A Robust Optimization Method Considering the Resilience of Distribution Networks

Abstract

The present invention discloses a robust optimization method considering the resilience of a distribution network. It belongs to the field of electric power energy and includes the following contents: obtaining the elastic response ability of the distribution network to faults and the restoration ability after faults, and establishing a performance model of a resilient distribution network system under extreme events; analyzing the changing trend of the system resilience curve in each stage, and combining the system load shedding amount under faults to construct a quantitative index of the distribution network resilience under extreme events; considering the flexible regulation ability of delayed data loads, and establishing a spatio-temporal transfer model of data loads; in order to cope with the losses brought by small-probability - high-loss extreme events to the distribution network, a robust optimization model considering energy storage and spatio-temporal flexible regulation of data loads is established from the perspective of enhancing the survivability of the distribution network; the rationality and feasibility of the robust optimization scheme are verified through examples. The present invention considers the influence of small-probability - high-loss extreme events on the distribution network planning, provides theoretical guidance for power grid planners, and improves the reliability and robustness of the planning scheme.

Description

Technical Field

[0001] The present invention belongs to the field of electric power energy. Specifically, it relates to a robust optimization method considering the resilience of the distribution network, which can provide theoretical guidance for system planners, take into account the impact of small-probability and high-loss extreme events during the planning stage, coordinate the regulation strategies of energy storage and data load in the distribution network, and improve the robustness and reliability of the distribution network planning. Background Art

[0002] In recent years, large-scale power outages in the power grid caused by extreme events such as natural disasters and cyber attacks have occurred frequently. Although the probability of such extreme events is small, the consequences once they occur are usually very serious. Therefore, how to take into account the impact of possible extreme events during the planning stage to ensure that the distribution network not only meets the reliable operation under normal conditions, but also improves its resilience under extreme events is of great significance.

[0003] On the other hand, with the continuous development of intelligence and informatization, the access volume of distributed resources in the distribution network has been continuously increasing, making the distribution network have higher flexibility and recovery ability. The rapid development of technologies such as distributed power sources, energy storage, and data centers provides an opportunity for improving the resilience of the distribution network, and also enables the distribution network to have sufficient flexible regulation resources to respond in a timely manner under fault conditions, improving the reliability and resilience of the system. Summary of the Invention

[0004] Technical Problem: Aiming at the fact that the existing distribution network planning scheme does not fully consider the impact of small-probability and high-loss extreme events, the technical problem to be solved by the present invention is: based on the power regulation characteristics of energy storage and the spatio-temporal flexible regulation ability of data load, to propose a robust optimization method considering the resilience of the distribution network, taking into account the elastic response ability of the distribution network under extreme events, and improving the robustness and reliability of the distribution network planning.

[0005] Technical Solution: To solve the above technical problem, the technical solution adopted by the present invention is: a robust optimization method considering the resilience of the distribution network, including the following steps:

[0006] Obtain the elastic response ability of the distribution network to cope with faults and the recovery ability after faults, and based on the resilience of the distribution network, establish a performance model of the resilient distribution network system under extreme events;

[0007] Analyze the changing trend of the resilience curve of the distribution network system at each stage, and combine the amount of load loss in the system under faults to construct a quantitative index of the resilience of the distribution network under extreme events; where each stage includes the normal operation stage before faults, the disturbance stage, the response stage, the recovery stage, and the final response stage;

[0008] Obtain the flexible regulation ability of delay-type data load and establish a spatio-temporal transfer model of data load;

[0009] Regarding the losses brought by low-probability and high-loss extreme events to the distribution network, based on the resilience distribution network system performance model and the spatio-temporal transfer model of data load, a robust optimization model considering energy storage and spatio-temporal flexible regulation of data load is established from the perspective of improving the viability of the distribution network;

[0010] The rationality and feasibility of the robust optimization scheme are verified based on the actual distribution network.

