Distributed resource cooperative control method based on layering and partitioning autonomy of power distribution network

Abstract

The invention belongs to the technical field of distributed resource control of a power distribution network, and particularly relates to a distributed resource cooperative control method based on layering and partitioning autonomy of the power distribution network. The method comprises the following steps: S100, proposing an active power grid layering and zoning control strategy oriented to feeder autonomy; s200, establishing a power distribution network area power regulation and control information model of hierarchical and distributed control; and S300, realizing distributed resource cooperative control based on layering and partitioning autonomy of the power distribution network. According to the distributed resource cooperative control method provided by the invention, the distributed photovoltaic, user load and distributed energy storage coordinated control in the transformer area is realized by combining a feeder-autonomous-oriented active power distribution network layered and partitioned control strategy and utilizing the edge computing capability of the transformer area intelligent fusion terminal, the distributed photovoltaic absorption capability is improved, and the distributed resource utilization rate is improved. Energy supply and demand imbalance caused by distributed photovoltaic randomness is stabilized, and the economic operation level of a power distribution area is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of distributed resource control in distribution networks, and in particular relates to a distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution networks. Background Art

[0002] The new active distribution network contains a large number of distributed resources, which are distributed in various substations and feeders. The controllability of the single unit is poor. If the traditional centralized control method is adopted, not only the communication burden is heavy, but also it is difficult to achieve the expected control target. With the acceleration of the construction of new power systems, the penetration rate of distributed resources in the distribution network continues to increase, and problems such as distributed photovoltaic consumption gradually emerge. The distribution substation is the main level of small-capacity distributed photovoltaic access. The problem of distributed photovoltaic consumption is more prominent, and the distribution of distribution substations is unbalanced. The main reason is the unreasonable distribution and regulation of substations. It is mainly manifested in the problems of excessive reverse power and voltage deviation in the distribution substation. It is urgent to empower the distribution substation through digitalization and communication technology to improve the control ability of the distribution substation, and to realize flexible, inclusive and efficient control methods through information models and state perception to help distributed photovoltaic consumption. Therefore, whether it is possible to provide a distribution substation power control information model that can be combined with hierarchical distributed control, and use the edge computing capabilities of the intelligent fusion terminal in the substation to realize the coordinated control of distributed photovoltaic, user load and distributed energy storage in the substation is a technical problem to be solved by the present invention. Summary of the invention

[0003] In view of the shortcomings of the prior art, the present invention provides a distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution networks. To solve the above technical problems, the technical solution adopted by the present invention is: A distributed resource collaborative control method based on hierarchical partition autonomy of distribution network, comprising: S100, proposing a hierarchical partition control strategy for active power grid oriented to feeder autonomy; S200, establishing a power regulation information model of distribution network substations for hierarchical distributed control; S300, realizing distributed resource collaborative control based on hierarchical partition autonomy of distribution network; In step S100, the active power grid hierarchical partition control strategy includes: S110, establishing an equivalent model for the station layer based on the internal autonomy principle, and passing the equivalent result to the feeder layer; S120, the feeder layer divides the area based on the network structure and resource operation characteristics; S130, establishing a cross-layer and cross-region main distribution network collaborative optimization scheduling model; S140, using the ADMM algorithm to optimize the algorithm, and obtaining the hierarchical partition dynamic power interaction results and distributed resource processing results; In step S300, the distributed resource collaborative control method includes: S310, area data collection; S320, distributed photovoltaic output prediction; S330, area load prediction; S340, determining the SOC constraint conditions for distributed energy storage regulation; S350, regulating the power balance of the distribution station area.

[0004] Furthermore, in step S110, the process of establishing an equivalent model for the substation layer based on the principle of internal autonomy and passing the equivalent result to the feeder layer is as follows: S111, based on the feeder-substation layered collaborative control strategy, determine the power exchange conditions between the feeder layer and the substation layer; S112, determine the substation objective function; S113, determine the substation constraint conditions.

[0005] Furthermore, in step S120, the process of performing regional division at the feeder layer based on the network structure and resource operation characteristics includes: S121, determining an electrical distance index; S122, determining a regional division method.

[0006] Furthermore, in step S130, the process of establishing a cross-layer and cross-region main and distribution network collaborative optimization scheduling model includes: S131, determining the feeder layer objective function; S132, determining the feeder layer constraint conditions.

[0007] Furthermore, the process of step S140 includes: S141, based on ADMM distributed solution algorithm; S142, based on SPPA-ADMM distributed solution algorithm.

[0008] Furthermore, in step S200, a hierarchical distributed control architecture of the distribution network is first established, and the control architecture includes a distributed application APP group, a communication network and a physical facility architecture.

[0009] Further, in step S200, the method for establishing the power regulation information model of the distribution network substation is as follows: S210, simulating the modeling of the information model between the device and the master station, the interoperability of logical nodes between devices, and the interactive control and data sharing based on IEC 61850 / CIM at the distributed application APP layer, and using model-driven model conversion to generate source code; S220, using the MDE method between the IEC 61850 model and the function block to configure the communication network, and automatically configure, generate and deploy the functions of the logical nodes. The server model, logical device model, logical node model, data object model, data attribute model and public data class model in IEC 61850 are respectively mapped to the device model, resource model, composite function block model, basic function block model, data type and structured data type of the function block model.

[0010] Furthermore, in step S350, the power balance control process of the distribution station area includes: 1) collecting historical data and real-time operation data of distributed photovoltaic and station area loads; 2) predicting the power generation at the next time respectively. With load power ; 3) Calculate the power deviation between power generation and load , as the input for planning the optimal path for energy storage charging and discharging; 4) At the same time, based on the distributed energy storage operation data and prediction data, an optimal path for energy storage charging and discharging is formed; 5) Control commands are issued to control the charging and discharging of energy storage, and the remaining SOC energy of energy storage is predicted. The data is calculated by the DPMC logic node; 6) According to the imbalance of power in the substation and the operation status of energy storage, the optimal path for energy storage charging and discharging is adjusted until the power balance in the substation is optimal.

[0011] The present invention provides a distributed resource collaborative control method based on hierarchical partition autonomy of distribution network, combined with the hierarchical partition control strategy of active distribution network for feeder autonomy, and utilizing the edge computing capability of intelligent fusion terminal in the substation area to realize the coordinated control of distributed photovoltaic, user load and distributed energy storage in the substation area, improve the absorption capacity of distributed photovoltaic, smooth the imbalance of energy supply and demand caused by the randomness of distributed photovoltaic, and improve the economic operation level of distribution substation area. The beneficial effects achieved by the present invention are as follows: (1) A logical node information model of distributed photovoltaic, distributed energy storage and public nodes is established through the same communication standard, and the data definition of photovoltaic, energy storage and intelligent fusion terminals is unified, so that there is no misunderstanding between them to communicate, and the data is transmitted without error, so as to realize the flexible coordinated control of distribution network; (2) The power deviation of the imbalance between distributed photovoltaic output and load demand is smoothed by distributed energy storage, and the energy storage charging and discharging strategy is optimized to smooth the power fluctuation, improve the absorption capacity of distributed photovoltaic, and improve the stable, safe and economic operation capacity of distribution substation area.

