Urban power grid load recovery method and device based on user requirements

Abstract

The invention provides an urban power grid load recovery method and device based on user demands, and is suitable for the technical field of power systems, and the method comprises the steps: obtaining a historical power load data set of each target user of a post-disaster city, and maintenance resources, a power distribution network topological structure and a power equipment parameter set of the post-disaster city; for each target user, based on the historical load data set, performing power utilization analysis on the target user to obtain a power utilization demand of the target user; inputting the power demands of all the target users, the power distribution network topological structure of the post-disaster city and the power equipment parameter set into a pre-constructed power distribution network post-disaster recovery model, and solving by taking strong maintenance timeliness as a target to obtain a scheduling strategy; the power distribution network post-disaster recovery model comprises a maintenance resource scheduling unit; and scheduling power resources and maintenance resources of the post-disaster city based on the scheduling strategy. The method can improve the recovery speed of the post-disaster urban power grid and reduce the economic loss.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power systems, and in particular relates to a method and device for urban power grid load recovery based on user demand. Background Art

[0002] After a natural disaster, the recovery of a city's distribution network is extremely important. It is directly related to the rapid recovery of urban infrastructure and directly affects the continuity of residents' lives and urban economic activities. In this process, fully considering the power supply security needs of users is the core of the recovery work. By rationally allocating human, material and financial resources, maintaining communication with users, timely updating the recovery progress, and formulating emergency measures and long-term plans, the disaster resistance and flexibility of the distribution network can be improved, and the impact of future disasters on the power supply system can be reduced.

[0003] In existing research on post-disaster recovery, historical data is usually screened or forecasted using tools to obtain time-series power curves, which are then used to describe user power demand. A mixed integer linear programming model for post-disaster distribution network topology optimization is then constructed using the distribution network branch flow model, and a series of optimization problems involving a large number of 0, 1 variables and continuous variables are solved using classic optimization solvers such as CPLEX (Concert Technology for Extending Linear Programming Extraction) or GUROBI (Gurobi Optimizer).

[0004] However, in the process of clarifying the power demand of key users after the disaster, the standard recurrent neural network cannot fully handle the time series characteristics of the input data, and will face the problem of gradient vanishing and gradient exploding. Classical optimization solvers have difficulties in solving such problems, such as low solution efficiency and inability to solve. Summary of the invention

[0005] The embodiment of the present invention provides a method and device for urban power grid load recovery based on user needs, so as to solve the problems of gradient disappearance, gradient explosion and low solution efficiency in the existing post-disaster recovery methods.

[0006] The present invention is achieved through the following technical solutions: In a first aspect, an embodiment of the present invention provides a method for restoring a load of an urban power grid based on user demand, comprising: Obtain the historical power load data set of each target user in the post-disaster city, as well as the maintenance resources, distribution network topology and power equipment parameter set of the post-disaster city; For each target user, based on the historical load data set of the target user, the power consumption of the target user is analyzed to obtain the power demand of the target user; The electricity demand of all target users, the distribution network topology of the city after the disaster, and the power equipment parameter set are input into the pre-built distribution network post-disaster recovery model, and the solution is carried out with the goal of strong maintenance timeliness to obtain the dispatching strategy; the distribution network post-disaster recovery model includes the power system operation unit, the topology reconstruction unit, and the maintenance resource dispatching unit; The power resources and maintenance resources of the post-disaster city are dispatched based on the dispatch strategy.

[0007] In a second aspect, an embodiment of the present invention provides a user-demand-based urban power grid load recovery device, comprising: An acquisition module is used to obtain the historical power load data set of each target user in the post-disaster city, as well as the maintenance resources, distribution network topology and power equipment parameter set of the post-disaster city; An analysis module is used to analyze the power consumption of each target user based on the historical load data set of the target user to obtain the power consumption demand of the target user; The solution module is used to input the power demand of all target users, the distribution network topology of the post-disaster city, and the power equipment parameter set into the pre-built distribution network post-disaster recovery model, and solve it with the goal of strong maintenance timeliness to obtain the dispatching strategy; the distribution network post-disaster recovery model includes the power system operation unit, the topology reconstruction unit, and the maintenance resource dispatching unit; The dispatching module is used to dispatch the power resources and maintenance resources of the post-disaster city based on the dispatching strategy.