[0011] A robust optimization method considering the resilience of the distribution network, and the resilience distribution network system performance model under the extreme event includes the following stages:

[0012] T0 - T1: Normal operation stage before the fault. In this stage, the system makes corresponding preparations and preventive measures for possible extreme events through reasonable resource allocation;

[0013] T1 - T2: Disturbance stage. The system encounters a disturbance fault at time T1. At this time, due to the failure of various elastic resources to respond in time, the elastic performance of the system decreases rapidly;

[0014] T2 - T4: Response stage. After a period of time after the disturbance occurs, the system degrades into a stable response state, and various elastic resources are ready to respond to the disturbance fault;

[0015] T4 - T5: Recovery stage. The elastic resources of the system respond to the disturbance fault, and the system performance recovers rapidly, but has not yet recovered to the normal state before the fault;

[0016] T5 - T7: Final response stage. The damaged infrastructure in the system is restored, and the system performance is restored to the normal operation state before the fault.

[0017] A robust optimization method considering the resilience of the distribution network, and the resilience quantification indexes of the distribution network under the extreme event include:

[0018] For the first stage, only consider the impact of extreme events on the viability of the distribution network, that is, it is necessary to consider the resilience quantification indexes of the distribution network in the disturbance stage and the response stage. Generally speaking, the most direct impact of extreme events on the distribution network is the reduction of system load. Considering the time integral of the load loss of the distribution network under extreme events as the resilience quantification index.

[0019]

[0020] Among them, p s is the probability of the occurrence of extreme event s, and S is the set of extreme events; is the reduction value of the active power of node i in the extreme event at time t; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integral with respect to time t.

[0021] Convert the above toughness quantification index into a quantifiable cost index, that is, the annual load shedding cost of the distribution network under the influence of extreme events:

[0022]

[0023] Among them, T ex is the average number of extreme events occurring in a year; is the cost of unit active power loss in the distribution network.

[0024] A robust optimization method considering the toughness of the distribution network, the delay-type data load includes:

[0025] The data load in CPDS is usually divided into two categories: delay-sensitive type and delay-tolerant type. The former requires real-time processing within a short time, and usually uses the M / M / 1 queuing model to model the queuing delay in a time period to ensure that the data load received by the data center in each time period must be processed within that time period; the latter has a high tolerance for the processing time and can be processed within the specified time. The delay-tolerant data load between different data centers can also achieve spatial transfer. Therefore, this type of data load has spatio-temporal adjustment characteristics. To simplify the model and without loss of generality, the present invention mainly considers the delay-tolerant load.

[0026] A robust optimization method considering the toughness of the distribution network, the data load flexible regulation ability includes:

[0027] The delay-tolerant data load in the data center has the potential for spatio-temporal flexible regulation. The data load between different front-end servers and computing nodes should satisfy the following constraints:

[0028]

[0029]

[0030]

[0031] Equation (1) means that the sum of the data loads regulated and allocated by each front-end server should be equal to the local user demand, where Load s,t represents the local user demand of the s-th front-end server at time t, Data l,s,t represents the data load allocated by the s-th front-end server to data center l at time t, and S is the total number of front-end servers; Equation (2) means that the data load of each data center should be equal to the sum of its own and the spatial transfer loads of other data centers, and N is the total number of data centers; Equation (3) means that the data load to be processed by each data center at each moment should be equal to the difference between the data load received by spatial transfer and the time transfer load, where Trans l,tIndicates the data load transferred by data center l at time t.

[0032] Data load flexible regulation ability, and the data load time transfer model includes:

[0033] Delay-tolerant loads do not require real-time processing of data loads and allow them to be processed after a certain delay. Therefore, the data load time transfer model is:

[0034]

[0035] Total l,t+1 = Total l,t + ΔData l,t Δt (5)

[0036] 0 ≤ Total l,t ≤ Total l,max (6)

[0037] In Equation (4), ΔData l,t Indicates the data load transferred by the l-th data center at time t. Data l,t and Trans l,t Are explained the same as in Equation (1); Equation (5) represents the relationship between the total data load storage at different times in the data center, and Δt represents the time interval from t to t + 1; Equation (6) constrains the upper and lower limits of the total data load storage in the data center, where Total l,max Represents the upper limit of the data load storage.

[0038] Data load flexible regulation ability, and the data load space transfer model includes:

[0039] Data loads can be flexibly transferred between different data centers, and its space transfer model is:

[0040]

[0041]

[0042] Since a single data center cannot transfer and absorb loads to any arbitrary data center simultaneously, Equation (8) is added for constraint.

