Partitioned energy local area network two-stage energy management method and device, and electronic equipment

Abstract

The application particularly relates to a two-stage energy management method and device for a partitioned energy local area network and electronic equipment, wherein the method comprises the following steps: acquiring running state data of each energy local area network in a target area; determining a local reference and / or a zero-interconnection tendency reference of each energy local area network at a current moment according to the running state data of each energy local area network; calculating network coordination of each energy local area network at the current moment based on the local reference and / or the zero-interconnection tendency reference of each energy local area network at the current moment, and obtaining executable power instructions of each energy local area network according to the network coordination of each energy local area network at the current moment, with the target of minimizing interconnection power and line loss; and performing energy distribution on the corresponding energy local area network according to the executable power instructions of each energy local area network. Thus, the problems of poor consistency in the first stage, high cross-area energy carrying loss and energy storage life overdraft in multi-bus building energy management are solved.

Description

Technical Field

[0001] This application relates to the field of smart building energy management technology, and in particular to a two-stage energy management method, device and electronic equipment for zoned energy local area networks. Background Technology

[0002] With the electrification of building energy consumption, the continuous increase in the penetration rate of distributed photovoltaics, and the growing importance of demand-side flexible resources, the traditional home-level energy management approach under a single bus or single energy form is difficult to directly transfer to commercial building scenarios with multiple buses and cross-regional coupling. Therefore, how to systematically reduce the dimensions of physical electrical topology and optimization decisions has become a key breakthrough for energy management of large-scale smart buildings.

[0003] The relevant technology employs a two-stage energy management strategy. The first stage involves independent rolling optimization within each energy local area network (LAN). Using short-term predicted solar irradiance and load data as input, it accumulates photovoltaic (PV) operation and maintenance costs, energy storage power usage costs, and equivalent grid purchase and sale costs. It also considers constraints such as interface converter power limits, energy storage charge and discharge limits, state-of-charge (SOC) operating ranges, and internal power balance. The bottom-up solution yields the current operating power and grid-connected interaction reference for each LAN. The second stage, while maintaining minimal deviation from the energy storage reference for each LAN, minimizes instantaneous power transfer between LANs to reduce unnecessary inter-regional energy transport and improve local self-consumption rate. It only considers the current interface power limits, energy storage feasible region, and the overall building power balance.

[0004] However, the relevant technologies still have several engineering shortcomings. First, there is insufficient consistency between stages. The energy storage unit and grid connection interaction references given in the zoning stage are often short-term corrections in the global coordination stage, lacking hard constraints and cost internalization for deviations between the reference and the execution. Second, the suppression of cross-regional energy transport is not sufficient. Most schemes only minimize the power transmitted between regions at the current moment, without incorporating line and converter losses, power flow capacity between buses, and ramping constraints into a unified model. This causes the interconnection power to fluctuate frequently under the influence of price fluctuations and prediction errors, resulting in ineffective circulating power and incidental losses. Third, the lack of consideration for equipment physical characteristics and lifespan factors increases lifecycle costs. Fourth, uncertainties and communication realities are not fully modeled. Problems such as prediction errors, communication delays, data packet loss, and local actuator failures lack calculable robust protection domains and degradation strategies, leading to a disconnect where the optimized solution is feasible under extreme conditions but not feasible in actual implementation. These issues urgently need to be addressed. Summary of the Invention

[0005] This application provides a two-stage energy management method, device, and electronic equipment for zoned energy local area networks to solve problems such as poor stage consistency, high cross-zone energy transport losses, and overdraft of energy storage life in multi-bus building energy management.

[0006] The first aspect of this application provides a two-stage energy management method for a partitioned energy local area network, including the following steps: Acquire operational status data for each energy local area network within the target area; Determine the local reference and / or zero interconnection tendency reference for each energy local area network at the current moment based on the operational status data of each energy local area network; Based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment, the network coordination of each energy local area network at the current moment is calculated, and with the goal of minimizing interconnection power and line loss, the executable power command of each energy local area network is obtained according to the network coordination of each energy local area network at the current moment. Energy is allocated to the corresponding energy local area network according to the executable power command of each energy local area network.

[0007] Optionally, acquiring the operational status data of each energy local area network within the target area includes: Based on a preset step size, the current intraday rolling forecast data and the current measured data of each energy local area network within the target area are obtained. The operating status data of each energy local area network is obtained based on the intraday rolling forecast data and the measured data at the current moment for each energy local area network within the target area.

[0008] Optionally, determining the local reference and / or zero interconnection tendency reference of each energy local area network (ELAN) at the current moment based on the operating status data of each ELAN includes: Construct a first objective function for each of the energy local area networks based on a preset first constraint condition; The first objective function is solved based on the operating status data of each energy local area network to obtain the optimal decision sequence within the rolling window; Based on the optimal decision sequence, determine the local reference and / or zero interconnection tendency reference for each of the energy local area networks at the current moment.

[0009] Optionally, the first objective function is: ; in, For the current moment, The length of the scroll window. The electricity price at the current moment. For the first A local area network of energy The power of constant interaction with the mainnet. To preset the step size, This refers to the penalty coefficient for photovoltaic power generation restrictions. To limit PV power generation, This is the energy storage degradation cost coefficient. Let be the energy storage charging power of the k-th local energy network at time τ. For the first A local area network of energy Energy storage and discharge power at any given time For the first A local area network of energy The state of charge of the energy storage at any given time. For the bias penalty weight, This is the grid-connected execution power written back from the previous moment.

[0010] Optionally, obtaining the executable power command for each energy local area network based on the network coordination of each energy local area network at the current time includes: Construct a second objective function based on the preset second constraints; Based on the network coordination of each energy local area network at the current time, the second objective function is solved to obtain the executable power command for each energy local area network.

[0011] Optionally, the second objective function is: ; in, For the first The grid-connected power of a local area network. For the first The energy storage unit of a local area network performs power. For from the first The energy flow from the local area network to the first The power of the interconnecting lines in a local energy region. It is a set of interconnected edges between energy local area networks. The absolute value penalty factor for tie line power. This is the line loss coefficient. For the first The equivalent resistance of the link between the j-th energy local area network and the j-th energy local area network. For E-LAN ​​sets, The power deviation penalty factor is applied to the grid connection. For reference and interaction with the mainnet, The power deviation penalty factor is applied to energy storage. This serves as a reference for charging and discharging the energy storage unit.

[0012] Optionally, after allocating energy to the corresponding energy local area network according to the executable power command of each energy local area network, the method further includes: The writeback reference for solving the first objective function at the next moment is determined based on the executable power instruction, and the weight of the deviation penalty term when solving the local reference is updated according to the deviation between the executable power instruction and the local reference.

[0013] A second aspect of this application provides a two-stage energy management device for a partitioned energy local area network, comprising: The acquisition module is used to acquire the operating status data of each energy local area network within the target area; The determination module is used to determine the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network; The calculation module is used to calculate the network coordination of each energy local area network at the current time based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current time, and to obtain the executable power command of each energy local area network based on the network coordination of each energy local area network at the current time with the goal of minimizing interconnection power and line loss. The allocation module is used to allocate energy to the corresponding energy local area network according to the executable power command of each energy local area network.