[0008] The embodiment of the present invention provides a method and device for urban power grid load recovery based on user demand, obtains the electricity demand of the corresponding target user through the historical load data set of each target user, pre-constructs a distribution network post-disaster recovery model including a power system operation unit, a topology reconstruction unit and a maintenance resource scheduling unit, inputs the electricity demand of all target users, the distribution network topology structure of the post-disaster city and all parameters in the power equipment parameter set into the pre-constructed distribution network post-disaster recovery model, obtains a scheduling strategy, and schedules the power resources and maintenance resources of the post-disaster city according to the scheduling strategy. The electricity demand of the target user is clarified, and in the process of scheduling, by determining the target user, considering the electricity demand of the target user, and taking strong maintenance timeliness as the goal, the speed of urban power grid recovery after the disaster can be effectively improved, and economic losses can be further reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0010] Figure 1 It is a flow chart of a method for urban power grid load restoration based on user demand provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of an application scenario of a method for urban power grid load restoration based on user needs provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a process flow of a target user post-disaster electricity demand prediction model based on LSTM in a method for urban power grid load restoration based on user demand provided by an embodiment of the present invention; Figure 4 It is an algorithm flow chart of solving the main problem and sub-problems of the Benders algorithm of the urban power grid load restoration method based on user demand provided by an embodiment of the present invention; Figure 5 It is a structural schematic diagram of a city power grid load recovery device based on user demand provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0011] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0012] Figure 1 is a flow chart of a method for urban power grid load restoration based on user demand provided by an embodiment of the present invention, with reference to Figure 1 , the urban power grid load restoration method based on user demand is described in detail as follows: S110, obtaining a historical power load data set of each target user in the post-disaster city, as well as a maintenance resource, a distribution network topology, and a power equipment parameter set in the post-disaster city.

[0013] Optionally, the target users refer to users who are given priority protection after a disaster and are confirmed in advance.

[0014] It should be noted that after an extreme event, power lines may break and towers may collapse, which may interrupt the power transmission channel and affect power supply. Figure 2 In an emergency situation, it is necessary to establish functional department centers such as emergency management center, overall command and decision-making center, maintenance center, meteorological center, transportation center, and data sharing center.

[0015] Among them, the maintenance center is used to collect fault conditions and maintenance resource information; the meteorological center is used to report post-disaster weather warning information, and meteorological conditions will affect the progress of the repair process; the transportation center is used to grasp the real-time traffic conditions of the transportation network and provide guidance for the maintenance team's travel routes.

[0016] These functional department centers will pass information to the data sharing center to provide information and data support for the decision-making center to formulate post-disaster recovery strategies. Among them, the data sharing center needs to ensure the security of shared data and maintain the privacy of key core data of each department.

[0017] After obtaining information, the decision-making center is used to decompose such large-scale mixed integer linear programming optimization problems, interactively and iteratively solve the decomposed main and sub-problems, and feed back the relevant optimization results to the data sharing center. The various department centers collaborate to complete the task of post-disaster recovery of the urban power grid in accordance with the optimization results of the data sharing center, so as to achieve the purpose of quickly restoring power supply and reducing economic losses.

[0018] S120: For each target user, based on the historical load data set of the target user, perform power consumption analysis on the target user to obtain the power consumption demand of the target user.

[0019] In this embodiment, the target user's electricity consumption analysis is performed through the LSTM (Long Short Term Memory) neural network model. The target user's electricity consumption analysis can also be performed through other methods, such as the auto-estimation model (ARIMA, Auto Regressive Integrated Moving Average), which is not specifically limited in this application.

[0020] In one possible implementation, for each target user, based on the historical load data set of the target user, an electricity consumption analysis is performed on the target user to obtain the electricity demand of the target user, including: for each target user, the following steps are performed: obtaining the electricity influencing factors of the post-disaster city, and constructing the feature vector of the target user based on the electricity influencing factors and the historical load data set of the target user; inputting the feature vector of the target user into a pre-constructed LSTM network model to obtain the electricity demand of the target user.

[0021] Among them, the factors affecting electricity include meteorological factors and disaster factors. Meteorological factors mainly include minimum temperature, air quality, sunshine hours, working days, weather type and other factors, while disaster factors mainly include the degree of disaster.

[0022] Optionally, for each target user, the input of the LSTM neural network model is the feature vector of the target user, and the feature vector of the target user is as follows:

[0023] in, For the i Target users in t The feature vector at time, is the minimum temperature, For air quality, is the number of sunshine hours, For working days, For weather type, For the degree of disaster, For the i The historical load data of each target user.

[0024] The output of the LSTM neural network model is the target user’s t Electricity demand at all times.

[0025] See also Figure 3 The network structure of the LSTM neural network model mainly includes an input layer, a hidden layer and an output layer. The input of the input layer is the feature vector of the target user that needs to be predicted. The hidden layer contains several node structures of LSTM. The design core of LSTM is the threshold mechanism, including input gate, forget gate, memory unit and output gate.

[0026] Among them, the input gate is mainly used to screen new information. Through the input gate, the information input into the memory unit state can be determined and the memory unit state at the current moment can be updated. It also determines which information in the user's electricity demand and electricity usage habits at the previous moment should be retained. Its mathematical model is as follows:

[0027]

[0028] in, is the input of the input gate at time t; σ is the Sigmoid function; and is the network weight of the input gate, is the bias of the input gate, for t -1 The user's electricity usage information is retained at all times; is the state of the candidate memory unit at time t; and is the memory unit network weight, Bias for the memory cell.

[0029] The forget gate is used to control the impact of the memory of the previous time step on the current time step. It determines which user power consumption information should be discarded. Its mathematical model is:

[0030] in, for t Always forget the input of the gate; and is the forget gate network weight, For the forget gate bias, the LSTM forget gate value can be selected between [0, 1] to achieve the use of LSTM to improve the gradient vanishing and gradient exploding situations in the process of processing user electricity demand information.