[0043] A robust optimization method considering the resilience of the distribution network, and the robust optimization model includes:

[0044] To address the losses caused by small-probability - high-loss extreme events to the distribution network, the present invention establishes a robust optimization model considering energy storage and spatio-temporal flexible regulation of data loads from the perspective of enhancing the survivability of the distribution network. The robust optimization model is as follows:

[0045]

[0046] such that

[0047] Ax ≤ d

[0048]

[0049]

[0050] where P is the planning set, O is the operation set, and F is the fault set; x is the planning decision vector, including all decision variables involved in the distribution network planning; y is the operation decision vector, including decision variables that can participate in the flexible scheduling during the operation stage of the distribution network; z is the fault scenario vector; a T , b T , c T are the coefficient matrices corresponding to the planning decision vector, operation decision vector, and fault scenario vector respectively; A, B, C, D, G are the coefficient matrices under the corresponding constraint conditions; f is the constant matrix corresponding to the equality constraint condition.

[0051] The above optimization model is a two-stage three-layer robust optimization model; the first stage is the investment stage, in which a reasonable distribution network investment plan is determined based on the probability distribution of limited severe scenarios, and the planning decision vector includes the location and capacity configuration of fixed energy storage; the second stage is the operation stage, and the operation decision variables include the spatio-temporal flexible scheduling plan of the data load, and the most severe scenario probability distribution is sought under the known investment plan in the first stage; based on this, the internal double-layer optimization problem is simulated and decoupled to solve, so as to minimize the annual comprehensive cost of the system under the most severe scenario probability distribution.

[0052] Robust optimization model, the planning set includes:

[0053] Energy storage investment cost:

[0054] where is the investment cost per unit capacity of energy storage, E n is the capacity of the nth energy storage, N E is the number of energy storage planning, y1 is the operation life of energy storage, and d is the discount rate.

[0055] Intelligent terminal investment cost: where is the investment cost of a single intelligent terminal, K is the number of intelligent terminal planning, y2 is the operation life of intelligent terminal, and d is the discount rate.

[0056] Robust optimization model, the operation set includes:

[0057] Energy storage operation cost:

[0058] Among them, is the operation scheduling cost of unit capacity energy storage, represents the charging power or discharging power of the i-th energy storage at time t, and T represents the total number of energy storage charging and discharging time periods.

[0059] Operating cost of the data center:

[0060] Among them, MP t represents the nodal marginal price of the distribution network at time t, N D is the number of data centers, represents the electric energy required for the data center to process unit data load per unit time.

[0061] Robust optimization model, the fault set includes:

[0062] Annual load shedding loss cost of the distribution network under the influence of extreme events:

[0063]

[0064] Among them, T e is the average number of extreme events occurring in a year; is the cost of unit active power loss in the distribution network, p s is the probability of the occurrence of extreme event s, and S is the set of extreme events; is the reduction value of the active power of node i at time t in the extreme event; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integration with respect to time t.

[0065] Robust optimization model, the planning set constraint conditions include:

[0066] (1) Constraints on the rated power and capacity of energy storage allowed to be installed at nodes;

[0067] (2) Constraints on the number of energy storage allowed to be installed in the distribution network;

[0068] (3) Constraints on the number of intelligent terminals allowed to be installed at nodes;

[0069] Robust optimization model, the operation set constraint conditions include:

[0070] (1) Power flow constraints (active power and reactive power constraints);

[0071] (2) Security constraints (voltage and current constraints);

[0072] (3) Energy storage constraints (energy storage state of charge constraint, energy storage capacity constraint, energy storage power balance constraint);

[0073] (4) Data payload constraints (data payload time transfer volume constraint, data payload space transfer volume constraint);

[0074] (5) Communication bandwidth constraint.

[0075] Beneficial effects: Compared with the prior art, the present invention has the following characteristics:

[0076] Based on a cyber-physical system of a distribution network containing a data center and distributed resources, the present invention proposes a method for integrated planning of energy storage and data center in a distribution network considering cyber-physical coupling. The planning scheme considers comprehensively utilizing the power and voltage regulation characteristics of distributed energy storage and the spatio-temporal transfer potential of the data payload of the data center, collaboratively plans the physical-side energy storage configuration, the spatio-temporal transfer mode of the data payload, and the information-side communication network topology, proposes a CPDS integrated planning model, realizes the minimization of the planning cost, and optimizes the communication topology of the distribution network. The present invention takes into account the influence of cyber-physical coupling, can improve the consumption capacity of distributed energy in the distribution network, and reduce the system power operation cost. Description of the drawings