[0012] The hierarchical and partitioned control strategy for active distribution network for feeder autonomy proposed in the present invention also has the following effects: (1) The proposed main transformer-feeder-substation hierarchical and partitioned coordinated control framework can make full use of multi-level and multi-regional adjustable resources to promote the consumption of new energy, improve the economy and low carbon of the entire system, and promote the economic, efficient and safe operation of the new distribution network. (2) A substation autonomous control optimization model with dynamic scene matching is proposed. The capacity of a single resource in the substation is small and scattered, which cannot meet the threshold of distribution system scheduling. In order to tap the potential of the substation to participate in grid scheduling, a substation equivalent model is proposed to achieve optimized autonomous operation under multiple objectives of the substation. Compared with the traditional autonomous control model, the proposed power model can comprehensively consider the operating characteristics of various distributed resources in the substation, network constraints and wind and solar consumption requirements, improve the utilization rate of distributed resources in the substation, and can achieve the effect of peak shaving and valley filling through the coordinated operation of photovoltaic storage. (3) The use of a distributed scheduling architecture fully taps the potential of low-carbon interactive support of distributed resources across voltage levels. The distributed coordinated control algorithm based on the ADMM algorithm used in it achieves the effects of internal autonomy within the region and coordinated control between regions, effectively solving the problem of low computing efficiency caused by centralized control. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 It is a schematic diagram of the power exchange relationship between the feeder layer and the substation layer of the present invention. Figure 2 It is a schematic diagram of the overall coordinated control architecture of the distribution network of the present invention. Figure 3 It is a flow chart of power balance control in the distribution station area of ​​the present invention. Figure 4 It is a schematic diagram of mapping the IEC61850 model and the functional blocks of the present invention. Figure 5 It is a schematic diagram of the logical nodes of the power control of the device of the present invention. Figure 6 It is a dataset diagram. DETAILED DESCRIPTION

[0014] In order to better understand the present invention, the following embodiments and drawings are used to further clarify the content of the present invention, but the protection content of the present invention is not limited to the following embodiments. Figure 1-6 As shown, the present invention mainly designs a distributed resource collaborative control method based on hierarchical partition autonomy of distribution network, including steps S100-S300. S100, proposes a hierarchical partition control strategy for active power grid oriented to feeder autonomy. S200, establishes a power regulation information model of distribution network substation for hierarchical distributed control. S300, realizes distributed resource collaborative control based on hierarchical partition autonomy of distribution network.

[0015] In step S100, the active power grid hierarchical partition control strategy includes: S110, establishing an equivalent model for the substation layer based on the principle of internal autonomy, and passing the equivalent result to the feeder layer. S120, the feeder layer divides the area based on the network structure and resource operation characteristics. S130, establishing a cross-layer and cross-region main distribution network collaborative optimization scheduling model. S140, using the ADMM algorithm to optimize the algorithm, and obtain the hierarchical partition dynamic power interaction results and distributed resource processing results.

[0016] In step S110 , the process of establishing an equivalent model for the station layer based on the internal autonomy principle and transmitting the equivalent result to the feeder layer includes S111 - S113 .

[0017] S111. Based on the feeder-substation hierarchical collaborative control strategy, the substation transmits the overall external active power and reactive power at each moment to the feeder layer based on the principle of autonomy. The active power and reactive power exchange relationship between the feeder layer and the substation layer is as follows: Figure 1 The feeder-station power exchange meets the following conditions: , Among them, M is the total number of stations; For the exchange of power variables on the feeder area side at the coupling node between the feeder layer and the substation area layer, is the power variable exchanged at the corresponding substation layer.

[0018] S112, determine the objective function of the substation area and optimize the resource allocation in the substation area. The substation layer resources are mainly distributed photovoltaic and energy storage, and the substation area aims to minimize the daily operating cost, which includes the penalty cost of abandoned light, the operation and maintenance cost of distributed energy storage, and the penalty cost of power interaction between the substation and the feeder layer. The objective function of the substation area is , , , .

[0019] In the formula, is the penalty cost coefficient for abandoned light, is the abandoned power of the lth photovoltaic in the area at time t, is the number of distributed photovoltaics in the area, and are the charging and discharging power of the sth distributed energy storage at time t, is the operation and maintenance cost of energy storage per unit power, is the amount of distributed energy storage within the substation, is the penalty cost coefficient for power interaction between the substation and the feeder side of the upper grid at time t, is the actual value of the interaction power between the station layer and the feeder layer at time t, and T is the time length of a typical scenario.

[0020] S113. Determine the area constraints, where the area constraints include power balance constraints, distributed energy storage constraints, and distributed photovoltaic output constraints.

[0021] The power balance constraints are: Where: is the actual active power output of distributed photovoltaic l in the area at time t; It is the total load demand within the substation at time t.

[0022] The distributed energy storage constraints are: .

[0023] Where: is the energy state of energy storage s at time t; , are the initial energy state and the end-of-cycle energy state of the energy storage s, respectively; Indicates the duration of energy storage charging and discharging; , are the charge and discharge efficiency respectively; , are the minimum and maximum values ​​of the state of charge, respectively; , They are the charge and discharge states of energy storage s at time t. In a time period, the energy storage can be in either the charge or discharge state, represented by 0 and 1; , are the maximum values ​​of charging power and discharging power of energy storage s respectively.

[0024] Distributed photovoltaic output constraints: , where: It is the maximum value of the distributed photovoltaic output in the area at time t.

[0025] In step S120, the process of performing regional division at the feeder layer based on the network structure and resource operation characteristics includes S121-S122.

[0026] S121, determine the electrical distance index, the electrical distance index includes electrical distance, active power and reactive power balance index, and comprehensive modularity partition index. The modularity function of the electrical distance index is defined as , , In the formula, is the modularity function; is a node set; is the weight of the edge connecting nodes i and j, is the sum of the weights of the edges connected to node i, is the sum of the weights of the edges connected to node j, is the sum of the network edge weights; when i and j are in the same partition, is 1, otherwise it is 0.

[0027] 1) Calculation method of electrical distance. Using voltage-reactive sensitivity Calculate the electrical distance, which reflects the degree of electrical coupling between network nodes. The voltage sensitivity of each node can be calculated using the Newton-Raphson method, and then the electrical distance between nodes can be calculated.