[0031] The memory cell state is similar to a conveyor belt, which runs through the entire LSTM neural model. Its function is to store the network information of the LSTM at the current moment and pass it downward. The calculation formula of the memory cell state is as follows:

[0032] in, is the state of the memory unit at time t-1; It is the updated state of the memory unit after passing through the input gate and the forget gate at time t.

[0033] The output gate determines the user electricity consumption information that the LSTM neural network model finally outputs and retains. It uses the Sigmoid function to weightedly add the input information at the current moment and the output information at the previous moment to obtain an initial output, and then scales the previously learned memory information through the tanh activation function, and then multiplies it with the initial output to finally obtain the output of the LSTM neural network model, that is, the electricity demand of the i-th target user at time t. The mathematical model of the output gate is as follows:

[0034]

[0035] in, is the state information retained at time t; and is the output gate network weight, Bias for the output gate.

[0036] The hidden layer uses the gradient calculated by the loss function to adjust the relevant parameters in the mathematical model through back propagation, that is, the Adam algorithm is used to update the model parameters, and the optimization is continuously iterated until the loss function reaches a convergence state. The number of hidden layers, the number of units in each layer, the maximum number of iterations, the initial learning rate, the number of iterations when the learning rate decreases, and the learning rate decrease factor in the LSTM model are optimized using the particle swarm algorithm. After the model training is completed, the output layer processes the results to obtain the electricity demand of users after the disaster.

[0037] By introducing LSTM to process the historical electricity consumption data of target users and using a gating mechanism to selectively retain the characteristics of the data during the transmission process, information can be maintained and updated for a long time, allowing the neural network to perform more accurate and differentiated electricity consumption characteristic analysis on target users and further clarify post-disaster electricity demand.

[0038] S130, input the electricity demand of all target users, the distribution network topology of the post-disaster city and the power equipment parameter set into the pre-built distribution network post-disaster recovery model, and solve it with the goal of strong maintenance timeliness to obtain the scheduling strategy; the distribution network post-disaster recovery model includes a power system operation unit, a topology reconstruction unit and a maintenance resource scheduling unit.

[0039] In a possible implementation, the maintenance resources of the post-disaster city include the maintenance team resources, the path resources and the traffic resources of the post-disaster city; the power equipment parameter set includes the equipment parameters of the normally operating power equipment and the equipment parameters of the faulty power equipment.

[0040] Optionally, the power equipment refers to equipment used in a distribution network or a microgrid for power supply, transmission or transformation, and does not include power equipment of target users of the power equipment.

[0041] Optionally, the power system operation unit in the distribution network post-disaster recovery model is mainly composed of distribution network operation constraints and microgrid operation constraints; the topology reconstruction unit is mainly obtained by representing the distribution network topology network model with radial characteristics using a virtual power balance constraint model; the maintenance resource scheduling unit can ensure that each damaged component can have a maintenance team responsible for maintenance, and the maintenance team that arrives at the node can leave on its own after the maintenance work is completed. All maintenance teams can depart from the maintenance site, and the time for the maintenance team to arrive at the damaged component is constrained to the repair status of the damaged component in each time period.

[0042] Optionally, the distribution network operation constraints in the power system operation unit can include distribution network node power balance constraints, power factor constraints that should be met when removing active and reactive loads, branch flow constraints that should be met by the power on the line and the voltage of each node, upper and lower limit constraints on the voltage of each node, and upper and lower limit constraints on the power output of the microgrid and the upper-level power grid.

[0043] The power balance constraint of the distribution network node is:

[0044] The power factor constraint that should be met when removing active and reactive loads is:

[0045] The branch power flow constraints that the power on the line and the voltage at each node should satisfy are:

[0046]

[0047] The upper and lower line constraints of each node voltage are:

[0048] The upper and lower limits of the power output of the upper power grid are:

[0049] in, For the upper power grid t Node of time i The active power output at For the upper power grid t Time Node i The reactive power output at For microgrids t Node of time i The active power output at For microgrids t Time Node i The reactive power output at for t Time Node i Active load removed at for t Time Node i The reactive load removed at For Node i The set of connected downlinks, For Node i The set of connected uplink lines, For target users t Time Node i The electricity demand at For target users t Time Node i The reactive load at For t Time Node i The power factor of the load, For branch l exist t Active power at the moment, For branch l exist t Reactive power at the moment, For branch l exist t The maximum apparent power at the moment, For Node i exist t The voltage value at the moment, For node j in t The voltage value at the moment, For branch l The resistance, For branch l Reactance, is the system reference voltage value, M is a large number related to the reference value of the system node voltage, and the line l exist t When there is no failure ,line l exist t When there is a fault at any time , is the upper limit constraint of power output of the upper grid and microgrid, is the lower limit constraint of power output of the upper grid and microgrid, For the upper grid and microgrid t Time Node i The active or reactive power output at the

[0050] The microgrid operation constraints in the power system operation unit include the upper and lower limit constraints of the distributed generator output power, the upper and lower limit constraints of the energy storage system charging and discharging power, the total balance constraints after load transfer within the microgrid, and the power balance constraints of each line in the microgrid.