[0014] Optionally, the acquisition module is specifically used for: Based on a preset step size, the current intraday rolling forecast data and the current measured data of each energy local area network within the target area are obtained. The operating status data of each energy local area network is obtained based on the intraday rolling forecast data and the measured data at the current moment for each energy local area network within the target area.

[0015] Optionally, the determining module is specifically used for: Construct a first objective function for each of the energy local area networks based on a preset first constraint condition; The first objective function is solved based on the operating status data of each energy local area network to obtain the optimal decision sequence within the rolling window; Based on the optimal decision sequence, determine the local reference and / or zero interconnection tendency reference for each of the energy local area networks at the current moment.

[0016] Optionally, the first objective function is: ; in, For the current moment, The length of the scroll window. The electricity price at the current moment. For the first A local area network of energy The power of constant interaction with the mainnet. To preset the step size, This refers to the penalty coefficient for photovoltaic power generation restrictions. To limit PV power generation, This is the energy storage degradation cost coefficient. Let be the energy storage charging power of the k-th local energy network at time τ. For the first A local area network of energy Energy storage and discharge power at any given time For the first A local area network of energy The state of charge of the energy storage at any given time. For the bias penalty weight, This is the grid-connected execution power written back from the previous moment.

[0017] Optionally, the computing module is specifically used for: Construct a second objective function based on the preset second constraints; Based on the network coordination of each energy local area network at the current time, the second objective function is solved to obtain the executable power command for each energy local area network.

[0018] Optionally, the second objective function is: ; in, For the first The grid-connected power of a local area network. For the first The energy storage unit of a local area network performs power. For from the first The energy flow from the local area network to the first The power of the interconnecting lines in a local energy region. It is a set of interconnected edges between energy local area networks. The absolute value penalty factor for tie line power. This is the line loss coefficient. For the first The equivalent resistance of the link between the j-th energy local area network and the j-th energy local area network. For E-LAN ​​sets, The power deviation penalty factor is applied to the grid connection. For reference and interaction with the mainnet, The power deviation penalty factor is applied to energy storage. This serves as a reference for charging and discharging the energy storage unit.

[0019] Optionally, after allocating energy to the corresponding energy local area network according to the executable power command of each energy local area network, the allocation module is further configured to: The writeback reference for solving the first objective function at the next moment is determined based on the executable power instruction, and the weight of the deviation penalty term when solving the local reference is updated according to the deviation between the executable power instruction and the local reference.

[0020] A third aspect of this application provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being configured to perform a two-stage energy management method for a partitioned energy local area network as described in the above embodiments.

[0021] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the two-stage energy management method for a partitioned energy local area network as described in the above embodiments.

[0022] Therefore, this application embodiment obtains the operational status data of each energy local area network (LAN) within the target area; determines the local reference and / or zero interconnection tendency reference of each energy LAN at the current moment based on the operational status data of each energy LAN; calculates the network coordination of each energy LAN at the current moment based on the local reference and / or zero interconnection tendency reference of each energy LAN at the current moment, and obtains the executable power command of each energy LAN based on the network coordination of each energy LAN at the current moment with the goal of minimizing interconnection power and line loss; and allocates energy to the corresponding energy LAN according to the executable power command of each energy LAN. This solves the problems of poor phase consistency, high cross-regional energy transport losses, and overdraft of energy storage life in multi-bus building energy management.

[0023] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0024] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application. Figure 2 This is a schematic diagram of the physical topology of a building-E-LAN ​​interconnection for a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; Figure 3 This is a schematic diagram of a two-stage energy management process for a partitioned energy local area network according to an embodiment of this application; Figure 4 This is a schematic diagram of the constraint-objective relationship for a stage-one local optimization of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; Figure 5 This is a schematic diagram of the second stage of network coordination in a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; Figure 6 This is a schematic diagram of the time axis and rolling consistency write-back of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; Figure 7 This is a schematic diagram illustrating the fault-tolerant and backup mode of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; Figure 8 This is a schematic diagram of a two-stage energy management device for a local area network (LAN) according to an embodiment of this application. Figure 9 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0025] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0026] Before introducing the two-stage energy management method for partitioned energy local area networks in the embodiments of this application, let's briefly introduce the two-stage energy management method for partitioned energy local area networks in related technologies.

[0027] Specifically, with the increasing electrification of building energy consumption, the rising penetration rate of distributed photovoltaic (PV) power, and the growing emphasis on demand-side flexibility, buildings simultaneously contain multiple types of controllable equipment (air conditioning / fresh air, lighting, DC and AC loads, Energy Storage Units (ESUs), etc.) and multi-source power routing (grid connection, power exchange between buses, and power flow within floors / functional areas). The home-level energy management approach based on a single bus or single energy form is difficult to directly transfer to the multi-bus, cross-regional coupling scenarios of commercial buildings: on the one hand, rule-based methods are easy to implement but expand exponentially with the number of devices and buses; on the other hand, while optimization methods such as MILP / MINLP and Deep Reinforcement Learning (DRL) can handle coupling and uncertainty, they face pain points such as huge decision dimensions, limited solution time, and uncontrollable safety during the learning period. Therefore, "systematically reducing the order of physical-electrical topology and optimization / learning dimensions" has become a key solution for large-scale smart building energy management.

[0028] In engineering practice, commercial buildings increasingly adopt a hierarchical DC power supply topology: High-Voltage Direct Current (HVDC), Medium-Voltage Direct Current (MVDC), and Low-Voltage Direct Current (LVDC) serve DC loads of different distance / capacity levels respectively; AC loads are connected to the DC bus via AC / DC converters; and each floor is equipped with photovoltaic arrays and energy storage converters of varying capacities. For ease of management, an "Energy Local Area Network (E-LAN)" is established centered around the "DC bus with energy storage": each E-LAN ​​aggregates PV, ESU, and local loads on its bus side; buildings interact with the main grid via bidirectional AC / DC converters, and E-LANs interact with each other via isolated bidirectional DC / DC converters, forming an energy cascading and power flow reconfiguration capability of "grid-bus-zone". This physical organization provides a natural structured boundary for subsequent zone optimization and cross-zone coordination.

[0029] The closest existing solution to this invention is a two-stage energy management strategy: its idea is to first partition the system and then coordinate it. The first stage involves independent rolling optimization within each E-LAN ​​(typically with a step size of 1 hour and several look-ahead steps). The input is short-term predicted irradiance / load (LSTM is used in the literature for short-term irradiance and load prediction). The objective function accumulates PV operation and maintenance costs, ESU power usage costs, and the cost of purchasing and selling electricity from the "equivalent grid." Constraints include interface converter power limits, ESU charging and discharging upper and lower limits, SOC operating range, and the power balance within each E-LAN ​​at any given time. To reduce computational complexity, the first stage proceeds from the bottom up, starting from the "end E-LAN" of the topology, obtaining the current execution power of each E-LAN ​​and the power reference for interaction with the main grid. The second stage, while maintaining the ESU reference from the first stage without significant deviation, minimizes the instantaneous power exchange (i.e., "internal flow") between E-LANs to reduce unnecessary cross-regional energy transport and improve local self-consumption rate. This stage only considers constraints such as the current interface power upper and lower limits, ESU power, SOC feasible area, and the total power balance of the entire building. The two-stage cascading allows the model size to increase linearly with the number of E-LANs, significantly alleviating the dimensionality explosion and computational burden of centralized unified optimization.