[0077] Figure 1 It is a block diagram of a method for integrated planning of energy storage and data center in a distribution network considering cyber-physical coupling according to the present invention;

[0078] Figure 2 It is a curve graph of the resilience performance of the distribution network at each stage according to the present invention;

[0079] Figure 3 It is a schematic diagram of the physical model and processing flow of the data center according to the present invention;

[0080] Figure 4 It is the IEEE-33 node distribution network model according to the present invention;

[0081] Figure 5 It is the typical daily photovoltaic output curve and load curve according to the present invention;

[0082] Figure 6 It is the typical daily data payload curve of the data center according to the present invention. Detailed implementation manners

[0083] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

[0084] A robust optimization method considering the resilience of a distribution network includes the following steps:

[0085] Obtain the elastic response ability of the distribution network to faults and the recovery ability after faults (distribution network resilience), and establish a system performance model of a resilient distribution network under extreme events;

[0086] Analyze the changing trend of the distribution network system resilience curve in each stage, and combine the system load shedding amount under faults to construct a distribution network resilience quantification index under extreme events; where each stage includes the normal operation stage before faults, the disturbance stage, the response stage, the recovery stage, and the final response stage;

[0087] Obtain the flexible regulation ability of delayed data loads, and establish a spatio-temporal transfer model for data loads;

[0088] Aiming at the losses brought by small-probability - high-loss extreme events to the distribution network, according to the system performance model of the resilient distribution network and the spatio-temporal transfer model of data loads, establish a robust optimization model considering energy storage and spatio-temporal flexible regulation of data loads from the perspective of improving the survivability of the distribution network;

[0089] Verify the rationality and feasibility of the robust optimization scheme based on the actual distribution network.

[0090] A robust optimization method considering the resilience of the distribution network, and the system performance model of the resilient distribution network under extreme events includes:

[0091] Under extreme events, the distribution network resilience performance curves in each stage are as Figure 2 shown, where:

[0092] T0 - T1: Normal operation stage before faults. In this stage, the system makes corresponding preparations and prevention for possible extreme events through reasonable resource allocation;

[0093] T1 - T2: Disturbance stage. The system encounters a disturbance fault at time T1, and at this moment, due to the failure of various elastic resources to respond in time, the elastic performance of the system rapidly decreases;

[0094] T2 - T4: Response stage. After a period of time after the disturbance occurs, the system degrades into a stable response state, and various elastic resources are ready to respond to the disturbance fault;

[0095] T4 - T5: Recovery stage. The elastic resources of the system respond to the disturbance fault, and the system performance rapidly recovers, but has not yet recovered to the normal state before the fault;

[0096] T5 - T7: Final response stage. The damaged infrastructure in the system is restored, and the system performance is restored to the normal operation state before the fault.

[0097] A robust optimization method considering the resilience of the distribution network, and the distribution network resilience quantification index under extreme events includes:

[0098] For the first stage, only consider the impact of extreme events on the viability of the distribution network, that is, it is necessary to consider the quantitative indicators of the resilience of the distribution network during the disturbance stage and the response stage. Generally speaking, the most direct impact of extreme events on the distribution network is the reduction of system load. Consider taking the time integral of the load loss of the distribution network under extreme events as the quantitative indicator of resilience.

[0099]

[0100] Among them, p s is the probability of the occurrence of extreme event s, and S is the set of extreme events; is the reduction value of the active power of node i at time t during the extreme event; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integral with respect to time t.

[0101] Convert the above-mentioned quantitative resilience index into a cost index that can be quantitatively calculated, that is, the annual load shedding loss cost of the distribution network under the influence of extreme events:

[0102]

[0103] Among them, T e is the average number of occurrences of extreme events in a year; is the cost of unit active power loss in the distribution network, p s is the probability of the occurrence of extreme event s, and S is the set of extreme events; is the reduction value of the active power of node i at time t during the extreme event; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integral with respect to time t.

[0104] A robust optimization method considering the resilience of the distribution network, the delay-type data load includes:

[0105] The data load in the CPDS is usually divided into two categories: delay-sensitive type and delay-tolerant type. The former requires real-time processing within a short time, and usually uses the M / M / 1 queuing model to model the queuing delay within a time period to ensure that the data load received by the data center within each time period must be processed within that time period; the latter has a high tolerance for the processing time and can be processed within the specified time. The delay-tolerant data load between different data centers can also achieve spatial transfer. Therefore, this type of data load has spatio-temporal adjustment characteristics. To simplify the model and without loss of generality, the present invention mainly considers the delay-tolerant load.