[0051] The upper and lower limits of the distributed generator output power are:

[0052] The upper and lower limits of the charging and discharging power of the energy storage system are:

[0053]

[0054]

[0055] The total balance constraint after load transfer in the microgrid is:

[0056]

[0057]

[0058]

[0059]

[0060] The power balance constraint of each line in the microgrid is:

[0061] in, For distributed generators i exist t The output power at the moment, is the upper limit of the output power of distributed generators, is the lower limit of the output power of distributed generators, For energy storage system j exist t Charging power at the moment, For energy storage system j exist t The discharge power at the moment, is the lower limit of the charging power of the energy storage system, is the upper limit of the charging power of the energy storage system, is the lower limit of the discharge power of the energy storage system, is the upper limit of the discharge power of the energy storage system, For energy storage system j exist t The charging power coefficient at the time, For energy storage system j exist t The discharge power coefficient at the moment, For energy storage system j The lower limit of the state of charge, For energy storage system j The upper limit of the state of charge, is the discharge efficiency coefficient, is the charging efficiency coefficient, For energy storage system j The rated energy value, is the time step, and N PCC For contact line lIn time t The input power and the collection, N cDG is the collection of distributed generators, N ESS is the collection of energy storage systems, N DR is the set of controllable loads participating in demand response. For controllable load m exist t The power of the moment, In order to control the load transfer amount within one load cycle, For controllable load at time t The lower limit of load transfer, For controllable load at time t The upper limit of load transfer, and For controllable load m exist t Auxiliary variables added at all times, For controllable load m exist t Scheduling cost at the time.

[0062] The topology reconstruction unit in the distribution network post-disaster recovery model is:

[0063]

[0064]

[0065] in, For power transmission lines l At the moment t The virtual power flow, is the number of power system nodes, To ensure the 0 / 1 variable of the radial structure of the distribution network, Indicates i Is the potential root node in t The moment is selected as the root node, the potential root node set R The two end nodes of the damaged line of the distribution network, For Node i The set of connected downlink / uplink lines, ε It is a collection of distribution network lines.

[0066] In a possible implementation, the maintenance resource scheduling unit includes:

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075] in, For maintenance sites and the collection of damaged components, In c For Site The maintenance team under its jurisdiction assembled, A 0 / 1 variable used to represent the maintenance team c Whether path selection has been performed, Repair teams from damaged components d Travel to damaged component s , The repair team did not remove the damaged components. d Travel to damaged component s , Whether the repair team can access the damaged components d If the repair team passes by the damaged component d , then the value is 1 if the repair team does not pass the damaged component d , then the value is 0, Reach damaged components for repair teams s The time of the location, Reach damaged components for repair teams d The time of the location, For maintenance team c Repair damaged components d The time required, Repair teams from damaged components d Location to damaged component s The driving time required to reach your location, is a 0 / 1 variable used to represent damaged components d Is it in t The moment is repaired, if t Moment, damaged components d is being repaired, then ,liket Moment, damaged components d If the repair work has not started or has been completed, , ε is a decimal number related to the maintenance time error accuracy allowed by the maintenance team scheduling model. To t The number of damaged components that have been repaired so far, is a 0 / 1 variable used to represent damaged components d Is it in k The moment is repaired, if k Moment, damaged components d is being repaired, then =1, if k Moment, damaged components d If the repair work has not started or has been completed, =0.

[0076] In a possible implementation, the electricity demand of all target users, the distribution network topology of the city after the disaster, and the power equipment parameter set are input into a pre-built distribution network post-disaster recovery model, and the solution is performed with the goal of strong maintenance timeliness to obtain a scheduling strategy, including: the electricity demand of all target users, the distribution network topology of the city after the disaster, and the power equipment parameter set are input into the distribution network post-disaster recovery model to obtain a mixed integer linear programming model, and the objective function and related constraints of the mixed integer linear programming model are constructed with the goals of small post-disaster load shedding losses, low microgrid operating costs, and strong maintenance timeliness; the objective function and related constraints are converted into a compact form, and based on the Benders algorithm, the mixed integer linear programming model is decomposed into a main model and a sub-model; the main model and the sub-model are solved to obtain a scheduling strategy.

[0077] In one possible implementation, the objective function is:

[0078]

[0079] in, i For the distribution network i nodes, t For the moment, j For the j Energy storage system, l It is the connection line between the distribution network and the microgrid. is the unit coefficient of load shedding loss, is the operating cost of distributed power generation, is the operating cost of the energy storage system, is the demand response cost, The cost of purchasing and selling electricity, is the load importance, which is related to the electricity demand. is the weight coefficient of the entire maintenance process, is the set of power system nodes, Gather for the maintenance station, is the set of damaged components, is the set of distributed generators, A collection of energy storage devices. is the load set participating in demand response, is the collection of connecting lines between the distribution network and the microgrid, is the distributed generation operation cost coefficient, and are the coefficients of the first-order term and constant term in the conversion expression of the distributed generation output power, is the energy storage system operation cost coefficient, is the demand response cost, for t The demand response power reduced by load shedding at any time, is the cost coefficient of the distribution network purchasing electricity from the upper grid, is the cost coefficient of the distribution network selling electricity to users, The amount of electricity purchased by the distribution network from the upper grid. The amount of electricity sold by the distribution network to users.