[0030] However, for energy management in multi-bus commercial buildings, the existing closest approach of "two-stage (zone-based rolling, global coordination)" alleviates the dimensionality explosion of centralized unified optimization, but still has several engineering shortcomings: First, insufficient consistency between stages. The ESU (energy storage unit) and bus-grid interaction reference given in the zoning stage are often briefly corrected in the global coordination stage, lacking hard constraints and cost internalization of "reference-execution deviation," leading to long-term SoC (state of charge) drift and intraday / cross-day "reverse compensation" phenomena, weakening the effectiveness of rolling optimization; Second, insufficient explicit suppression of cross-zone energy transport. Most schemes only minimize the power exchange at the current moment, without incorporating line / converter losses, inter-bus power flow capacity, and ramping constraints into the unified model. As a result, "internal flow" fluctuates frequently under the drive of price / prediction errors, generating ineffective circulating power and incidental losses; Third, lack of equipment physics and lifespan considerations. The power limits, ramp-up, minimum start-up and shutdown of interface converters, as well as the derating and degradation costs of energy storage, are often simplified or ignored. While short-term costs may decrease, long-term TCO (Total Cost of Ownership) may increase. Fourth, uncertainties and communication realities are not adequately modeled. Prediction errors, communication delays / packet loss, and local failures to operate are often handled with empirical margins in existing processes. The lack of calculable robust / tube protection domains and degradation strategies leads to a disconnect between "feasible solutions and infeasible implementation" during extreme periods.

[0031] This application addresses the aforementioned problems by proposing a two-stage energy management method for zoned energy local area networks (LANs). In this method, embodiments of this application acquire operational status data for each LAN within a target area; based on the operational status data of each LAN, determine the local reference and / or zero interconnection tendency reference for each LAN at the current moment; based on the local reference and / or zero interconnection tendency reference for each LAN at the current moment, calculate the network coordination for each LAN at the current moment, and, with the goal of minimizing interconnection power and line losses, obtain the executable power command for each LAN based on the network coordination at the current moment; and allocate energy to the corresponding LAN based on the executable power command for each LAN. This solves the problems of poor phase consistency, high cross-zone energy transport losses, and overdraft of energy storage life in multi-bus building energy management.

[0032] Specifically, Figure 1 This is a flowchart illustrating a two-stage energy management method for a partitioned energy local area network provided in an embodiment of this application.

[0033] like Figure 1 As shown, the two-stage energy management method for this partitioned energy LAN includes the following steps: In step S101, the operating status data of each energy local area network within the target area is obtained.

[0034] Optionally, in some embodiments, obtaining the operating status data of each energy local area network within the target area includes: obtaining the current intraday rolling forecast data and the current measured data of each energy local area network within the target area based on a preset step size; and obtaining the operating status data of each energy local area network based on the current intraday rolling forecast data and the current measured data of each energy local area network within the target area.

[0035] Understandably, this is done by using a fixed step size. (Preset step size) Acquire the intraday rolling forecast data (including irradiance forecast, load forecast and electricity price forecast within a future time window) and the measured data (including actual photovoltaic output, actual load power, energy storage state of charge, energy storage temperature and DC bus voltage, etc.) of each energy local area network at the current moment; integrate the two to form a complete operating status dataset, which will be used as the input parameters for determining the local reference in the future.

[0036] In step S102, the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment are determined based on the operating status data of each energy local area network.

[0037] Among them, the local reference refers to the current operating baseline obtained after each energy local area network is solved independently in the first phase of partition rolling optimization, and the zero interconnection tendency reference refers to the expected value of the power exchanged between the current interconnection lines and other energy local areas networks after each energy local area network is solved independently in the first phase of partition rolling optimization, and its value is zero or close to zero.

[0038] Specifically, each E-LAN ​​(Energy Local Area Network) independently solves for a length of To address the scrolling issue, we obtain the local reference at the current moment: (ESU charging and discharging) (Interaction with the mainnet) and optional zero-interconnection tendency reference During this phase, the constraints and costs of this E-LAN ​​are used to ensure linear growth in scale.

[0039] Optionally, in some embodiments, determining the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network includes: constructing a first objective function for each energy local area network based on preset first constraints; solving the first objective function based on the operating status data of each energy local area network to obtain the optimal decision sequence within a rolling window; and determining the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the optimal decision sequence.

[0040] Understandably, the first preset constraint includes power balance (within the E-LAN): ; (In Phase 1, interconnect power is not explicitly allocated, tending towards zero internal current, which is fine-tuned in Phase 2).

[0041] SoC recursion and feasible corridors: ; Power and ramp boundary (interface / converter): ; After solving, take the current time reference (local reference and / or zero interconnection tendency reference):

[0042] .

[0043] A first objective function is constructed for each energy local area network (LAN) based on a preset first constraint. The first objective function is used to comprehensively minimize the deviation between the electricity purchase cost, photovoltaic curtailment penalty, energy storage degradation cost, and the write-back reference from the previous time step within a rolling window. The first objective function is solved based on the operating status data of each energy LAN to obtain the optimal decision sequence at each time step within the rolling window. The decision value corresponding to the current time step is extracted from the optimal decision sequence and used as the local reference for each energy LAN at the current time step. The local reference includes the energy storage unit charging and discharging power reference, the power interaction with the main grid reference, and the zero interconnection tendency reference.

[0044] Optionally, in some embodiments, the first objective function is: ; in, For the current moment, The length of the scroll window. The electricity price at the current moment. For the first A local area network of energy The power of constant interaction with the mainnet. To preset the step size, This refers to the penalty coefficient for photovoltaic power generation restrictions. To limit PV power generation, This is the energy storage degradation cost coefficient. Let be the energy storage charging power of the k-th local energy network at time τ. For the first A local area network of energy Energy storage and discharge power at any given time For the first A local area network of energy The state of charge of the energy storage at any given time. For the bias penalty weight, This is the grid-connected execution power written back from the previous moment.

[0045] Understandably, the first objective function includes minimizing the electricity purchase cost, PV generation curtailment penalty, ESU degradation, and reference bias, among which... A degenerate approximation for piecewise convexity (penalizing deep cycles / high multiplication). For reference when writing back to the previous step (to maintain consistency). The first objective function drives each energy local area network to make decisions at the current moment that achieve a comprehensive optimum in terms of electricity purchase economy, photovoltaic grid integration rate, energy storage lifespan friendliness, and inter-stage consistency.

[0046] In step S103, based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current time, the network coordination of each energy local area network at the current time is calculated, and with the goal of minimizing interconnection power and line loss, the executable power command of each energy local area network is obtained according to the network coordination of each energy local area network at the current time.

[0047] In the second stage of the two-stage energy management method, network coordination refers to the process of uniformly determining the grid-connected power, energy storage unit power, and tie-line power of each energy local area network, which is a small power network consisting of multiple energy local area networks and their interconnection topology within the target area, by solving a global optimization problem.

