,p is the influence coefficient of wind direction on tower failure, v crit,p and v abs,p is the critical wind speed of the pole, η norm,p is the normal operation failure probability of the line, μ p is the sensitivity coefficient of the pole to wind speed; Step 1.4: Equivalent the component failure rate model to a series model of the line failure rate model and the pole failure rate model In formula (5), p h (v) is the failure rate of the h-th branch, mh and yh are the number of lines and poles in the h-th branch, respectively.
3. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 1 is characterized in that: The step 2 comprises: Step 2.1: Sampling of extreme weather scenarios K extreme weather scenarios are extracted by Monte Carlo method, and the duration T of extreme weather is described by normal distribution. d : T d ~N(μ,σ 2 )(6) In formula (6), μ and σ are the expectation and standard deviation of extreme weather duration, respectively; Step 2.2: Scene reduction Step 2.2.1, determine K extreme weather scenarios as the number of clusters, and randomly extract k initial cluster centers; Step 2.2.2: Use K-means to iteratively cluster until the clustering converges and obtain k extreme weather scenarios: Cluster centers at the cth iteration for: In formula (7), is the kth value of the cth iteration c Cluster centers, h e For the cth iteration The e-th node in For the cth iteration The number of nodes in ; The sum of the distances from each node to its corresponding cluster center in the cth iteration is f c : In formula (8), is the distance from the e-th node to its corresponding cluster center in the c-th iteration; Step 2.3: Create a fault scenario Latin hypercube sampling is used to extract load, distributed generation output and distribution network fault scenarios; Normal distribution is used to describe the load and distributed generation output: In formula (9), P t DG and P t L are the actual values of the distributed generation output and load at time t, and are the predicted values of distributed generation output and load at time t respectively.
4. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 1 is characterized in that: The step 3 comprises: Step 3.1: Select power supply reliability indicators The reliability of power supply is characterized from three aspects: time, power shortage and reliability rate; Step 3.2: Select elastic resilience index The elastic recovery force is characterized from three aspects: time, speed and recovery rate; Step 3.3: Calculate the objective weight of the indicator The actual data of indicators in different evaluation scenarios are used to calculate the indicator entropy value and the objective weight of each indicator.
5. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 4 is characterized in that: The step 3.1 includes: Step 3.1.1: Establish the fault average power outage time indicator T SAIDI In formula (10), t i is the time of the ith fault outage in the distribution network, and N is the number of fault outages; Step 3.1.2: Establish the average power outage and power shortage index E AENS In formula (11), D i is the power shortage caused by the ith power outage; Step 3.1.2: Establish the average power supply reliability index R ASAI In formula (12), t s The statistical period.
6. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 4 is characterized in that: The step 3.2 includes: Step 3.2.
1. Establish load recovery time index T RCL In formula (13), is the scheduling time for the mobile energy storage to move to the i-th node, is the fault repair time on the i-th node; Step 3.2.2: Establish load recovery speed index V RSL In formula (14), T r is the fault recovery time, n1 is the total number of system load nodes, P r,i,j is the load power restored by the i-th node in the j-th time period; Step 3.2.3: Establish the power shortage rate index R ESR In formula (15), ω i is the load weight value, P i L is the total load of the ith node, P ln,i,t is the load of the normal operation of the i-th node at time t, and Δt represents the length of the time period.
7. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 4 is characterized in that: The step 3.3 includes: Step 3.3: Calculate the objective weight of the indicator The actual data of indicators in different evaluation scenarios are used to calculate the indicator entropy value and the objective weight of each indicator; Step 3.3.1: Construct the evaluation index matrix: Assuming there are a scenarios to be evaluated and b evaluation indicators, the evaluation indicator matrix Z is quantitatively constructed: In formula (16), the element z ij Indicates the value of the jth indicator in the i-th evaluation scenario; Step 3.3.2, calculate the index proportion: For the jth indicator, calculate the proportion p of the i-th evaluation scenario under this indicator ij ; Step 3.3.3, calculate the index entropy value: Calculate the entropy value e of the jth indicator according to the entropy formula j : Step 3.3.4, calculate the index entropy weight: Calculate the entropy weight ω of the jth indicator based on the indicator entropy value j :
8. The mobile energy storage planning method considering the reliability and flexibility of the distribution network under extreme weather conditions according to claim 1 is characterized in that: The specific content of step 5 is as follows: Step 5.1: Determine the population size S, maximum number of iterations Q, learning factors c1, c2, and particle flight speed [-v m , v m ], determine the optimization planning model parameters and set the planning cycle; Step 5.2, input the initial random particles: the output capacity value of the mobile energy storage in each fault scenario, and record the position of each particle; Step 5.3: Calculate the j-th velocity v of particle i at the t-th iteration ij (t) and position x ij (t) performing iterative updates; v ij (t)=wv ij (t-1)+c1r1[p best,ij (t-1)-x ij (t-1)]+c2r2[g best,j (t-1)-x ij (t-1)] (27) x ij (t)=x ij (t-1)+v ij (t) (28) In formulas (27) and (28), w is the inertia weight coefficient, c1 and c2 are learning coefficients, r1 and r2 are random numbers uniformly distributed between [0,1], and p best,ij (t) and g best,j (t) are the best positions experienced by particle i and all particles before the tth iteration respectively; If the j-th velocity v of particle i at the t-th iteration ij (t)≥v m , then let v ij (t) = v m ; if v ij (t) <v m , then let v ij (t) = -v m , continue to iterate; Step 5.4: Calculate the center of the point set G In formula (29), x i is the i-th particle in the search space, and m is the total number of particles in the search space Step 5.5: Calculate the reverse solution of the current particle Step 5.6: Use the objective function (20) and (21) in the optimization planning model as the fitness function in the particle swarm algorithm, calculate and compare the fitness value f best (x ij ), Update the particles selectively, retain the optimal fitness value and the corresponding optimal position of the individual particles, and then determine the global optimal position; Step 5.7, determine whether the current mobile energy storage output value is within the limit range, or whether the current iteration number exceeds the maximum iteration number. If so, output the current mobile energy storage output value, and the calculation ends; Otherwise, return to step 5.3 and continue the calculation.