[0080] The objective function is converted into a compact form:

[0081] The relevant constraints are converted into a compact form:

[0082]

[0083] in, and are discrete and continuous decision variables, i.e., the dispatching instructions for various post-disaster recovery resources of urban distribution networks, microgrids, and maintenance teams. and is the vector form of the dual variables associated with the continuous and discrete decision variable equality and inequality constraints, and is the coefficient vector of the discrete variable vector and the continuous variable vector, A , D is the coefficient matrix of the equality or inequality constraints involving 0 / 1 binary variables in compact form, and is the coefficient vector corresponding to the equality or inequality constraints of discrete variables, E , G and H is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints satisfied by both continuous and discrete variables.

[0084] In one possible implementation, the main model and the sub-model are solved to obtain a scheduling strategy, including: converting the sub-model into a dual expression and related constraints of the dual expression; performing iterative solving based on the main model, the dual expression and related constraints of the dual expression to obtain an optimal scheduling strategy.

[0085] Optionally, when constructing the objective function, each objective, such as small post-disaster load shedding losses, low microgrid operating costs, and strong maintenance timeliness, will correspond to a sub-model, and at the same time, the sub-models corresponding to each objective are similar.

[0086] Optionally, the objective function of the sub-model includes dual variables associated with relevant constraints in the urban distribution network post-disaster recovery model and discrete variables of decision variables of the urban distribution network post-disaster recovery model.

[0087] In one possible implementation, the sub-model is:

[0088]

[0089]

[0090] in, is the discrete variable decision result of the urban distribution network post-disaster recovery model transferred from the main problem, t For the moment, and is the vector form of the dual variables associated with the continuous and discrete decision variable equality and inequality constraints, q is the optimal objective function value, E , G , H is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables.

[0091] Among them, the objective function of the sub-model at any time is:

[0092] It should be noted that the corresponding constraints of the sub-model are the constraints that the dual variables associated with the corresponding constraints in the distribution network post-disaster recovery model should satisfy, and the mathematical relationship of the dual expression needs to be satisfied. The relevant constraints of the sub-model are as follows:

[0093] Optionally, the objective function of the main model is the optimal objective function value of the discrete variables and sub-problems in the decision variables of the distribution network post-disaster recovery model q composition.

[0094] In one possible implementation, the main model is:

[0095]

[0096]

[0097] in, t For the moment, k is the number of iterations, and For the submodel at time t and k The vector form of the dual variable of the post-disaster recovery resource scheduling decision variable at the iteration, q is the optimal objective function value, Provide dispatch instructions for various post-disaster recovery resources for urban distribution networks, microgrids, and maintenance teams. is the coefficient vector of the discrete variable vector, A , D is the coefficient matrix of the equality or inequality constraints involving 0 / 1 binary variables in compact form, and is the coefficient vector corresponding to the equality or inequality constraints of discrete variables, G is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints satisfied by both continuous and discrete variables.

[0098] Among them, the objective function of the main model is:

[0099] Among them, the first constraint of the main model is:

[0100] The second constraint of the main model is:

[0101] The third constraint of the main model is:

[0102] See also Figure 4 , the solution process is as follows: First, solve the objective function of the relaxed main model. At this time, there is no need to add the last two constraints of the main model. A set of solutions is obtained ( , ).

[0103] Secondly, substitute this set of solutions into the sub-model and calculate and The value of and Determine whether it is an unbounded solution or a bounded solution. and If the solution is unbounded, that is, the dual variable associated with the corresponding constraint in the post-disaster recovery model of the distribution network cannot be obtained, then it is necessary to add the third constraint of the main model to the main model, that is, to update the lower bound of the range of the post-disaster recovery model of the distribution network, and recalculate the objective function of the relaxed main model; like and is a bounded solution, that is, , at this time, we can get the extreme point of the dual subproblem, that is, the candidate optimal solution.

[0104] Finally, compare and , if the two are equal, terminate the loop and output the solution at this time; if they are not equal, it is necessary to add the optimal cut constraint, which is the second constraint of the main model, that is, to update the upper bound of the distribution network post-disaster recovery model range, and recalculate the objective function of the main model, and iterate until the upper and lower bounds of the distribution network post-disaster recovery model are constantly approaching and the optimal solution is obtained.

[0105] It should be noted that the optimal solution refers to the optimal scheduling strategy for various resources such as urban distribution networks, microgrids, and maintenance teams.

[0106] Since the number of bounded and unbounded solutions generated by the sub-model in the process of solving the distribution network post-disaster recovery model is limited, and new feasibility constraints and optimality constraints are generated in each iteration, and the constraints in the main model of the distribution network post-disaster recovery model are updated, therefore, after a limited number of iterations, the solution obtained will converge to the optimal solution, that is to say, the optimal scheduling strategy for various resources such as urban distribution networks, microgrids, and maintenance teams will be solved.

[0107] By constructing a mixed integer linear programming model for distribution network topology optimization, and taking small post-disaster load shedding losses, low microgrid operating costs, and strong maintenance timeliness as the goals, the objective function and related constraints of the mixed integer linear programming model for distribution network topology optimization can be constructed. Rapid reconstruction of the distribution network after a disaster can be achieved, and the scheduling plan can be quickly solved through the efficient Benders decomposition algorithm. The optimization problem containing a large number of decision variables is decomposed into a main model and sub-models, which helps to quickly and accurately find the optimal solution of the model, that is, the optimal scheduling strategy.