[0106] A robust optimization method considering the resilience of the distribution network, the data load flexible regulation ability includes:

[0107] The delay-tolerant data load in the data center has the potential for spatio-temporal flexible regulation, and its spatio-temporal adjustment method is asFigure 3 As shown in the figure. The data load between different front-end servers and computing nodes should satisfy the following constraints:

[0108]

[0109]

[0110]

[0111] Equation (1) indicates that the sum of the data loads regulated and allocated by each front-end server should be equal to the local user demand, where Load s,t represents the local user demand of the s-th front-end server at time t, and Data l,s,t represents the data load allocated by the s-th front-end server to data center l at time t, and S is the total number of front-end servers; Equation (2) indicates that the data load of each data center should be equal to the sum of its own and the space transfer loads of other data centers, and N is the total number of data centers; Equation (3) indicates that the data load to be processed by each data center at each moment should be equal to the difference between the data load received by space transfer and the time transfer load, where Trans l,t represents the amount of data load transferred by data center l at time t.

[0112] Data load flexible regulation ability, the data load time transfer model includes:

[0113] The delay-tolerant load does not require real-time processing of the data load and allows it to be processed after a certain delay. Therefore, the data load time transfer model is:

[0114]

[0115] Total l,t+1 = Total l,t + ΔData l,t Δt (5)

[0116] 0 ≤ Total l,t ≤ Total l,max (6)

[0117] In Equation (4), ΔData l,t represents the amount of data load transferred by the l-th data center at time t, and the explanations of Data l,t and Trans l,t are the same as those in Equation (1); Equation (5) represents the relationship between the total data load storage of the data center at different times, and Δt represents the time interval from t to t + 1; Equation (6) constrains the upper and lower limits of the total data load storage of the data center, where Total l,max represents the upper limit of the data load storage.

[0118] The ability to flexibly regulate the data payload, and the data payload spatial transfer model includes:

[0119] The data payload can be flexibly transferred between different data centers, and its spatial transfer model is:

[0120]

[0121]

[0122] Since a single data center cannot transfer out and absorb the load to any arbitrary data center, Equation (8) is added for constraint.

[0123] A robust optimization method considering the resilience of the distribution network, and the robust optimization model includes:

[0124] To cope with the losses brought by small-probability - high-loss extreme events to the distribution network, the present invention establishes a robust optimization model considering the spatio-temporal flexible regulation of energy storage and data payload from the perspective of enhancing the viability of the distribution network. The robust optimization model is as follows:

[0125]

[0126] s.t.

[0127] Ax ≤ d

[0128]

[0129]

[0130] Wherein, P is the planning set, O is the operation set, and F is the fault set; x is the planning decision vector, including all decision variables involved in the distribution network planning; y is the operation decision vector, including decision variables that can participate in the flexible scheduling during the operation stage of the distribution network; z is the fault scenario vector; a T , b T , c T are the coefficient matrices corresponding to the planning decision vector, operation decision vector, and fault scenario vector respectively; A, B, C, D, G are the coefficient matrices under the corresponding constraint conditions; f is the constant matrix corresponding to the equality constraint condition.

[0131] The above optimization model is a two-stage and three-layer robust optimization model. The first stage is the investment stage, in which a reasonable distribution network investment plan is determined based on the probability distribution of limited severe scenarios. The planning decision vector includes the location and capacity configuration of fixed energy storage. The second stage is the operation stage, and the operation decision variables include the spatio-temporal flexible scheduling plan of data load, and the worst-case scenario probability distribution is sought under the known investment plan in the first stage. Based on this, the internal double-layer optimization problem is simulated and decoupled to solve, so as to minimize the annual comprehensive cost of the system under the worst-case scenario probability distribution.

[0132] Robust optimization model, and the planning set includes:

[0133] Energy storage investment cost:

[0134] Among them, is the investment cost per unit capacity of energy storage, E n is the capacity of the nth energy storage, N E is the number of energy storage planning, y1 is the operation life of energy storage, and d is the discount rate.

[0135] Intelligent terminal investment cost:

[0136] Among them, is the investment cost of a single intelligent terminal, K is the number of intelligent terminal planning, y2 is the operation life of intelligent terminal, and d is the discount rate.