[0028] Where: represents the change in the amplitude of node voltage i; is the reactive power variation of node j; , They represent the influence of reactive power change of node j on voltage amplitude of node i and node j respectively; It reflects the ratio change of the degree of influence of node i voltage when the reactive power of node j changes; The larger the value, The smaller the value, the greater the impact of the reactive power change of node j on the voltage amplitude of node i.

[0029] In actual situations, the electrical coupling between two nodes is not only related to the two nodes themselves, but also to the power injected by other nodes in the network and the node position. If there are N nodes in the distribution network, then under the comprehensive influence of other nodes, the electrical distance between nodes i and j is for: ; Where: , Respectively represent nodes When the reactive power changes, the ratio of the affected voltages at node i and node j changes.

[0030] Considering the system electrical distance index, the value of the edge weight is determined, that is: .

[0031] 2) Calculation method of active / reactive power balance index. The reactive power balance index reflects the ability of reactive power local balance, as follows: Where: Indicates the number of feeder layer areas, For Region The net reactive power value at time t is, For Region The maximum reactive power provided at time t is: For Region The reactive power load demand value at time t.

[0032] The active power balance index is an index based on typical time-varying scenarios of the network, as follows: .

[0033] Where: is the net active power value of the region at time t; For Region The maximum active power provided at time t is: For Region Active power demand value at time t.

[0034] 3) Calculation method of comprehensive modularity zoning index. The comprehensive modularity zoning index is: Where: are the weight coefficients corresponding to different indicators.

[0035] S122. Determine the regional division method. Due to the time-varying nature of the source-load output, the regional division results will also change in real time. In this embodiment, the time scenario with the highest photovoltaic penetration rate in a day, i.e., 13:00, is selected for regional division. At this time, the photovoltaic penetration rate is 120%. The weights corresponding to the three indicators of modularity, active power balance, and reactive power balance can be selected according to the dispatching requirements of the dispatching center. In this example, the weights of the modularity, active power balance, and reactive power balance are 0.6, 0.2, and 0.2, respectively.

[0036] Genetic algorithm is used for regional division, with the maximum number of iterations set to 500, the population size set to 40, the crossover probability (0.3-0.6), and the mutation probability (0.1, 0.5). To ensure convergence, the two best individuals after each iteration are set not to perform crossover and mutation operations. Finally, the curve of the comprehensive regional division index changing with the number of iterations is obtained. After about 172 iterations, the comprehensive partition index reaches the maximum value of 0.869.

[0037] In step S130, the process of establishing a cross-layer and cross-region main and distribution network collaborative optimization scheduling model includes S131-S132.

[0038] S131. Determine the objective function of the feeder layer. Considering the safety and economy of the power system operation, the daily operating cost of each region is The minimum is the goal, and the decision variable is the output value of the adjustable resources in each area. Then the objective function of the feeder area is: , In the formula, For Region The penalty cost of abandoning new energy; For Region Energy storage operation and maintenance costs; For Region The main grid electricity purchase cost; For Region network loss cost.

[0039] 1) Penalty costs for abandoning new energy Where: For Region The number of PV plants inside; The unit power generation loss cost of the photovoltaic power station; is the predicted value of active output of the photovoltaic power station at node j at time t; is the actual active output value of the photovoltaic power station at node j at time t; is the abandoned power value of the PV power station at node j at time t.

[0040] 2) Energy storage power station operation cost—— Where: For Region The number of internal energy storage plants; is the operating cost of the unit energy storage power station; , is the charging power and discharging power of the energy storage station at node j at time t.

[0041] 3) Main grid electricity purchase cost—— Where: is the area at time t Active power purchased from the upper-level main grid; is the real-time electricity price of the power grid at time t.

[0042] 4) Network loss cost—— Where: Indicates area Number of internal nodes; The area at time t Total network loss; is the unit network loss cost; is the resistance value of the line between nodes i and j; is the current flowing between nodes i and j, are the active and reactive power flowing from upstream node i to node j respectively; is the voltage at node i.

[0043] S132. Determine feeder-layer constraints, where the feeder-layer constraints include Distflow power flow constraints, power balance constraints, energy storage power station charge state constraints, node voltage constraints, photovoltaic power station output constraints, and main grid-feeder interaction power upper and lower limit constraints.

[0044] 1) Distflow Constraints Where: , They respectively indicate that node i belongs to the first (last) node set of the branch with node j as the last (first) node; is the reactance value of the line between nodes i and j; , is the active and reactive power flowing from upstream node i to node j at time t; , is the active power and reactive power transferred between nodes j and k at time t; , They represent the net injected active power and reactive power of node j at time t respectively; is the current of branch ij at time t; , are the upstream and downstream node voltages at time t respectively.

[0045] Since the following equation is a nonlinear constraint, it needs to be converted into an inequality constraint according to the second-order cone relaxation principle: 2) Power balance constraints Where: , are the active power and reactive power of the load at node j at time t respectively; The power purchased by node j from the main grid at time t; is the reactive power output of the photovoltaic at node j at time t.

[0046] 3) Constraints on the state of charge of energy storage power stations Where: is the energy state of the energy storage power station at node j at time t; , is the charging and discharging efficiency of the energy storage station at node j; , are the initial energy state and the end-of-cycle energy state of the energy storage power station at node j, respectively; , are the maximum charging power and discharging power of the energy storage power station at node j respectively; , are the minimum and maximum state of charge at node j, respectively; , They are the charging and discharging states of the energy storage power station at node j at time t. In a time period, the energy storage power station is only in the charging or discharging state, which are represented by 0 and 1 respectively.

[0047] 4) Node voltage constraints Where: and are the minimum and maximum limits of the square of the voltage at node i, respectively. is the voltage amplitude at the substation outlet, is the rated voltage amplitude of the system.

[0048] 5) Output constraints of photovoltaic power stations Where: , It is the maximum value of active and reactive output of the PV power station at node j.

[0049] 6) Main grid-feeder interaction power upper and lower limit constraints Where: is the feeder area at time t Interaction power with the main network; , Feeder area Minimum and maximum values ​​of active power exchange with the main grid; , Feeder area Minimum and maximum values ​​for reactive power exchange with the mains.

[0050] In step S140, the ADMM algorithm is used to perform optimization calculations, and the process of finally obtaining the hierarchical partition dynamic power interaction results and the distributed resource output results includes steps S141-S142.

[0051] S141, based on ADMM distributed solution algorithm. In order to achieve friendly interaction between regions and improve the problem of low computational efficiency caused by centralized control, an inter-region distributed solution algorithm based on ADMM is proposed. For the partitioned region a, the distributed optimization model based on the partition coordination principle is , In the formula, is the objective function of region a, is the equality constraint for region a, represents the decision variable for region a; is an inequality constraint for region a. There exists an inequality between region a and its adjacent region b. Boundary equality constraints, used for convergence conditions of inter-regional distributed coordination optimization.