[0108] S140, dispatching power resources and maintenance resources of the post-disaster city based on the dispatching strategy.

[0109] Optionally, after obtaining the optimal scheduling strategy, Figure 2 The decision center shown in the figure feeds back the optimal dispatching strategy to the data sharing center. Each department obtains the corresponding department dispatching strategy through the data sharing center and performs post-disaster recovery of the urban power grid according to the dispatching strategy. The urban distribution network dispatches its own resources to restore the distribution network load according to the instructions corresponding to the optimal dispatching strategy; the microgrid dispatches microgrid resources to assist the distribution network in load recovery according to the instructions corresponding to the optimal dispatching strategy; the maintenance team repairs the line according to the path-related strategy in the optimal dispatching strategy, and uploads the repair information to the data sharing center in a timely manner, providing a reference for the overall power supply recovery of the urban distribution network.

[0110] The electricity demand of the corresponding target user is obtained through the historical load data set of each target user, and a distribution network post-disaster recovery model including a power system operation unit, a topology reconstruction unit, and a maintenance resource scheduling unit is pre-constructed. The electricity demand of all target users, the distribution network topology structure of the post-disaster city, and all parameters in the power equipment parameter set are input into the pre-constructed distribution network post-disaster recovery model to obtain a scheduling strategy, and the power resources and maintenance resources of the post-disaster city are scheduled according to the scheduling strategy. The electricity demand of the target user can be clearly identified. By determining the target user, the electricity demand of the target user is considered, and the maintenance timeliness, small post-disaster load shedding loss, and low microgrid operation cost are taken as the goals. The scheduling plan is quickly solved using the efficient Benders decomposition algorithm, which can effectively speed up the recovery of the post-disaster urban distribution network and reduce economic losses and maintenance costs.

[0111] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0112] Corresponding to the urban power grid load restoration method based on user demand in the above embodiment, Figure 5A schematic structural diagram of a user demand-based urban power grid load recovery device provided by an embodiment of the present invention is shown. For ease of explanation, only the parts related to the embodiment of the present invention are shown.

[0113] See also Figure 5 , the urban power grid load recovery device 5 based on user demand in the embodiment of the present invention may include: An acquisition module 51 is used to acquire a historical power load data set of each target user in the post-disaster city, as well as maintenance resources, distribution network topology and power equipment parameter set in the post-disaster city; An analysis module 52 is used to analyze the power consumption of each target user based on the historical load data set of the target user to obtain the power consumption demand of the target user; The solution module 53 is used to input the power demand of all target users, the distribution network topology of the post-disaster city and the power equipment parameter set into the pre-built distribution network post-disaster recovery model, and solve it with the goal of strong maintenance timeliness to obtain a scheduling strategy; the distribution network post-disaster recovery model includes a power system operation unit, a topology reconstruction unit and a maintenance resource scheduling unit; The scheduling module 54 is used to schedule the power resources and maintenance resources of the city after the disaster based on the scheduling strategy.

[0114] In one possible implementation, the analysis module 52 is specifically used to: for each target user, perform the following steps: obtain the power influencing factors of the post-disaster city, and construct the feature vector of the target user based on the power influencing factors and the historical load data set of the target user; input the feature vector of the target user into a pre-built LSTM network model to obtain the power demand of the target user.

[0115] In a possible implementation, the maintenance resources of the post-disaster city include the maintenance team resources, the path resources and the traffic resources of the post-disaster city; the power equipment parameter set includes the equipment parameters of the normally operating power equipment and the equipment parameters of the faulty power equipment.

[0116] In a possible implementation, the maintenance resource scheduling unit includes:

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126] in, For maintenance sites and the collection of damaged components, In c For Site The maintenance team under its jurisdiction assembled, A 0 / 1 variable used to represent the maintenance team c Whether path selection has been performed, Repair teams from damaged components d Travel to damaged component s , The repair team did not remove the damaged components. d Travel to damaged component s , Whether the repair team can access the damaged components d If the repair team passes by the damaged component d , then the value is 1 if the repair team does not pass the damaged component d , then the value is 0, Reach damaged components for repair teams s The time of the location, Reach damaged components for repair teams d The time of the location, For maintenance team c Repair damaged components d The time required, Repair teams from damaged components d Location to damaged component s The driving time required to reach your location, is a 0 / 1 variable used to represent damaged components d Is it in t The moment is repaired, if t Moment, damaged components d is being repaired, then ,like t Moment, damaged components d If the repair work has not started or has been completed, , εis a decimal number related to the maintenance time error accuracy allowed by the maintenance team scheduling model. To t The number of damaged components that have been repaired so far, is a 0 / 1 variable used to represent damaged components d Is it in k The moment is repaired, if k Moment, damaged components d is being repaired, then =1, if k Moment, damaged components d If the repair work has not started or has been completed, =0.