[0137] Robust optimization model, and the operation set includes:

[0138] Energy storage operation cost:

[0139] Among them, is the operation and scheduling cost per unit capacity of energy storage, represents the charging power or discharging power of the ith energy storage at time t, and T represents the total number of energy storage charging and discharging periods.

[0140] Data center operation cost:

[0141] Among them, MP t represents the nodal marginal price of the distribution network at time t, N D is the number of data centers, represents the electric energy required for the data center to process a unit of data load per unit time.

[0142] Robust optimization model, and the fault set includes:

[0143] Annual load shedding loss cost of the distribution network under the influence of extreme events:

[0144]

[0145] Among them, T e is the average number of extreme events occurring in a year; is the cost of unit active power loss in the distribution network, and p s is the probability of extreme event s occurring, and S is the set of extreme events; is the reduction value of the active power of node i at time t in the extreme event; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integral with respect to time t.

[0146] Robust optimization model, and the constraint conditions of the planning set include:

[0147] (1) Constraints on the rated power and capacity of energy storage allowed to be installed at nodes;

[0148] (2) Constraints on the number of energy storage allowed to be installed in the distribution network;

[0149] (3) Constraints on the number of intelligent terminals allowed to be installed at nodes;

[0150] Robust optimization model, and the constraint conditions of the operation set include:

[0151] (1) Power flow constraints (active power and reactive power constraints);

[0152] (2) Safety constraints (voltage and current constraints);

[0153] (3) Energy storage constraints (energy storage state of charge constraints, energy storage capacity constraints, energy storage power balance constraints);

[0154] (4) Data load constraints (data load time transfer amount constraints, data load space transfer amount constraints);

[0155] (5) Communication bandwidth constraints.

[0156] The following details an optional implementation manner of the present invention.

[0157] In an example of the present invention: applying the above topological identification method to the modified IEEE-33 node distribution network model as Figure 4 shown. The rated voltage is 12.66 kV, and the rated active power of the distribution network is 4000 kW.

[0158] Among them, distributed photovoltaics are installed at nodes 2, 6, 10, 13, 18, 22, 26, 29, and 33. The maximum installed capacity of the node photovoltaic is 500 kW, and the maximum installed capacity of the energy storage is 200 kW·h. The power curves of the photovoltaic system and the load are as Figure 5As shown, the power factor of the load is 0.95. Taking typhoon as an example for extreme events, it is assumed that the average annual occurrence times of typhoon are 10 times, and the moving speed is 30 km / h. Figure 2 Starting from time T1 in Figure 2 , the links 13-14, 26-27, and 7-8 are attacked in sequence, and the maximum number of faults in the distribution line at a certain time section is 3.

[0159] Data centers IDC1-IDC4 are respectively installed at nodes 5, 10, 18, and 26. Assuming that the fault recovery time of the distribution network generally does not exceed 2 hours, the data load scheduling is in units of 15 minutes.

[0160] In addition, other parameter settings in the collaborative robust optimization model of energy storage and data load considering the improvement of distribution network resilience are shown in the following table.

[0161] Table 1 Parameter settings

[0162]

[0163] The results of the energy storage capacity configuration plan and the data load scheduling plan are as follows:

[0164] Table 2 Results of the energy storage capacity configuration plan

[0165]

[0166] Table 3 Annual comprehensive cost of robust optimization

[0167]

[0168] Table 4 Space-time transfer volume of data load in each time period

[0169]

[0170] The collaborative robust optimization of energy storage and data load considering the improvement of distribution network resilience is realized through a computer simulation program, and finally the schematic diagram of the planning results as shown in Figure 6 can be obtained.

[0171] In the description of this specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0172] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and all such changes and improvements fall within the scope of the present invention claimed.