[0048] Specifically, based on the local reference and zero interconnection tendency reference of each energy local area network (LAN) obtained in the first stage solution, and combined with the interconnection topology between multiple energy LANs in the target area, the solution is performed with the objective of minimizing the sum of the absolute values ​​of the power of the tie lines between each energy LAN and the sum of the line losses generated on the tie lines. This yields the executable power command for each energy LAN at the current time. The executable power command includes the power executed by the energy storage unit, the power executed by grid connection, and the power of the tie lines between energy LANs, thus realizing the whole-network coordinated fine-tuning based on the partition optimization results.

[0049] Optionally, in some embodiments, obtaining the executable power command of each energy local area network (ELAN) based on the network coordination of each ELAN at the current moment includes: constructing a second objective function based on a preset second constraint; and solving the second objective function based on the network coordination of each ELAN at the current moment to obtain the executable power command of each ELAN.

[0050] Understandably, Phase Two introduces "scenario / tube" constraints or power reservations to cover prediction errors and communication anomalies; the edge implements "retry-degradation-hot handover". The preset second constraint includes the instantaneous power balance of each E-LAN ​​(the positive sign is conventionally "externally supplemented"): ; Communication capacity and symmetrical power: ; ESU and grid connection execution boundary: ; Optional robust allowance / pipe radius: ; in, This is the set of prediction errors (box set / ellipsoid set).

[0051] While maintaining the Phase 1 reference "without significant deviation", solve for network coordination at the current moment, aiming to minimize interconnect power and line loss while tracking the Phase 1 reference; output an executable solution. Power of each connecting line .

[0052] Optionally, in some embodiments, the second objective function is: ; in, For the first The grid-connected power of a local area network. For the first The energy storage unit of a local area network performs power. For from the first The energy flow from the local area network to the first The power of the interconnecting lines in a local energy region. It is a set of interconnected edges between energy local area networks. The absolute value penalty factor for tie line power. This is the line loss coefficient. For the first The equivalent resistance of the link between the j-th energy local area network and the j-th energy local area network. For E-LAN ​​sets, The power deviation penalty factor is applied to the grid connection. For reference and interaction with the mainnet, The power deviation penalty factor is applied to energy storage. This serves as a reference for charging and discharging the energy storage unit.

[0053] Understandably, the second objective function includes minimizing interconnect power and line loss while tracking the Phase 1 reference.

[0054] In step S104, energy is allocated to the corresponding energy local area network according to the executable power command of each energy local area network.

[0055] Specifically, executable power commands are sent to the edge controllers of each energy local area network. The edge controllers convert the commands into execution commands for each device and send them to the corresponding local actuators. After the actuators complete their actions, they return an execution response to the upper-level controller. The upper-level controller verifies the execution response. When the verification passes and the execution result is within the preset deviation range, the energy allocation is confirmed to be successful. When an actuator fails to operate, overshoots, or a communication timeout is detected, a degradation strategy is triggered, and the system enters a local self-balancing backup operation mode.

[0056] Optionally, in some embodiments, after allocating energy to the corresponding energy local area network according to the executable power command of each energy local area network, the method further includes: determining the write-back reference for solving the first objective function at the next moment based on the executable power command, and updating the weight of the deviation penalty term when solving the local reference according to the deviation between the executable power command and the local reference.

[0057] Understandably, writing back the actual execution volume of Phase Two as the boundary and tracking reference of "Phase One at the next moment" (with deployment factors / penalty weights updated synchronously) suppresses the reverse compensation of "release first, charge later / buy first, sell later". The execution solution for Phase Two... Written as the next moment, phase one Weight of “bias penalty” ( (Adaptive), and update SoC initial values. When a "release before charging / purchase before selling" pattern is detected, the corresponding weight is increased and the local ramp-up upper limit is tightened to achieve consistent rolling maintenance across the entire process.

[0058] Therefore, this invention, while maintaining the scalability of "rolling partitions first, then fine-tuning the global process," fills the gaps in five aspects: consistency, physical properties, lifetime, robustness, and feasibility. Specifically, in the partitioning phase, hard constraints such as SoC corridors, rate derating envelopes, and converter ramping are introduced, and the "deviations that may be caused by subsequent coordination" are pre-loaded into the objective function as deployment factors / ad valorem default terms to suppress the reverse compensation of "release before charging / buy before selling." In the global coordination phase, an explicit line / interface loss model and communication capacity constraints are used to unify "minimum power transfer" and "minimum loss / minimum cost." In terms of time, rolling consistency maintenance is adopted (the coordination results are written back as the reference and boundary for the next partitioning phase) to avoid policy drift caused by cross-phase decoupling. In terms of uncertainty handling, scenario / pipeline robustness is introduced to ensure executability under prediction errors and communication anomalies, and a fault-tolerant channel of "retry-degradation-hot switching" is provided. Thus, this invention achieves: ① significantly reducing cross-regional energy backflow and improving local self-consumption rate without increasing the central solver burden; ② balancing cost optimization and equipment lifespan friendliness, suppressing deep loops and high-rate behavior; ③ ensuring consistent solution across stages, robustly handling prediction and communication disturbances, and possessing engineering implementation and large-scale deployment capabilities.

[0059] It should be noted that the degradation approximation in this application embodiment adopts a piecewise convex approach: ; Upper bound of multiplier Adaptive derating based on SoC / temperature (lookup table or linear interpolation) serves as a hard upper bound in both Phase 1 and Phase 2, avoiding short-term optimization that could lead to overextension of lifespan.

[0060] To facilitate those skilled in the art to further understand the two-stage energy management method for partitioned energy local area networks in the embodiments of this application, the following is combined with... Figures 2 to 7 The embodiments shown will be described in detail.

[0061] Specifically, this invention proposes a two-stage energy management method, system, and storage medium for multi-bus smart buildings, using Energy Local Area Networks (E-LANs) as atomic units, employing a "partition-based rolling first—global then fine-tuning" approach. The overall concept is: on the edge controller, each E-LAN ​​is independently looked up in a minute-level closed loop with a step size of... The process involves partitioned rolling optimization (forming a reference for grid connection and energy storage at the current moment), followed by simultaneous instantaneous network coordination (minimizing cross-regional power transmission under constraints of interconnection capacity / line loss and tracking the partitioned reference). The execution results are then written back as the reference and boundary for the next step, maintaining consistency and executability across stages. The entire process unifies physical constraints such as power balance, SoC corridors, rate derating, grid connection / interconnection capacity, and ramp-up into a feasible domain. It incorporates electricity purchase costs, cross-regional power transmission penalties, and segmented convex degradation costs into the objective function, and includes scenario / pipeline robustness and "retry-degradation-hot switching" fault tolerance to ensure operational feasibility even under prediction errors, communication jitter, and local failures to operate. Compared to centralized unified optimization or only partitioned control, this invention significantly reduces cross-regional energy backhaul and electricity purchase costs without significantly increasing central computing power, improves local self-consumption rate, and balances equipment lifespan and grid connection constraints.