[0127] In one possible implementation, the solution module 52 is specifically used to: input the electricity demand of all target users, the distribution network topology of the city after the disaster, and the power equipment parameter set into the distribution network post-disaster recovery model to obtain a mixed integer linear programming model, and construct the objective function and related constraints of the mixed integer linear programming model with the goals of small post-disaster load shedding losses, low microgrid operating costs, and strong maintenance timeliness; convert the objective function and related constraints into a compact form, and based on the Benders algorithm, decompose the mixed integer linear programming model into a main model and a sub-model; solve the main model and the sub-model to obtain a scheduling strategy.

[0128] In one possible implementation, the objective function is:

[0129]

[0130] in, i For the distribution network i nodes, t For the moment, j For the j Energy storage system, l It is the connection line between the distribution network and the microgrid. is the unit coefficient of load shedding loss, is the operating cost of distributed power generation, is the operating cost of the energy storage system, is the demand response cost, The cost of purchasing and selling electricity, is the load importance, which is related to the electricity demand. is the weight coefficient of the entire maintenance process, is the set of power system nodes, Gather for the maintenance station, is the set of damaged components, is the set of distributed generators, A collection of energy storage devices. is the load set participating in demand response, is the collection of connecting lines between the distribution network and the microgrid, is the distributed generation operation cost coefficient, and are the coefficients of the first-order term and constant term in the conversion expression of the distributed generation output power, is the energy storage system operation cost coefficient, is the demand response cost, for t The demand response power reduced by load shedding at any time, is the cost coefficient of the distribution network purchasing electricity from the upper grid, is the cost coefficient of the distribution network selling electricity to users, The amount of electricity purchased by the distribution network from the upper grid. The amount of electricity sold by the distribution network to users.

[0131] In a possible implementation, the solving module 52 is further used to: convert the sub-model into a dual expression and related constraints of the dual expression; perform iterative solving based on the main model, the dual expression and related constraints of the dual expression to obtain the optimal scheduling strategy.

[0132] In one possible implementation, the sub-model is:

[0133]

[0134]

[0135] in, is the discrete variable decision result of the urban distribution network post-disaster recovery model transferred from the main problem, t For the moment, and is the vector form of the dual variables associated with the continuous and discrete decision variable equality and inequality constraints, q is the optimal objective function value, E , G , H is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables.

[0136] In one possible implementation, the main model is:

[0137]

[0138]

[0139] in, t For the moment, k is the number of iterations, and For the submodel at time t and k The vector form of the dual variable of the post-disaster recovery resource scheduling decision variable at the iteration, q is the optimal objective function value, Provide dispatch instructions for various post-disaster recovery resources for urban distribution networks, microgrids, and maintenance teams. is the coefficient vector of the discrete variable vector, A , D is the coefficient matrix of the equality or inequality constraints involving 0 / 1 binary variables in compact form, and is the coefficient vector corresponding to the equality or inequality constraints of discrete variables, G is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints satisfied by both continuous and discrete variables.

[0140] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0141] Those of ordinary skill in the art will appreciate that the templates, units, and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0142] If the module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned embodiments of the urban power grid load recovery method based on user needs can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal, telecommunication signal and software distribution medium.

[0143] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for urban power grid load restoration based on user demand, characterized in that: include: Obtaining a historical power load data set of each target user in the post-disaster city, as well as a maintenance resource, a distribution network topology, and a power equipment parameter set in the post-disaster city; For each target user, based on the historical load data set of the target user, the power consumption of the target user is analyzed to obtain the power demand of the target user; The power demand of all target users, the distribution network topology of the post-disaster city and the power equipment parameter set are input into a pre-built distribution network post-disaster recovery model, and the solution is performed with the goal of strong maintenance timeliness to obtain a scheduling strategy; the distribution network post-disaster recovery model includes a power system operation unit, a topology reconstruction unit and a maintenance resource scheduling unit; The power resources and maintenance resources of the post-disaster city are dispatched based on the dispatching strategy.

2. The urban power grid load restoration method based on user demand according to claim 1, characterized in that: The power demand of all target users, the distribution network topology of the post-disaster city, and the power equipment parameter set are input into the pre-built distribution network post-disaster recovery model, and the solution is performed with the goal of strong maintenance timeliness to obtain a scheduling strategy, including: The power demand of all target users, the distribution network topology of the post-disaster city, and the power equipment parameter set are input into the distribution network post-disaster recovery model to obtain a mixed integer linear programming model, and the objective function and related constraints of the mixed integer linear programming model are constructed with the goal of small post-disaster load shedding loss, low microgrid operation cost, and strong maintenance timeliness; Convert the objective function and the related constraints into a compact form, and decompose the mixed integer linear programming model into a main model and sub-models based on the Benders algorithm; The main model and the sub-model are solved to obtain a scheduling strategy.