[0052] The steps of the interval distributed coordination optimization algorithm based on the ADMM algorithm are shown in steps 1-6 below.

[0053] Step 1: Initialization: set the initial values ​​of global variables according to the measured data of the distribution network, and set the initial values ​​of the Lagrange multipliers and voltage compensation parameters of all regional boundary data to 0.

[0054] Step 2: Establish the Lagrangian augmentation function corresponding to each regional sub-target. Take region a as an example. The region adjacent to region b is region a. There is a boundary equality constraint relationship between region a and region b. Then the corresponding equations of region a and region b are , as follows: Where: k is the number of iterations, is the penalty parameter, is the control variable inside region b; , is the autonomous objective function within regions a and b; , are the variables of interaction between region a and region b obtained by the k-th iterative optimization calculation within region a and region b respectively; , are the fixed reference values ​​of the k+1th iteration of region a and region b respectively; , are the Lagrange multipliers of regions a and b corresponding to the kth iteration respectively.

[0055] Step 3: Update the decision variables in each sub-region. Start the iteration from k=1. In the k+1th iteration, optimize and solve the Lagrangian augmented function in each region in parallel to obtain the decision variable value in the region with the corresponding iteration step. , , and the coupling variable value at the regional boundary can be obtained : Step 4: Update the boundary variables between each region based on step 3. Based on the boundary values ​​obtained in step 3 , the above formula can be used to calculate the average value of the coupling branch state and use it as the reference value for the next iteration.

[0056] Step 5: Update the Lagrange multiplier of each region, that is: In the formula, , are the Lagrange multipliers of regions a and b corresponding to the k+1th iteration respectively.

[0057] Step 6: Determine whether the algorithm has converged. The convergence criterion is the original residual and the dual residual Whether it tends to 0. The definition is as follows, where is the convergence accuracy. If the convergence condition is met, the algorithm ends, otherwise returns to step 2.

[0058] S142, based on the SPPA-ADMM distributed solution algorithm, the SPPA-ADMM solution algorithm adaptively updates the spectral penalty parameters based on the spectral step size obtained according to the gradient descent method. Unlike the adaptive ADMM distributed solution algorithm based on residual balance, the SPPA-ADMM distributed solution algorithm does not require manual setting of parameters for penalty parameter update, which improves the convergence speed and is effective for optimization economic dispatch problems of different scales. The solution steps are as follows: Steps 1-10.

[0059] Step 1: Initialization: set the initial values ​​of global variables according to the measured data of the distribution network, and set the initial values ​​of the Lagrange multipliers and voltage compensation parameters of all regional boundary data to 0.

[0060] Step 2: Establish the corresponding Lagrangian augmented function for each area divided by the feeder. Take area a as an example. The area adjacent to area a is area b. There is a boundary equality constraint relationship between area a and area b. Then the corresponding formula between area a and area b is , as follows: Where: k is the number of iterations, is the penalty parameter for the kth iteration, , are the control variables within region a and region b respectively; , is the autonomous objective function within regions a and b; , are the variables of interaction between region a and region b obtained by the k-th iterative optimization calculation within region a and region b respectively; , are the fixed reference values ​​of the kth iteration of region a and region b respectively; , are the Lagrange multipliers of regions a and b corresponding to the kth iteration respectively.

[0061] Step 3: Control variables Based on fixed reference value and Lagrange multipliers is updated inside region a. At the same time, region b receives the interaction variables between regions. , update the control variables inside region b , and obtain the coupling variable value at the region boundary.

[0062] .

[0063] Step 4: To obtain the spectral step size of the problem, an auxiliary variable of the Lagrange multiplier , update the auxiliary variables according to the following formula : .

[0064] Step 5: Update the Lagrange multiplier of each region, that is: In the formula, , are the Lagrange multipliers of regions a and b corresponding to the k+1th iteration respectively.

[0065] Step 6: The steepest descent and minimum gradient of the model in region a are and , the steepest descent and minimum gradient of the model in region b are and , calculated as follows: , , , .

[0066] Step 7: and The spectral step length is and , can be obtained by the mixed step rule as follows: .

[0067] Step 8: and express , and The special penalty parameter is updated only when the correlation is greater than the threshold. , .

[0068] Step 9: Finally, update the spectrum penalty parameter based on the spectrum step size as shown below: Step 10: Determine whether the algorithm has converged. The convergence criterion is the original residual and the dual residual Whether it tends to 0. Defined as follows, where is the convergence accuracy. If the convergence condition is met, the algorithm ends, otherwise returns to step 2. .

[0069] In step S200, the overall coordinated control architecture of the distribution network is as follows: Figure 2 As shown, a hierarchical distributed control architecture of the distribution network is first established, and the control architecture includes a distributed application APP group, a communication network and a physical facility architecture. The physical facility architecture is a distribution station area, and the physical facility architecture includes a low-voltage distribution network. The intelligent fusion terminal is arranged in the distributed application APP group; the low-voltage distribution network manages the operation control of the distribution station area through the intelligent fusion terminal, and the distribution station area is connected to distributed photovoltaics. The photovoltaic, load and energy storage data of the station area are gathered at the intelligent fusion terminal through the communication network for calculation and processing, so as to form a control strategy to manage the power balance of the station area.

[0070] In step S200, the method for establishing the power regulation information model of the distribution network area is: S210, simulation, based on IEC 61850 / CIM, modeling the information model between the device and the master station, the interoperability of logical nodes between devices, interactive control and data sharing at the distributed application APP layer, using model-driven model conversion to generate source code; configuring the communication network in the device supporting the coordinated power balance control, and automatically configuring, generating and deploying the functions of the logical nodes, using the model to define the peripheral information of the control mode and the information interaction method, to achieve interoperability and convenient device interaction; S220. Use the MDE method between the IEC 61850 model and the function block to configure the communication network, and automatically configure, generate and deploy the functions of the logical nodes. The server model, logical device model, logical node model, data object model, data attribute model and common data class model in IEC 61850 are respectively mapped to the device model, resource model, composite function block model, basic function block model, data type and structured data type of the function block model, that is, all the logical nodes of the distribution equipment and the information data flows between them are regarded as a resource model, and each logical node is the corresponding function block model.