[0062] Furthermore, this invention divides the multi-bus smart building into several Energy Local Area Networks (E-LANs). Each E-LAN ​​is centered on a DC bus, aggregating local photovoltaic arrays (PV), energy storage units (ESU), peak-shaving loads (HVAC / lighting / DC loads), and AC / DC interfaces with the upstream power grid. Different E-LANs are interconnected via isolated bidirectional DC / DC converters to form a controllable bus interconnection. The building's main incoming line is equipped with a grid connection point capable of bidirectional metering, and tie switches / current limiters are provided when necessary. Sensing and metering coverage includes: bus power / voltage, current direction, PV output, load power, ESUSoC / temperature / rate, tie line power and status. The edge controller (industrial computing unit) handles two-stage energy management and execution; cloud services provide day-ahead / intraday forecasts and parameter feedback.

[0063] The uplink signals include those from each E-LAN. Meter-level measurements and rolling forecasts are performed. Downlink signals include grid-connected power reference, charging and discharging power of each ESU, E-LAN ​​interconnect power, and power settings / operation levels of peak-shifting loads. In case of link anomalies, the edge independently closes the loop and enters a "zero interconnection / local self-balancing priority" backup mode.

[0064] Furthermore, such as Figure 2 As shown, Figure 2 This is a schematic diagram of the physical topology of a building-E-LAN ​​interconnection according to an embodiment of the two-stage energy management method of a partitioned energy local area network provided in this application. It shows the main incoming line, multiple DC buses (E-LANs), their respective PV / ESU / loads, and bidirectional DC / DC interconnections, with measurement and control points marked. The building is connected to the external power grid via the main incoming line, which is equipped with a bidirectional metering grid connection point and a matching grid-connected converter responsible for bidirectional power exchange between the building and the external power grid. The building contains multiple energy local area networks (such as E-LAN ​​A, E-LAN ​​B, E-LAN ​​C, etc.), each centered around a DC bus. The bus aggregates local distributed photovoltaic arrays (PV), energy storage units (ESU), and various AC / DC loads (including air conditioning, lighting, DC loads, etc.). The PV, energy storage, and loads within each energy local area network are all connected to this bus through their respective converters. Different energy local area networks (LANs) are interconnected in a controllable manner through isolated bidirectional DC / DC converters, forming a multi-level power supply architecture of "main incoming line - DC bus - zone". Data collected by the measurement points is transmitted to the edge controller via uplink signals, providing the measured data required for real-time operation status and rolling forecasting for two-stage energy management. At the same time, the executable power commands generated by the edge controller are sent to each grid-connected converter, energy storage converter, and tie-line converter via downlink signals, realizing the energy distribution and coordinated control of each energy LAN.

[0065] Furthermore, such as Figure 3 As shown, Figure 3This is a schematic diagram of a two-stage energy management process for a partitioned energy local area network (LAN) according to an embodiment of this application. It illustrates the complete closed-loop control process from data acquisition to execution feedback, arranged from S1 to S6 from top to bottom. Stage 1 (partitioned rolling optimization) and Stage 2 (global instantaneous coordination) are highlighted with bold frames, and the consistency write-back link is marked with a dashed arrow on the right. The process involves acquiring intraday rolling forecast data (including irradiance forecast, load forecast, and electricity price forecast) and the real-time state of charge of energy storage units; each LAN independently solves a rolling optimization problem of length H to obtain the local reference of each LAN at the current moment; based on the local reference of each LAN and the interconnection topology, minimizing interconnection power and line losses while tracking the Stage 1 reference, the executable power command for each LAN at the current moment is obtained; the executable power command is sent to each actuator, and command response and action verification are completed; when an actuator fails to operate or exceeds its limit, a deceleration is triggered. The system employs a multi-stage strategy: The grid-connected power of Phase 2 is written back along the right-hand write-back link as the write-back reference for Phase 1 at the next moment. The deviation penalty weight is adaptively updated based on the deviation between the current grid-connected power and the grid-connected reference. Simultaneously, the energy storage unit's power is used to update the initial state of charge of the energy storage unit, suppressing the reverse compensation phenomenon of "discharge before charging / purchase before selling," and maintaining consistency between phases. Scenario / pipeline robust constraints are introduced to address prediction errors and communication anomalies, and a three-stage fault-tolerant mechanism of "retry-degradation-hot switching" is implemented to ensure the system can still operate safely and reliably under extreme conditions.

[0066] Furthermore, such as Figure 4 As shown, Figure 4This diagram illustrates the constraint-objective relationship of a phase one local optimization of a two-stage energy management method for a partitioned energy local area network (E-LAN) according to an embodiment of this application. It shows the display of power balance, SoC corridors, rate / ramp, and cost items (electricity purchase, power curtailment, degradation, and reference deviation) within the E-LAN. Phase one constraints include power balance constraints, energy storage state of charge (SoC) corridor and recursive constraints, rate and ramp constraints, and photovoltaic power curtailment constraints: the power balance constraint requires that the load power equals photovoltaic output minus curtailed power plus grid-connected power plus discharge power minus charging power; the SoC corridor and recursive constraints update the SoC at the next moment based on charge / discharge power, charge / discharge efficiency, and rated capacity, and limit it within a preset upper and lower operating range; the rate and ramp constraints impose limits on the upper limit of energy storage charge / discharge power (adaptively derating with SoC and temperature) and the ramp rate of power interaction with the main grid; the photovoltaic power curtailment constraint limits the curtailed power to between zero and the current photovoltaic output. The objective function for Phase 1, while satisfying the above constraints, involves a weighted summation and minimization of the electricity purchase cost, photovoltaic curtailment penalty, energy storage degradation cost, and deviation penalty from the previous time-to-time reference at each moment within the rolling window H. The electricity purchase cost is related to grid-connected power and real-time electricity price, encouraging electricity purchase from the grid when prices are low. The curtailment penalty is used to avoid curtailment. The degradation cost adopts a piecewise convex function form, penalizing deep cycling and high-rate charging / discharging behaviors to suppress overextension of energy storage lifespan. The reference deviation penalty term tracks the grid-connected execution power written back from the previous time-to-time, maintaining consistency between phases. After solving this local optimization problem, the current time-to-time energy storage charging / discharging power reference and grid-connected power reference are extracted from the optimal decision sequence and used as input for Phase 2.

[0067] Furthermore, such as Figure 5 As shown, Figure 5 This is a schematic diagram of the second stage of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application; it shows a small network composed of E-LAN ​​nodes and interconnecting edges; the goal is "less cross-connection / lower loss / closer to reference"; the capacity is marked on the edges. With loss coefficient Each E-LAN ​​node (e.g., E-LAN ​​A, E-LAN ​​B, E-LAN ​​C) has a grid connection point and a corresponding grid connection execution power. The E-LANs are interconnected via tie lines formed by isolated bidirectional DC / DC converters. Transmission capacity limits and equivalent line resistance r_ij (i.e., loss coefficients) are marked on the tie lines. Interconnection power is constrained by tie line capacity and satisfies symmetry constraints. The optimization objective of Phase Two is "less cross-regional transmission, lower losses, and closer to the reference": on the one hand, minimizing the sum of the absolute values ​​of tie line power and the sum of line losses, i.e., the "less cross-regional transmission" objective shown on the vertical axis of the figure, thereby suppressing unnecessary cross-regional energy transport and reducing transmission losses; on the other hand, minimizing the sum of squared deviations between the grid connection execution power and energy storage execution power of each E-LAN ​​relative to the local reference given in Phase One, i.e., the "following the reference" objective shown on the horizontal axis of the figure, ensuring that the coordination results of Phase Two do not significantly deviate from the zonal optimization intentions of each E-LAN. By jointly optimizing these two objectives, Phase 2 rationally allocates interconnection power among E-LAN ​​nodes under the constraints of tie line capacity and loss coefficient, thereby achieving coordinated determination of the grid-connected execution power, energy storage execution power, and tie line power of each E-LAN ​​at the current moment. This reduces cross-regional energy backhaul and transmission losses while maintaining consistency with the zonal reference.