Claims

1. A robust optimization method considering the resilience of the distribution network, characterized in that, It includes the following steps: Obtain the elastic response ability of the distribution network to faults and the recovery ability after faults, and establish a performance model of a resilient distribution network system under extreme events; Analyze the changing trend of the resilience curve of the distribution network system in each stage, and combine the load shedding amount of the system under faults to construct a quantitative index of the resilience of the distribution network under extreme events; Among them, each stage includes the normal operation stage before the fault, the disturbance stage, the response stage, the recovery stage, and the final response stage; Obtain the flexible regulation ability of the delayed data load, and establish a spatio-temporal transfer model of the data load; Aiming at the losses brought by small-probability and high-loss extreme events to the distribution network, according to the performance model of the resilient distribution network system and the spatio-temporal transfer model of the data load, establish a robust optimization model considering energy storage and spatio-temporal flexible regulation of the data load from the perspective of improving the survivability of the distribution network; Verify the rationality and feasibility of the robust optimization scheme based on the actual distribution network; The quantitative index of the resilience of the distribution network under the extreme event includes: For the first stage, only consider the impact of extreme events on the survivability of the distribution network, that is, it is necessary to consider the quantitative index of the resilience of the distribution network in the disturbance stage and the response stage; Generally speaking, the most direct impact of extreme events on the distribution network is the reduction of the system load. Consider taking the time integral of the load loss of the distribution network under extreme events as the quantitative index of resilience; where p s is the probability of the occurrence of the extreme event s, and S is the set of extreme events; is the reduction value of the active power of node i in the extreme event during time period t; N is the set of nodes in the distribution network; T1 and T4 represent different state stages; dt represents the integration with respect to time t; Convert the above quantitative index of resilience into a cost index that can be quantitatively calculated, that is, the annual load shedding loss cost of the distribution network under the influence of extreme events: Among them, T e is the average number of extreme events occurring in a year; is the cost of unit active power loss in the distribution network, p s is the probability of extreme event s occurring, and S is the set of extreme events; is the reduction value of the active power of node i in the extreme event at time t; N is the set of distribution network nodes; T1 and T4 are different state time periods; dt represents the integration with respect to time t; The delayed data load includes: The data loads in the cyber-physical distribution system (CPDS) are usually divided into two categories: delay-sensitive and delay-tolerant. The former requires real-time processing within a short time. Usually, the M / M / 1 queuing model is used to model the queuing delay within a time period to ensure that the data loads received by the data center in each time period must be processed within that time period; The latter has a high tolerance for the processing time and can be processed within the specified time. The delay-tolerant data loads between different data centers can also achieve spatial transfer.

2. A robust optimization method considering the resilience of a distribution network according to claim 1, characterized in that, The performance model of the resilient distribution network system under the extreme event includes the following stages: T0-T1: Normal operation stage before the fault; In this stage, the system makes corresponding preparations and preventions for possible extreme events through reasonable resource allocation; T1-T2: Disturbance stage; The system encounters a disturbance fault at time T1. At this moment, due to the failure of various elastic resources to respond in time, the elastic performance of the system rapidly decreases; T2-T4: Response stage; After a period of time after the disturbance occurs, the system degrades into a stable response state, and various elastic resources are ready to respond to the disturbance fault; T4-T5: Recovery stage; The elastic resources of the system respond to the disturbance fault, and the system performance rapidly recovers, but has not yet recovered to the normal state before the fault; T5-T7: Final response stage; The damaged infrastructure in the system is restored, and the system performance is restored to the normal operation state before the fault.

3. A robust optimization method considering the resilience of the distribution network according to claim 1, characterized in that, The flexible regulation ability of the data load includes: The delay-tolerant data loads in the data center have the potential for spatio-temporal flexible regulation, and the data loads between different front-end servers and computing nodes should satisfy the following constraints: Equation (1) indicates that the sum of the data loads regulated and allocated by each front-end server should be equal to the local user demand, where Load s,t represents the local user demand of the s-th front-end server at time t, and Data l,s,t represents the data load allocated by the s-th front-end server to data center l at time t. S is the total number of front-end servers; Equation (2) indicates that the data load of each data center should be equal to the sum of its own and the space transfer loads of other data centers. N is the total number of data centers; Equation (3) indicates that the data load to be processed by each data center at each moment should be equal to the difference between the data load received by space transfer and the time transfer load, where Trans l,t represents the amount of data load transferred by data center l at time t.

4. A robust optimization method considering the resilience of a distribution network according to claim 3, characterized in that The time transfer model of the data load includes: Delay-tolerant loads do not require real-time processing of data loads and allow them to be processed after a certain delay. Therefore, the time transfer model of data loads is as follows: Total l,t+1 = Total l,t + ΔData l,t Δt (5) 0≤Total l,t ≤Total l,max (6) Formula (4) ΔData l,t represents the data load of the l-th data center transferred at time t. Data l,t and Trans l,t are the same as those in Formula (1); Formula (5) represents the relationship of the total data load stored at different times in the data center, and Δt represents the time interval from t to t+1; Formula (6) restricts the upper and lower limits of the total data load stored in the data center, where Total l,max represents the upper limit of the data load storage volume.