[0071] IEC 61850 adopts object-oriented modeling ideas to establish data objects, namely logical nodes (logical nodes, LN), for application function entities (current protection, measurement values, etc.) and equipment (circuit breakers, transformers, etc.). Logical nodes are the basic components of IEDs. Each node contains data DATA (or data object Data Object-DO), and each data contains data attributes DataAttribute-DA. The data attribute value is the main information that needs to be paid attention to, such as switch position, set value, measurement value, control value, etc. The operation model of DER mainly includes: ①DER resources, which describe the basic electrical information of distributed energy, such as rated capacity, status, function, etc., inherited from the DERResourceLN logical node; ②Electrical connection point instance, which records the electrical connection points of DER to associate measurement information and "close the loop" energy management function, inherited from the ElectricalReferencePoint LN object; ③Operational function, which controls the specific behavior of the energy management function logical node to complete the relevant DER resources, inherited from OperationalFunctionLN; ④Energy management function logical node, which realizes the control and management of DER resources. The information model of power control in the distribution area involves logical nodes including DGEN (distributed energy operation), DECP (electrical connection characteristics), DPMC (power management control), MMXU (measurement node), ZINV (inverter function), MMET (meteorological monitoring), DSTO (storage electrical operation attributes), DBAT (battery system), DBTC (charging management), DMDR (multiple distributed energy combinations or microgrid hybrids), MMTR (metering), DVVR (voltage and reactive power regulation), DLOD (load), DPVC (photovoltaic array controller), etc. The logical nodes of equipment power control are as follows: Figure 5 shown.

[0072] The power control function of the distributed distribution substation is abstractly described through the information model, and the electrical quantity information and control process information are organized and formalized to form a data structure accepted by power workers, equipment manufacturers and users. This enables the query, reading, exchange, sharing and effective management and maintenance of data such as power control of distribution substations. Among them, the IEC 61850-7-420 standard not only details the logical nodes of distributed energy, but also specifies the interaction logic and transmission mode of the logical nodes. Based on Figure 2 The overall coordinated control architecture of the distribution network can use logical nodes to organize the various process steps of the control architecture, such as the data flow process of power control in the substation: MMXU→DGEN→DVVR→DPMC→prediction→optimized path→power allocation. Among them, DGEN and DSTO establish functional mapping for power equipment and carry the operation and control data of the equipment, and have interfaces for power equipment control data; MMXU, as a general logical node, carries the measurement and upload data of electrical equipment; DVVR, DBTC, DMDR and other functional algorithms carry specific electrical quantity regulation; DPMC carries specific prediction, optimization path and charge and discharge control algorithms. Combining the power regulation function with the logical node can build a control framework for the control algorithm implementation, and the process of the IEC 61850-7-420 standard logical node sorts out the interaction issues of device nodes, clarifies the flow and functionality of data, and guides the problems of algorithm results and process data being difficult to interact with power equipment, thereby improving the feasibility of the coordinated control algorithm.

[0073] Specifically, in this embodiment, in order to realize the control of photovoltaic load power balance in the distribution station area, the CDC (common data class), DataSetlmitate data set, DAType data attribute type, LNFB logical node function block, LNStructs logical node structure and other model libraries are defined based on IEC61850 / CIM. Among them, the DataSetlmitate data set (DATA-SET) is an ordered DATA DataAttributes ObjectReference group (called a data set member). For the convenience of the device side, the logical nodes of the application example are organized into a single set. Both the device side and the server know the DATA-SETObjectReference members and order, so only the DATA-SET name and the current value of the referenced DATA DataAttribute need to be transmitted. In order to realize the power regulation of the substation in the example, the data set defines the state output data set of the DGEN (distributed energy operation) logic node, the state output data set of the DMDR (multiple distributed energy combinations or microgrids) logic node, the reactive power control output data set of the DPMC (DER management and control) logic node, the state output data set of the DSTO (battery energy storage properties) logic node, the request data set of the DVVR (voltage reactive power regulation) logic node, and the basic data output data set of the DECP (electrical connection characteristics). Figure 6 shown.

[0074] Furthermore, the information model function blocks are formed into LNFB logical node function blocks, which provide a basis for data communication and interoperability within and between devices. The example constructs logical node function blocks such as DGEN, DMDR, DPMC, DSTO, DVVR, MMXU, DECP, DBTA, etc. At the same time, according to the communication requirements of power control, the Pub (publish) and Sub (subscribe) function blocks of the DDS communication protocol are constructed, including DDSPublisher_DGEN, DDSPublisher_DPMC, DDSPublisher_DSTO, DDSPublisher_DVVR, DDSSubscriber_DGEN, DDSSubscriber_DPMC, DDSSubscriber_DSTO, DDSSubscriber_DVVR and other function blocks.

[0075] According to the actual communication interaction mode between the electrical equipment of distributed photovoltaic, distributed energy storage, substation load and other objects, the example establishes a functional block to realize the functions of receiving, sending control data and uploading operation data, and carries the communication function between the local and remote centers. At the same time, the central node control algorithm functional block is constructed, the communication process is established, and the linear optimization iterative process is used to continuously find the optimal method for absorbing the distributed photovoltaic power within the time interval, forming a control strategy for regulating the charging and discharging of distributed energy storage, and sending the control target to the user-side functional block operation control to achieve substation power balance.

[0076] In step S300, the distributed resource collaborative control method includes steps S310-S350.

[0077] S310, area data collection, the intelligent fusion terminal collects distributed photovoltaic data, area load data and energy storage data. The above data include real-time data and historical data such as voltage, current, active power, reactive power, frequency of distributed photovoltaic and area load, as well as real-time data such as voltage, current, active power, reactive power, frequency of energy storage.

[0078] S320, distributed photovoltaic output prediction, first the intelligent fusion terminal requests the distributed photovoltaic electrical quantity data, and combines the weather data, the data of the nearest day, the data of similar days, and the historical power generation data to calculate the short-term prediction value of the target distributed photovoltaic output.

[0079] The steps for distributed photovoltaic output prediction are: ① Establish a historical database. According to the source, data structure, data volume and other related fields of distributed photovoltaic output data, comprehensively and systematically collect photovoltaic output data in the substation area as a historical database for data processing and establishment of prediction models. Historical data includes historical processing data and historical meteorological data.

[0080] ② Determine the influencing factors. Analyze the key factors that affect photovoltaic output, including solar radiation intensity, weather type, temperature, and humidity.

[0081] ③ Data preprocessing: Data preprocessing includes removing attribute data, identifying and correcting bad data, standardizing data, adding influencing factors and reorganizing data to improve data quality.

[0082] ④ Establish a short-term photovoltaic output prediction model. The modeling process of the photovoltaic output prediction model includes designing the model structure, setting the model parameters, and iterative training. The pre-processed data is input into the designed network, and the optimal short-term photovoltaic output prediction model is finally obtained through the training process of extracting features, iterative training, and correcting errors.