[0068] Furthermore, such as Figure 6 As shown, Figure 6 This diagram illustrates the timeline and rolling consistency write-back of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application. It shows the execution order of Stage 1 (partitioned rolling optimization) and Stage 2 (global instantaneous coordination) on the timeline, as well as the correlation effect of the consistency write-back mechanism on decisions at adjacent moments. The timeline advances from the previous moment to the current moment t and future moments t+1, t+2, etc., with a preset step size Δt. At each moment, Stage 2 is executed first to complete network coordination for the current moment, outputting the grid-connected execution power and the energy storage unit execution power. Subsequently, while executing control, the grid-connected execution power of Stage 2 is written back as the write-back reference for Stage 1 at the next moment, and the energy storage unit execution power is used to update the initial value of the energy storage state of charge at the next moment. Simultaneously, the deviation penalty weight is adaptively updated based on the execution deviation at the current moment. When the system detects that the energy storage unit has a reverse compensation mode of "discharging before charging" or "purchasing electricity before selling electricity", the system increases the deviation penalty weight and tightens the upper bound of the grid connection ramp constraint. This ensures that the rolling optimization of the next time step fully considers the actual execution situation of the previous time step when solving the problem, thus avoiding strategy drift between partition optimization and global coordination.

[0069] Furthermore, such as Figure 7 As shown, Figure 7This diagram illustrates the fault-tolerance and backup mode of a two-stage energy management method for a partitioned energy local area network according to an embodiment of this application. It shows the "retry-degradation-hot switchover" path triggered by communication failure / failure to operate. The backup strategy is "zero interconnection / local self-balancing priority + grid connection limiting". Specifically, when the communication link is abnormal or the actuator fails to operate, the retry mechanism is first triggered to quickly resend timed-out or unresponsive commands. If the retry is successful, the normal execution process is restored. If the retry fails, the degradation mode is entered, the power of the interconnection line between the energy local area networks is set to zero at the current moment, and the grid connection power amplitude is limited. The strategy of "zero interconnection, local self-balancing priority, and grid connection limiting" ensures that the building can still operate safely under abnormal conditions. When the abnormality is resolved, the hot switchover mechanism is triggered. After verifying that the system parameters are normal, the system smoothly recovers to the normal scheduling mode, avoiding power surges during the recovery process. Meanwhile, the system records and interprets fault tolerance trigger times, causal chains, and strategy selections with timestamps, providing a basis for subsequent analysis and compliance audits. The version signing and rollback module adopts a dual-version snapshot mechanism, saving rollback points when parameters change or strategies are updated, and quickly reverting to the previous stable version when parameter anomalies or strategy failures are detected. Through the coordinated work of the above fault tolerance chain and supporting mechanisms, this invention can still ensure the continuous executable operation of the system under conditions of communication jitter, prediction errors, and local equipment anomalies, significantly improving the engineering robustness and implementation reliability of the energy management method.

[0070] The embodiments of this application can also compare the characteristics of the two-stage energy management method and related technologies of the partitioned energy local area network.

[0071] Specifically, compared to the publicly disclosed "two-stage (partition rolling first, global then coordination)" scheme, this invention, without changing the overall "partition-coordination" framework, introduces key mechanisms such as inter-stage rolling consistency maintenance, explicit loss and capacity-constrained interconnection optimization, lifetime-friendly energy storage constraints and costs, and scenario / pipeline robustness and fault-tolerant loops. Therefore, under the same prediction and computing power conditions, this invention can stably transform "partition reference" into "executable solution," suppressing reverse compensation of "discharge before charging / purchase before selling"; at the network layer, it reduces ineffective cross-regional energy transport and improves local consumption with the joint objective of "minimum cross-regional power transfer + minimum loss"; simultaneously, it embeds both rate derating and degradation costs to avoid sacrificing battery life for short-term cost savings; and in disturbance and communication anomaly scenarios, it relies on robust protection domains and a "retry-degradation-hot switching" fault-tolerant chain to ensure solution consistency and continuous, deployable operation.

[0072] From a causal chain perspective, the power balance, SoC corridor, interconnect capacity / ramp, and line loss proposed in this application's embodiments jointly define the feasible domain of "can be done / cannot be done"; the consistency write-back forwards the execution quantity of stage two as the reference and boundary of stage one, enabling "what to do" and "what can be done" to gradually converge within the rolling window; the segmented convex degradation cost and temperature / SoC dynamic derating quantify "whether it's worth doing" into the objective function, achieving cost minimization and lifetime friendliness simultaneously; the robustness and fault tolerance module ensures that the above mechanisms still hold under uncertain and abnormal conditions. Thus, this invention provides verifiable, computable, and executable systematic improvements to address four types of weak links (insufficient stage consistency, insufficient cross-regional energy transport suppression, lack of equipment physics and lifetime, and lack of modeling of uncertainty and communication reality).

[0073] Furthermore, the embodiments of this application have the following advantages: Strong execution consistency: The closed-loop write-back and ramp / corridor linkage of reference-execution significantly reduce intraday / cross-day drift and reverse compensation, making the rolling optimal "calculated and achievable".

[0074] Optimal synergy between cost and lifespan: Minimum cross-transmission + explicit loss modeling reduces power purchase and internal flow losses; segmented convex degradation + derated envelope suppresses deep circulation and high rate, optimizing total lifespan cost (TCO).

[0075] Linear scalability and real-time friendliness: Decomposed using E-LAN ​​as the atomic unit, the scale of Phase 1 increases linearly with the number of partitions; Phase 2 only solves the small network power distribution at the current moment, satisfying minute-level online calculation time.

[0076] Robust, deployable, and auditable: The scenario / pipeline robustness ensures feasibility under prediction errors, the fault-tolerant chain handles refusal to operate / disconnection, and the accompanying version and log traceability are suitable for large-scale engineering deployment and compliance auditing.