5. A robust optimization method considering the resilience of a distribution network according to claim 3, characterized in that The spatial transfer model of the data load includes: Data loads can be flexibly transferred between different data centers, and its spatial transfer model is as follows: Since a single data center cannot transfer and absorb loads to any arbitrary data center simultaneously, Equation (8) is added for constraint.

6. A robust optimization method considering the resilience of a distribution network according to claim 1, characterized in that The robust optimization model includes: To address the losses brought by small-probability and high-loss extreme events to the distribution network, the present invention establishes a robust optimization model considering the spatio-temporal flexible regulation of energy storage and data loads from the perspective of enhancing the survivability of the distribution network. The robust optimization model is as follows: s.t. Ax ≤ d where $P$ is the planning set, $O$ is the operation set, and $F$ is the fault set; $x$ is the planning decision vector, including all decision variables involved in the distribution network planning; $y$ is the operation decision vector, including decision variables that can participate in the flexible scheduling during the operation stage of the distribution network; $z$ is the fault scenario vector; $a$ T , $b$ T , $c$ T are the coefficient matrices corresponding to the planning decision vector, operation decision vector, and fault scenario vector respectively; $A$, $B$, $C$, $D$, and $G$ are the coefficient matrices under the corresponding constraint conditions; $f$ is the constant matrix corresponding to the equality constraint conditions; The above optimization model is a two-stage three-layer robust optimization model. The first stage is the investment stage, where a reasonable distribution network investment plan is determined based on the probability distribution of limited severe scenarios. The planning decision vector includes the location and capacity configuration of fixed energy storage. The second stage is the operation stage, where the operation decision variables include the spatio-temporal flexible scheduling plan of data loads, and the most severe scenario probability distribution is sought under the known investment plan in the first stage. Based on this, the internal double-layer optimization problem is simulated and decoupled to minimize the annual comprehensive cost of the system under the most severe scenario probability distribution. The planning set includes: Energy storage investment cost: Among them, is the investment cost of unit capacity energy storage, E n is the nth energy storage capacity, N E is the number of energy storage planning, y1 is the operation years of energy storage, and d is the discount rate; Intelligent terminal investment cost: Among them, is the investment cost of a single intelligent terminal, K is the planned number of intelligent terminals, y2 is the operation years of the intelligent terminal, and d is the discount rate; The operation set includes: Energy storage operation cost: Among them, is the operation and scheduling cost of unit-capacity energy storage, represents the charging power or discharging power of the i-th energy storage at time t, and T represents the total number of energy storage charging and discharging time periods; Data center operating cost: Among them, MP t represents the nodal marginal price of the distribution network at time t, N D is the number of data centers, represents the electric energy required for the data center to process a unit of data load per unit time; The fault set includes: The annual load shedding loss cost of the distribution network under the influence of extreme events: Among them, T e is the average number of extreme events occurring in a year; is the cost of active power loss per unit in the distribution network, p s is the probability of extreme event s occurring, and S is the set of extreme events; is the reduction value of active power of node i at time t in the extreme event; N is the set of distribution network nodes; T1 and T4 correspond to different state time periods in claim 2; dt represents the integral with respect to time t.

7. A robust optimization method considering the resilience of a distribution network according to claim 6, characterized in that, The constraint conditions of the planning set include: (1) The rated power and capacity constraints of energy storage allowed to be installed at nodes (2) The number constraints of energy storage allowed to be installed in the distribution network (3) The number constraints of intelligent terminals allowed to be installed at nodes.

8. The robust optimization method considering the resilience of the distribution network according to claim 6, wherein The constraint conditions of the operation set include: (1) Power flow constraints (active power and reactive power constraints) (2) Security constraints (voltage and current constraints) (3) Energy storage constraints (energy storage state of charge constraint, energy storage capacity constraint, energy storage charge-discharge balance constraint) (4) Data load constraints (data load time transfer amount constraint, data load spatial transfer amount constraint) (5) Communication bandwidth constraint.

Images