[0083] The photovoltaic output power is predicted based on the clearness index of photovoltaic irradiation. According to the ten-day historical data of photovoltaic output, the photovoltaic output data is statistically analyzed and the average clearness index at each time is taken into account to obtain the photovoltaic output prediction result within the current time.

[0084] In step S320, the distributed photovoltaic output power is predicted using the calculation formula , Where t is the current time, is the time step of prediction; To predict k time intervals added to the future; is the clearness index; is the predicted PV power; is the maximum value of the historical PV power data at the same time within the j-day period; and The calculation formula is: In the formula This application is , take another .about To calculate the clearness index, this embodiment uses the ratio of the measured photovoltaic power 3 hours before the prediction time to the maximum photovoltaic power at the same time 10 days ago.

[0085] ⑤ Photovoltaic output prediction. Photovoltaic output, real-time meteorological data, and historical load data are obtained as parameter inputs of the load prediction model, and the ultra-short-term photovoltaic real-time output prediction value and the short-term photovoltaic daily output prediction value are calculated through the prediction algorithm.

[0086] S330, substation load forecast, the intelligent fusion terminal requests the electrical quantity data of the substation low-voltage intelligent switch, and calculates the short-term forecast value of the substation load based on the historical data of the substation load.

[0087] The steps of load forecasting are: ① Establish a historical database. Analyze the relevant fields according to the source, data structure, and data volume of the required data, and comprehensively and systematically collect the load data of the substation area as a historical database for data processing and establishment of prediction models.

[0088] ② Determine the influencing factors. Analyze the key influencing factors of different substations based on their operating characteristics.

[0089] ③ Load data preprocessing: Data preprocessing includes removing attribute data, identifying and correcting bad data, data standardization, price adjustment influencing factors and data reorganization to improve data quality.

[0090] ④ Establish a short-term load forecasting model. The modeling process of the load forecasting model includes designing the model structure, setting the model parameters, and iterative training. The pre-processed data is input into the designed network, and the optimal short-term load forecasting model is finally obtained through the training process of extracting features, iterative training, and correcting errors.

[0091] The historical load data is divided into the comprehensive calculation results of the long-term time scale impact and the short-term time scale impact. According to the time interval with the predicted data, the weight of the long-term time scale and the short-term time scale in the prediction result is dynamically adjusted to obtain more reliable load forecast data and adjust the forecast in time according to the occurrence of load peaks and valleys, so as to respond quickly to load changes.

[0092] The calculation formula for the load forecast of the substation area is: The exponential smoothing factor is calculated using the formula: , Where t is the current time and ∆t is the predicted time step; To predict k time intervals added to the future; is the predicted load power, is the actual historical load power; Represents the weight of the actual load power at time t to the predicted value, Represents the impact of the actual load power at time t on the previous day and the actual load power at time t 7 days ago on the predicted value.

[0093] ⑤ Load forecasting. Obtain real-time load data and historical load data as parameter inputs for the load forecasting model, and calculate ultra-short-term real-time load forecast values ​​and short-term daily load forecast values ​​through the forecasting algorithm.

[0094] S340, determining the SOC constraint conditions for distributed energy storage control, adjusting the charging and discharging strategy of the energy storage according to the SOC constraint conditions, improving the load capacity of the energy storage system and optimizing the operation of the distribution network.

[0095] Distributed energy storage technology provides solutions for power systems with a high proportion of new energy. The problem of accommodating distributed power sources in the distribution network can be solved by exerting regional autonomy through the distributed energy storage cluster control strategy to increase the proportion of distributed power sources. The SOC of the energy storage device is an important factor affecting the participation of energy storage in regulation. The charging and discharging strategy of energy storage is adjusted according to the energy storage SOC to improve the load capacity of the energy storage system and optimize the operation of the distribution network.

[0096] Rated capacity of distributed energy storage The real-time state of charge of the energy storage is determined when the equipment is manufactured. As shown below: ; The following formula can also be used for calculation: Where: The remaining capacity of energy storage at the current moment; , are the energy changes of energy storage charging or discharging at time t-1-t respectively; Power absorbed for energy storage; The power released by the energy storage.

[0097] S350: Regulate the power balance of the distribution station area, use the edge computing capabilities of the intelligent fusion terminal to generate the energy storage charging and discharging optimization path to control the distributed energy storage charging and discharging, and then adjust the energy storage charging and discharging optimization path according to the energy storage predicted remaining energy and the actual remaining energy.

[0098] like Figure 3 As shown, in step S350, the power balance control process of the distribution station area includes: 1) Collect historical data and real-time operation data of distributed photovoltaic and regional loads. The above data is measured by the MMXU of each device through the DECP logical node. The DGEN and DLOD carry the real-time control interaction and historical operation data of the power equipment, and the DMDR integrates them into a single power equipment for DPMC processing; 2) Predict the power generation power at the next time respectively With load power , the data is collected by the DPMC logical node and the prediction data of the process is calculated; 3) Then the power deviation between power generation and load is calculated , as the input for planning the optimal path for energy storage charging and discharging, the DPMC logic node collects the predicted data of the previous process and calculates the intermediate process data; 4) At the same time, based on the distributed energy storage operation data and predicted data, the optimal path for energy storage charging and discharging is formed. The MMXU and DECP collect data, and the operation data of the energy storage logic node is formed by combining the DSTO and DBAT of the energy storage. The data flows to the DPMC for processing, and the control command is formed and sent back to the DSTO; 5) Control commands are issued to control the charging and discharging of energy storage, and the remaining SOC energy of energy storage is predicted. The data is calculated by the DPMC logic node; 6) According to the imbalance of power in the substation and the operation status of energy storage, the optimal path for energy storage charging and discharging is adjusted until the power balance in the substation is optimal. The DPMC issues control instructions to DBAT and DSTO for execution, and continuously feeds back the control results.

[0099] The control goal of the energy storage charging and discharging optimization path is to use energy storage to absorb as much distributed photovoltaic power as possible when the reverse power transmission limit is exceeded, and to allocate as little energy storage charging power as possible when the reverse power transmission limit is not exceeded, so as to balance the photovoltaic load power in the distribution station area. It is also necessary to ensure that the energy storage SOC is within a favorable range. Where: Charging power for the battery, is the battery discharge power, is the amount of abandoned light, is the reverse power, Supply power to the grid, is the energy storage capacity, It is the coefficient of each power in the optimization target. By adjusting the coefficient allocation, the tendency of the objective function for different power allocations can be changed.

[0100] The constraints of energy storage regulation are: in: The surplus power is the difference between the power generated by photovoltaic power and the power used by the load. When the surplus power is positive, it has three usage orientations: power returned to the grid, energy storage charging power and abandoned light; when the surplus power is negative, it has two usage orientations: energy storage discharge power and grid supply load power. The above formula shows the power allocation constraints under different situations.