[0077] Therefore, this application adopts a two-stage energy management framework (linearly scalable) with E-LAN ​​atomization; taking "Energy Local Area Network (E-LAN)" as the smallest unit: Stage 1 performs independent rolling optimization (window H) in each E-LAN, and Stage 2 performs global instantaneous coordination (network layer power allocation) at the current moment. This achieves linear scaling with scale without centralized explosive modeling. This application employs inter-stage rolling consistency maintenance and reverse compensation suppression, writing the execution results of Stage 2 (grid connection and energy storage power) back to Stage 1 in the next time period. Combined with reference tracking penalties and ramp / lock boundaries, it avoids strategy drift of "discharge before charging / purchase before selling," ensuring consistency between solution and line. This application employs explicit loss and capacity-constrained interconnection power optimization, simultaneously considering boundaries such as interconnection capacity, line / converter losses, and symmetrical power flow in Stage 2, aiming for "minimum interconnection power + minimum loss + tracking Stage 1 reference," suppressing ineffective cyclic power and improving local self-consumption rate. This application's embodiments employ a lifetime-friendly energy storage collaboration and robust / fault-tolerant deployment mechanism. It uses a two-stage decision-making process with segmented convex degradation cost and SoC / temperature derating envelope dual embedding to balance short-term costs and battery life. Furthermore, it introduces a scenario / tube robust protection domain and a three-stage fault tolerance mechanism of "retry-degradation-hot switching" to ensure executability under prediction errors and communication anomalies.

[0078] Furthermore, without altering the overall principle of "using E-LAN ​​as the atomic unit, rolling the partitions first, then fine-tuning the global model, and combining consistent write-back," the solvers, model details, and device configurations of Phase 1 and Phase 2 can be replaced without affecting the achievement of the invention's objectives. Specifically, the partition rolling optimization in Phase 1 can be replaced by adaptive MPC, MILP / MINLP, dynamic programming, or heuristics (tabu search, genetic algorithm); degradation costs can be achieved by the equivalent full-cycle (EFC) model of rainflow counting, a semi-empirical formula containing the Arrhenius temperature factor, or a second-order cone (SOCP) / piecewise linear approximation; the derating and power upper bound can be determined by table lookup interpolation or by piecewise linear constraints fitted by the manufacturer's curve. In addition to centralized quadratic programming, Phase 2 network coordination can employ distributed decoupling using ADMM / Lagrange relaxation, or Tube-MPC, Distributed Optimization Reduction (DRO), and chance-constrained methods to handle forecast uncertainty and interconnection capacity fluctuations. The "minimum interconnection + loss" objective can also be addressed using equivalent upper bounds on line losses and weighted averages. The loss function can be represented by a 1-norm or piecewise linear loss function. Maintaining "solution-line consistency" is not limited to 2-norm tracking penalties and ramp tightening; it can also be achieved using a reference locking window, an execution bias budget, or soft / hard bandwidth linkage rules.

[0079] At the physical and system level, power coupling between E-LANs is not limited to bidirectional DC / DC; it can also be achieved through low-voltage AC interconnection (AC / DC / AC) or common bus coupling. Different topologies such as star, chain, and ring can all be solved by the same coordination layer. Energy storage units are not limited to electrochemical batteries and can be replaced by cold / heat storage (ice storage, water storage), flywheels, compressed air, or hydrogen energy storage. Correspondingly, SoC can be replaced by equivalent states such as "cold / hot inventory" or "gas cavity energy share," and enter the two-stage model with the same corridor and derating envelope. On the load side, it can be extended to peak-shifting loads such as elevator group control, charging piles, and cold / hot aisle scheduling in data centers. Their power-comfort (or service level) constraints are entered into zonal rolling and global coordination in the same way. Communication and architecture can also be varied: cloud-based desensitization can be implemented using "full edge closed loop + local small-batch training," or cloud-edge collaboration using "centralized training + lightweight inference." Field protocols can be interchanged between BACnet / IP, Modbus-TCP, MQTT, etc. In addition to "retry-degradation-hot switching," the fault tolerance chain can also include primary / backup hot switching, bypass direct connection, and a local security policy whitelist. The prediction module is not limited to LSTM / GBRT; it can use a hybrid physical-statistical approach (grey-box RC network + Kalman filter), quantile regression, or temporal Bayesian methods. Measurement and settlement objectives can be expanded from "minimum cost" to a weighted or hierarchical objective of "cost + carbon emissions + demand." DR can interface with price-based, capacity-based, and event-based targets, all achieved through stage two constraints and objective parameterization.

[0080] The aforementioned substitutions only change the algorithm, model approximation, or device configuration, but do not alter the key technical path of this invention: partitioned rolling optimization is performed using E-LAN ​​as the smallest unit, network layer coordination is carried out at the current moment, and the execution data is written back to maintain phase consistency; simultaneously, power balancing, state corridors, capacity / ramp-up, and lifetime / loss are explicitly incorporated into the model, and robust / fault-tolerant mechanisms ensure executability. Therefore, these variations are all equivalent substitutions and extensions to this invention and fall within the protection scope of this invention.

[0081] According to the two-stage energy management method for zoned energy local area networks proposed in this application, the method acquires the operating status data of each energy local area network within the target area; determines the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network; calculates the network coordination of each energy local area network at the current moment based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment, and obtains the executable power command of each energy local area network based on the network coordination of each energy local area network at the current moment with the goal of minimizing interconnection power and line loss; and allocates energy to the corresponding energy local area network according to the executable power command of each energy local area network. This solves the problems of poor phase consistency, high cross-zone energy transport losses, and overdraft of energy storage life in multi-bus building energy management.

[0082] Next, referring to the accompanying drawings, a two-stage energy management device for a partitioned energy local area network according to an embodiment of this application is described.

[0083] Figure 8 This is a block diagram of a two-stage energy management device for a partitioned energy local area network according to an embodiment of this application.

[0084] like Figure 8 As shown, the two-stage energy management device 10 of the partitioned energy local area network includes: an acquisition module 100, a determination module 200, a calculation module 300, and an allocation module 400.

[0085] The acquisition module 100 is used to acquire the operating status data of each energy local area network within the target area; The determination module 200 is used to determine the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network; The calculation module 300 is used to calculate the network coordination of each energy local area network at the current time based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current time, and to obtain the executable power command of each energy local area network based on the network coordination of each energy local area network at the current time with the goal of minimizing interconnection power and line loss. The allocation module 400 is used to allocate energy to the corresponding energy LAN according to the executable power command of each energy LAN.

[0086] Optionally, the acquisition module 100 is specifically used to: acquire the current intraday rolling forecast data and the current measured data of each energy local area network in the target area based on a preset step size; and obtain the operating status data of each energy local area network based on the current intraday rolling forecast data and the current measured data of each energy local area network in the target area.

[0087] Optionally, the determining module 200 is specifically used for: constructing a first objective function for each energy local area network based on a preset first constraint; solving the first objective function based on the operating status data of each energy local area network to obtain the optimal decision sequence within the rolling window; and determining the local reference and / or zero interconnection tendency reference for each energy local area network at the current moment based on the optimal decision sequence.

[0088] Optionally, the first objective function is: ; in, For the current moment, The length of the scroll window. The electricity price at the current moment. For the first A local area network of energy The power of constant interaction with the mainnet. To preset the step size, This refers to the penalty coefficient for photovoltaic power generation restrictions. To limit PV power generation, This is the energy storage degradation cost coefficient. Let be the energy storage charging power of the k-th local energy network at time τ. For the first A local area network of energy Energy storage and discharge power at any given time For the first A local area network of energy The state of charge of the energy storage at any given time. For the bias penalty weight, This is the grid-connected execution power written back from the previous moment.