3. The urban power grid load restoration method based on user demand according to claim 2, characterized in that: The objective function is: in, i For the distribution network i nodes, t For the moment, j For the j Energy storage system, l It is the connection line between the distribution network and the microgrid. is the unit coefficient of load shedding loss, is the operating cost of distributed power generation, is the operating cost of the energy storage system, is the demand response cost, The cost of purchasing and selling electricity, is the load importance, which is related to the power demand. is the weight coefficient of the entire maintenance process, is the set of power system nodes, Gather for the maintenance station, A collection of damaged components is the set of distributed generators, A collection of energy storage devices. is the load set participating in demand response, is the collection of connecting lines between the distribution network and the microgrid, is the distributed generation operation cost coefficient, and are the coefficients of the first-order term and constant term in the conversion expression of the distributed generation output power, is the energy storage system operation cost coefficient, is the demand response cost, for t The demand response power reduced by load shedding at any time, is the cost coefficient of the distribution network purchasing electricity from the upper grid, is the cost coefficient of the distribution network selling electricity to users, The amount of electricity purchased by the distribution network from the upper grid. The amount of electricity sold by the distribution network to users.

4. The urban power grid load restoration method based on user demand according to claim 2, characterized in that: The main model and the sub-model are solved to obtain a scheduling strategy, including: Converting the sub-model into a dual expression and related constraints of the dual expression; Based on the main model, the dual expression and related constraints of the dual expression, an iterative solution is performed to obtain an optimal scheduling strategy.

5. The urban power grid load restoration method based on user demand according to claim 1, characterized in that: The maintenance resource scheduling unit comprises: in, For maintenance sites and the collection of damaged components, In c For Site The maintenance team under its jurisdiction assembled, A 0 / 1 variable used to represent the maintenance team c Whether path selection has been made, Repair teams from damaged components d Travel to damaged component s , The repair team did not remove the damaged components. d Travel to damaged component s , Whether the repair team can access the damaged components d If the repair team passes the damaged component d , then the value is 1 if the repair team does not pass by the damaged component d , then the value is 0, Reach damaged components for repair teams s The time of the location, Reach damaged components for repair teams d The time of the location, For maintenance team c Repair damaged components d The time required, Repair teams from damaged components d Location to damaged component s The driving time required to reach your location, is a 0 / 1 variable used to represent damaged components d Is it in t The moment is repaired, if t Moment, damaged components d is being repaired, then ,like t Moment, damaged components d If the repair work has not started or has been completed, , ε is a decimal number related to the maintenance time error accuracy allowed by the maintenance team scheduling model. To t The number of damaged components that have been repaired so far, is a 0 / 1 variable used to represent damaged components d Is it in k The moment is repaired, if k Moment, damaged components d is being repaired, then =1, if k Moment, damaged components d If the repair work has not started or has been completed, =0.

6. The urban power grid load restoration method based on user demand according to claim 1, characterized in that: For each target user, based on the historical load data set of the target user, the target user is analyzed for power consumption to obtain the power consumption demand of the target user, including: For each target user, perform the following steps: Obtaining the power influencing factors of the post-disaster city, and constructing a feature vector of the target user based on the power influencing factors and the historical load data set of the target user; The feature vector of the target user is input into the pre-built LSTM network model to obtain the electricity demand of the target user.

7. The urban power grid load restoration method based on user demand according to claim 1, characterized in that: The maintenance resources of the post-disaster city include maintenance team resources of the post-disaster city, path resources of the post-disaster city and traffic resources of the post-disaster city; The electric power equipment parameter set includes equipment parameters of normally operating electric power equipment and equipment parameters of faulty electric power equipment.

8. The urban power grid load restoration method based on user demand according to claim 2, characterized in that: The sub-models are: in, is the discrete variable decision result of the urban distribution network post-disaster recovery model transferred from the main problem, t For the moment, and is the vector form of the dual variables associated with the continuous and discrete decision variable equality and inequality constraints, q is the optimal objective function value, E , G , H is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables.

9. The urban power grid load restoration method based on user demand according to claim 2, characterized in that: The main model is: in, t For the moment, k is the number of iterations, and For the submodel at time t and k The vector form of the dual variable of the decision variable for post-disaster recovery resource scheduling at the iteration, q is the optimal objective function value, Provide dispatch instructions for various post-disaster recovery resources for urban distribution networks, microgrids, and maintenance teams. is the coefficient vector of the discrete variable vector, A , D is the coefficient matrix of the equality or inequality constraints involving 0 / 1 binary variables in compact form, and is the coefficient vector corresponding to the equality or inequality constraints of discrete variables, G is the coefficient matrix containing the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints on continuous variables, is the coefficient vector corresponding to the equality or inequality constraints satisfied by both continuous and discrete variables.

10. A load recovery device for a city power grid based on user needs, characterized in that: include: An acquisition module, used to acquire a historical power load data set of each target user in a post-disaster city, as well as maintenance resources, distribution network topology, and power equipment parameter sets of the post-disaster city; An analysis module is used to analyze the power consumption of each target user based on the historical load data set of the target user to obtain the power consumption demand of the target user; A solution module is used to input the power demand of all target users, the distribution network topology of the post-disaster city and the power equipment parameter set into a pre-built distribution network post-disaster recovery model, and solve it with the goal of strong maintenance timeliness to obtain a scheduling strategy; the distribution network post-disaster recovery model includes a power system operation unit, a topology reconstruction unit and a maintenance resource scheduling unit; The scheduling module is used to schedule the power resources and maintenance resources of the post-disaster city based on the scheduling strategy.

Images