[0101] in: For energy storage charging efficiency, is the energy storage discharge efficiency, and the above formula represents the constraints of energy storage charging and discharging.

[0102] Where: is the minimum energy storage capacity, is the maximum energy storage capacity, is the maximum discharge power of energy storage, is the maximum charging power of energy storage, is the discharge power allowed by the energy storage, To supply maximum grid power, is the maximum reverse power.

[0103] In order to ensure that the system complies with the set feed-in and / or load limits, the previously optimized charging / discharging power is continuously adjusted based on the difference between the last prediction and the current measurement. The specific power of energy storage charging and discharging can be determined based on the positive or negative surplus power.

[0104] In the formula is the predicted battery discharge power, is the predicted battery charging power, and Based on the predicted photovoltaic output and load power, the final energy storage charging and discharging power is controlled according to the above formula. It needs to be balanced with photovoltaic power generation, load power, and energy storage power to ensure the safe and stable operation of equipment and power grid.

[0105] In the energy storage charging and discharging optimization control, the power deviation value between the photovoltaic output and the load in the substation, as well as the voltage and harmonic deviation values ​​are calculated in combination with the control requirements of the main station, and combined with the charging and discharging optimization path of the distributed energy storage to generate the energy storage charging and discharging optimization control strategy, and the control instructions are sent to the distributed energy storage to control its charging and discharging.

[0106] Among them, the optimal charging and discharging path of distributed energy storage is: first, the predicted power is calculated through the historical data and real-time data of photovoltaic and load; secondly, the power deviation of photovoltaic and load is calculated through the prediction results; and then the optimal charging and discharging path of energy storage is generated by combining the return power limit of the substation and the energy storage operating conditions.

[0107] The constraints of energy storage regulation also include: in is the power deviation between photovoltaic power generation and load in the area: . It is the power absorbed or released by the power grid under the restriction of reverse power transmission, which needs to be balanced with photovoltaic power generation power, load power and energy storage power to ensure the safe and stable operation of equipment and power grid.

[0108] Energy storage charging and discharging optimization control strategy: The control strategy is generated by the intelligent integrated terminal. The control goal is to use energy storage to absorb as much distributed photovoltaic power as possible when the reverse power transmission limit is exceeded, and to allocate as little energy storage charging power as possible when the reverse power transmission limit is not exceeded, so that the photovoltaic load power in the distribution station area is balanced, and the energy storage SoC must be kept within a favorable range. The charging and discharging optimization path based on distributed energy storage controls the energy storage charging and discharging start and stop time and charging and discharging capacity to adjust the deviation value between photovoltaic and load power, evaluate whether the power in the station area is balanced based on the deviation value, and adjust the energy storage charging and discharging optimization path in combination with the remaining capacity of energy storage. The control strategy has a process of power prediction-deviation calculation-generation of energy storage charging and discharging optimization path-control charging and discharging-optimization path adjustment with an execution time step of 15 minutes.

[0109] Energy storage charging and discharging optimization control steps: ① Distributed photovoltaic output prediction. ② Substation load prediction. ③ Generate the power deviation value between photovoltaic output and load. ④ Distributed energy storage simulation, generate the energy storage charging and discharging optimization path. ⑤ Carry out substation power deviation control. Generate the energy storage charging and discharging optimization control strategy, that is, adjust the energy storage charging and discharging according to the power deviation value between photovoltaic output and load in the best way to balance the power deviation between photovoltaic output and load.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Other modifications or equivalent substitutions made to the technical solution of the present invention by ordinary technicians in the field should be included in the scope of the claims of the present invention as long as they do not depart from the spirit and scope of the technical solution of the present invention.

Claims

1. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network, characterized in that: include: S100, propose a hierarchical and partitioned control strategy for active power grids oriented to feeder autonomy; S200, establish a power regulation information model for distribution network substations with hierarchical distributed control; S300, realize the coordinated control of distributed resources based on hierarchical and partitioned autonomy of distribution networks; In step S100, the active power grid hierarchical partition control strategy includes: S110. Establish an equivalent model for the substation layer based on the principle of internal autonomy, and pass the equivalent results to the feeder layer; S120. The feeder layer divides the area based on the network structure and resource operation characteristics; S130. Establish a cross-layer and cross-area main distribution network collaborative optimization scheduling model; S140. Use the ADMM algorithm to optimize the algorithm and obtain the hierarchical and partitioned dynamic power interaction results and distributed resource processing results; In step S300, the distributed resource collaborative control method includes: S310, data collection of the substation, the intelligent fusion terminal collects distributed photovoltaic data, substation load data and energy storage data; S320, distributed photovoltaic output prediction, the intelligent fusion terminal first requests the electrical quantity data of distributed photovoltaic, and combines weather data, nearby day data, similar day data and historical power generation data to calculate the short-term prediction value of the target distributed photovoltaic output; S330, substation load prediction, the intelligent fusion terminal requests the electrical quantity data of the substation low-voltage intelligent switch, and calculates the short-term prediction value of the substation load based on the historical data of the substation load; S340, determine the SOC constraint conditions for distributed energy storage regulation, adjust the charging and discharging strategy of energy storage according to the SOC constraint conditions, improve the load capacity of the energy storage system and optimize the operation of the distribution network; S350, regulate the power balance of the distribution station, use the edge computing capability of the intelligent fusion terminal to generate the energy storage charging and discharging optimization path to control the distributed energy storage charging and discharging, and then adjust the energy storage charging and discharging optimization path according to the predicted remaining energy and the actual remaining energy of the energy storage.

2. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 1, characterized in that: In step S110, the process of establishing an equivalent model for the station layer based on the internal autonomy principle and transmitting the equivalent result to the feeder layer is as follows: S111, based on the feeder-station layer coordinated control strategy, determine the power exchange conditions of the feeder layer and the station layer, ， Among them, M is the total number of stations; For the exchange of power variables on the feeder area side at the coupling node between the feeder layer and the substation area layer, is the power variable exchanged at the corresponding substation layer; S112, determine the area objective function and optimize the resource allocation in the area. The area objective function is: In the formula, is the penalty cost coefficient for abandoned light, is the abandoned power of the lth photovoltaic in the area at time t, is the number of distributed photovoltaics in the area, and are the charging and discharging power of the sth distributed energy storage at time t, is the operation and maintenance cost of energy storage per unit power, is the amount of distributed energy storage within the substation area, is the penalty cost coefficient for power interaction between the substation and the feeder side of the upper grid at time t, is the actual value of the interaction power between the station layer and the feeder layer at time t, and T is the time length of the typical scenario; S113. Determine the area constraints, where the area constraints include power balance constraints, distributed energy storage constraints, and distributed photovoltaic output constraints.

3. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 2, characterized in that: In step S120, the process of dividing the feeder layer into regions based on the network structure and resource operation characteristics includes: S121. Determine an electrical distance index, wherein the electrical distance index includes an electrical distance, an active power and reactive power balance index, and a comprehensive modularity partition index: The modularity function of the electrical distance index is defined as ， ， In the formula, is the modularity function; is a node set; is the weight of the edge connecting nodes i and j, is the sum of the weights of the edges connected to node i, is the sum of the weights of the edges connected to node j, is the sum of the network edge weights; when i and j are in the same partition, is 1, otherwise it is 0; S122. Determine the regional division method, and select the time scenario with the highest photovoltaic penetration rate in a day for regional division; then determine the weights of the modularity index, active power balance index and reactive power balance index, and the weights of the three indicators are selected according to the dispatching needs of the dispatching center; use genetic algorithm to divide the region, set the maximum number of iterations, population size, crossover probability and mutation probability, where the two best individuals after each iteration do not perform crossover and mutation operations, and finally obtain the maximum value of the comprehensive zoning index based on the curve of the comprehensive regional division changing with the number of iterations.

4. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 3, characterized in that: In step S130, the process of establishing a cross-layer and cross-region main and distribution network collaborative optimization scheduling model includes: S131, determine the feeder layer objective function, based on the daily operating cost The minimum is the goal, and the decision variable is the output value of the adjustable resources in each area. Then the objective function of the feeder area is ， In the formula, For Region Penalty costs for abandoning new energy sources; For Region Energy storage operation and maintenance costs; For Region The main grid electricity purchase cost; For Region The network loss cost; S132. Determine feeder-layer constraints, where the feeder-layer constraints include Distflow power flow constraints, power balance constraints, energy storage power station charge state constraints, node voltage constraints, photovoltaic power station output constraints, and main grid-feeder interaction power upper and lower limit constraints.

5. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 4, characterized in that: In step S140, the ADMM algorithm is used to perform optimization calculations, and the process of finally obtaining the hierarchical partition dynamic power interaction results and the distributed resource output results includes: S141, based on the ADMM distributed solution algorithm, for the partition area a, the distributed optimization model based on the partition coordination principle is ， In the formula, is the objective function of region a, is the equality constraint for region a, represents the decision variable for region a; is an inequality constraint for region a. There exists an inequality between region a and its adjacent region b. , , Boundary equality constraints are used as convergence conditions for distributed coordination optimization among regions; S142, based on the SPPA-ADMM distributed solution algorithm, the SPPA-ADMM solution algorithm adaptively updates the spectral penalty parameters based on the spectral step size obtained according to the gradient descent method. Different from the adaptive ADMM distributed solution algorithm based on residual balance, the SPPA-ADMM distributed solution algorithm improves the convergence speed.

6. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 5, characterized in that: In step S200, a hierarchical distributed control architecture of the distribution network is first established, wherein the control architecture includes a distributed application APP group, a communication network and a physical facility architecture, wherein the physical facility architecture is a distribution substation, and the physical facility architecture includes a low-voltage distribution network, and the intelligent fusion terminal is arranged in the distributed application APP group; the low-voltage distribution network manages the operation control of the distribution substation through the intelligent fusion terminal, and the distribution substation is connected to distributed photovoltaics, and the photovoltaic, load and energy storage data of the substation are gathered at the intelligent fusion terminal through the communication network for calculation and processing, so as to form a control strategy to manage the power balance of the substation.

7. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 6, characterized in that: In step S200, the method for establishing the power regulation information model of the distribution network area is: S210, simulation modeling of information model between equipment and master station, interoperability of logical nodes between equipment, interactive control and data sharing based on IEC 61850 / CIM at the distributed application APP layer, and use of model driver to convert model and generate source code; Configure the communication network in the equipment supporting the coordinated power balance control, and automatically configure, generate and deploy the functions of the logical nodes, and use the model to define the peripheral information of the control mode and the information interaction method to achieve interoperability and convenient device interaction; S220. Use the MDE method between the IEC 61850 model and the function block to configure the communication network, and automatically configure, generate and deploy the functions of the logical nodes. The server model, logical device model, logical node model, data object model, data attribute model and common data class model in IEC 61850 are respectively mapped to the device model, resource model, composite function block model, basic function block model, data type and structured data type of the function block model, that is, all the logical nodes of the distribution equipment and the information data flows between them are regarded as a resource model, and each logical node is the corresponding function block model.

8. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 7, characterized in that: In step S320, the distributed photovoltaic output power is predicted using the calculation formula ， Where t is the current time, is the time step of prediction; To predict k time intervals added to the future; is the clearness index; is the predicted PV power; is the maximum value of the historical PV power data at the same time within the j-day period; and The calculation formula is: In the formula .

9. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 8, characterized in that: In step S330, the load forecast of the substation area adopts the calculation formula: The exponential smoothing factor is calculated using the formula: ， Where t is the current time and ∆t is the predicted time step; To predict k time intervals added to the future; is the predicted load power, is the actual historical load power; Represents the weight of the actual load power at time t to the predicted value, Represents the impact of the actual load power at time t on the previous day and the actual load power at time t 7 days ago on the predicted value.

10. A distributed resource collaborative control method based on hierarchical and partitioned autonomy of distribution network according to claim 9, characterized in that: In step S350, the power balance control process of the distribution station area includes: 1) Collect historical data and real-time operation data of distributed photovoltaic and regional loads. The above data is measured by the MMXU of each device through the DECP logical node, and the DGEN and DLOD carry the real-time control interaction and historical operation data of the power equipment. The DMDR integrates the data into a single power equipment for DPMC processing; 2) Predict the power generation at the next time With load power ,The data is collected by the DPMC logical node and the prediction data of the process is calculated; 3) Then calculate the power deviation between power generation and load , as the input for planning the optimal path for energy storage charging and discharging, the data is collected by the DPMC logic node from the previous process prediction data and the intermediate process data is calculated; 4) At the same time, based on the distributed energy storage operation data and prediction data, an energy storage charging and discharging optimization path is formed. The MMXU and DECP collect data, and the operation data of the energy storage logic node is formed by combining the DSTO and DBAT of the energy storage. The data flows to the DPMC for processing, and the control command is formed and sent back to the DSTO; 5) Send control commands to control the charging and discharging of energy storage, and predict the remaining SOC energy of energy storage. The data is calculated by the DPMC logical node; 6) According to the imbalance of power in the substation and the operation status of energy storage, the energy storage charging and discharging optimization path is adjusted until the power balance in the substation is optimal. DPMC issues control instructions to DBAT and DSTO for execution, and continuously feeds back the control results.

Images