[0089] Optionally, the calculation module 300 is specifically used for: constructing a second objective function based on preset second constraints; solving the second objective function according to the network coordination of each energy local area network at the current time, and obtaining the executable power command of each energy local area network.

[0090] Optionally, the second objective function is: ; in, For the first The grid-connected power of a local area network. For the first The energy storage unit of a local area network performs power. For from the first The energy flow from the local area network to the first The power of the interconnecting lines in a local energy region. It is a set of interconnected edges between energy local area networks. The absolute value penalty factor for tie line power. This is the line loss coefficient. For the first The equivalent resistance of the link between the j-th energy local area network and the j-th energy local area network. For E-LAN ​​sets, The power deviation penalty factor is applied to the grid connection. For reference and interaction with the mainnet, The power deviation penalty factor is applied to energy storage. This serves as a reference for charging and discharging the energy storage unit.

[0091] Optionally, after allocating energy to the corresponding energy local area network according to the executable power command of each energy local area network, the allocation module 400 is further configured to: determine the write-back reference for solving the first objective function at the next moment based on the executable power command, and update the weight of the deviation penalty term when solving the local reference according to the deviation between the executable power command and the local reference.

[0092] It should be noted that the foregoing explanation of the two-stage energy management method embodiment for partitioned energy local area networks also applies to the two-stage energy management device for partitioned energy local area networks in this embodiment, and will not be repeated here.

[0093] According to the two-stage energy management device for zoned energy local area networks proposed in this application, the embodiment acquires the operating status data of each energy local area network within the target area; determines the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network; calculates the network coordination of each energy local area network at the current moment based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment, and obtains the executable power command of each energy local area network based on the network coordination of each energy local area network at the current moment with the goal of minimizing interconnection power and line loss; and allocates energy to the corresponding energy local area network according to the executable power command of each energy local area network. This solves the problems of poor phase consistency, high cross-zone energy transport losses, and overdraft of energy storage life in multi-bus building energy management.

[0094] Figure 9 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 901, the processor 902, and the computer program stored on the memory 901 and capable of running on the processor 902.

[0095] When the processor 902 executes the program, it implements the two-stage energy management method for partitioned energy local area networks provided in the above embodiments.

[0096] Furthermore, electronic devices also include: Communication interface 903 is used for communication between memory 901 and processor 902.

[0097] The memory 901 is used to store computer programs that can run on the processor 902.

[0098] The memory 901 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0099] If the memory 901, processor 902, and communication interface 903 are implemented independently, then the communication interface 903, memory 901, and processor 902 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0100] Optionally, in a specific implementation, if the memory 901, processor 902, and communication interface 903 are integrated on a single chip, then the memory 901, processor 902, and communication interface 903 can communicate with each other through an internal interface.

[0101] The processor 902 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0102] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described two-stage energy management method for a partitioned energy local area network.

[0103] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0104] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0105] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0106] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0107] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

Claims

1. A two-stage energy management method for a partitioned energy local area network, characterized in that, Includes the following steps: Acquire operational status data for each energy local area network within the target area; Determine the local reference and / or zero interconnection tendency reference for each energy local area network at the current moment based on the operational status data of each energy local area network; Based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment, the network coordination of each energy local area network at the current moment is calculated, and with the goal of minimizing interconnection power and line loss, the executable power command of each energy local area network is obtained according to the network coordination of each energy local area network at the current moment. Energy is allocated to the corresponding energy local area network according to the executable power command of each energy local area network.

2. The method according to claim 1, characterized in that, The acquisition of operational status data for each energy local area network within the target area includes: Based on a preset step size, the current intraday rolling forecast data and the current measured data of each energy local area network within the target area are obtained. The operating status data of each energy local area network is obtained based on the intraday rolling forecast data and the measured data at the current moment for each energy local area network within the target area.

3. The method according to claim 1 or 2, characterized in that, The process of determining the local reference and / or zero interconnection tendency reference of each energy local area network (ELAN) at the current moment based on the operational status data of each ELAN includes: Construct a first objective function for each of the energy local area networks based on a preset first constraint condition; The first objective function is solved based on the operating status data of each energy local area network to obtain the optimal decision sequence within the rolling window; Based on the optimal decision sequence, determine the local reference and / or zero interconnection tendency reference for each of the energy local area networks at the current moment.

4. The method according to claim 3, characterized in that, The first objective function is: ； in, For the current moment, The length of the scroll window. The electricity price at the current moment. For the first A local area network of energy The power of constant interaction with the mainnet. To preset the step size, This refers to the penalty coefficient for photovoltaic power generation restrictions. To limit PV power generation, This is the energy storage degradation cost coefficient. Let be the energy storage charging power of the k-th local energy network at time τ. For the first A local area network of energy Energy storage and discharge power at any given time For the first A local area network of energy The state of charge of the energy storage at any given time. For the bias penalty weight, This is the grid-connected execution power written back from the previous moment.

5. The method according to claim 1, characterized in that, The step of obtaining the executable power command for each energy local area network based on the network coordination of each energy local area network at the current time includes: Construct a second objective function based on the preset second constraints; Based on the network coordination of each energy local area network at the current time, the second objective function is solved to obtain the executable power command for each energy local area network.

6. The method according to claim 5, characterized in that, The second objective function is: ； in, For the first The grid-connected power of a local area network. For the first The energy storage unit of a local area network performs power. For from the first The energy flow from the local area network to the first The power of the interconnecting lines in a local energy region. It is a set of interconnected edges between energy local area networks. The absolute value penalty factor for tie line power. This is the line loss coefficient. For the first The equivalent resistance of the link between the j-th energy local area network and the j-th energy local area network. For E-LAN ​​sets, The power deviation penalty factor is applied to the grid connection. For reference and interaction with the mainnet, The power deviation penalty factor is applied to energy storage. This serves as a reference for charging and discharging the energy storage unit.

7. The method according to claim 1, characterized in that, After allocating energy to the corresponding energy local area network according to the executable power command of each energy local area network, the method further includes: The writeback reference for solving the first objective function at the next moment is determined based on the executable power instruction, and the weight of the deviation penalty term when solving the local reference is updated according to the deviation between the executable power instruction and the local reference.

8. A two-stage energy management device for a zoned energy local area network, characterized in that, include: The acquisition module is used to acquire the operating status data of each energy local area network within the target area; The determination module is used to determine the local reference and / or zero interconnection tendency reference of each energy local area network at the current moment based on the operating status data of each energy local area network; The calculation module is used to calculate the network coordination of each energy local area network at the current time based on the local reference and / or zero interconnection tendency reference of each energy local area network at the current time, and to obtain the executable power command of each energy local area network based on the network coordination of each energy local area network at the current time with the goal of minimizing interconnection power and line loss. The allocation module is used to allocate energy to the corresponding energy local area network according to the executable power command of each energy local area network.

9. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the two-stage energy management method for a partitioned energy local area network as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the two-stage energy management method for partitioned energy local area networks as described in any one of claims 1-7.

Images