A data center energy dispatching method, system, and electronic equipment

Abstract

This invention discloses an energy dispatching method, system, and electronic device for a data center, relating to the field of data processing technology. The method includes the following steps: acquiring battery cell operating data in a solar power plant energy storage system and extracting energy accumulation nodes based on the battery cell operating data; identifying the current loop extension direction of the energy accumulation nodes based on a preset wire connection relationship to obtain an energy echo zone; cutting off the echo channel corresponding to the energy echo zone according to a preset cutting frequency to obtain a first cutting zone; extracting the residual current attenuation characteristics of the first cutting zone and restoring the channel of the first cutting zone based on the residual current attenuation characteristics to obtain a second cutting zone; and dynamically adjusting the timing path of the internal energy flow of the battery cell based on the second cutting zone. This solves the problem of local energy accumulation during the charge-discharge conversion process of the energy storage system, which leads to current oscillations in some nodes.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and more specifically, to a data center energy scheduling method, system, and electronic equipment. Background Technology

[0002] With the widespread application of renewable energy, solar power plants, as an important form of green energy, have been widely deployed globally. However, solar power generation is characterized by intermittency and instability; its output is affected by factors such as weather, season, and time, making it difficult to achieve a stable power supply. Therefore, how to rationally dispatch energy in solar power plants and achieve a balance between power generation, energy storage, and load has become a crucial issue in the operation of solar power plants.

[0003] However, the topology of existing solar energy storage devices is usually fixed during the design phase. Their wire connections, switch matrix layout, and current flow paths are all statically constructed, lacking the ability to dynamically adapt during operation. When local energy accumulation occurs in the energy storage unit during charge-discharge conversion, the current often flows repeatedly along a predetermined fixed path, failing to promptly dissipate energy through structural reconfiguration or path diversion, thus creating a local loop echo phenomenon. When this echo effect persists, local nodes experience current oscillations. Even if the energy accumulation is partially eliminated, the oscillations continue due to factors such as wire impedance, parasitic capacitance, and phase delay, preventing complete energy release. Prolonged energy echoes not only reduce energy storage conversion efficiency but also cause node temperature rise, increased local losses, and disturbances to overall current stability. To address these problems, this invention proposes a solution. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an energy scheduling method, system, and electronic device for a data center. By regulating the energy echo region of battery cells, the problem of current oscillations occurring at some nodes due to local energy accumulation during the charging and discharging conversion process of the energy storage system is solved.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An energy scheduling method for a data center includes the following steps: acquiring battery cell operation data in a solar power station energy storage system, and extracting energy accumulation nodes based on the battery cell operation data; identifying the current loop extension direction of the energy accumulation nodes based on a preset wire connection relationship to obtain an energy echo zone; cutting off the echo channel corresponding to the energy echo zone according to a preset cutting-off frequency to obtain a first cutting-off zone; extracting the residual current attenuation characteristics of the first cutting-off zone, and restoring the channel of the first cutting-off zone based on the residual current attenuation characteristics to obtain a second cutting-off zone; and dynamically adjusting the timing path of the internal power flow of the battery cell based on the second cutting-off zone.

[0007] In a preferred embodiment, the step of extracting energy accumulation nodes based on battery cell operating data specifically involves: extracting multi-channel voltage and current sequences of battery cells based on battery cell operating data to obtain a battery operating state matrix; extracting the charge accumulation rate and discharge fluctuation amplitude of each battery cell based on the battery operating state matrix; calculating the energy difference per unit time based on the charge accumulation rate and discharge fluctuation amplitude, and identifying initial energy accumulation points based on the energy difference; obtaining the physical connection topology between battery cells, and performing cluster analysis on the initial energy accumulation points based on the physical connection topology to obtain energy accumulation regions; evaluating the energy density of the energy accumulation regions, and filtering the initial energy accumulation points based on the evaluation results to obtain energy accumulation nodes.

[0008] In a preferred embodiment, the step of identifying the current loop extension direction of the energy accumulation node based on the preset conductor connection relationship to obtain the energy echo zone specifically involves: acquiring the conductor connection topology of the solar power station energy storage system, and traversing along the forward and reverse directions of the conductor connection topology starting from the energy accumulation node to obtain several current flow paths; calculating the impedance value and current phase delay value of each current flow path, and identifying the closed-loop current loop based on the impedance value and current phase delay value; calculating the active power density vector between the nodes of the closed-loop current loop, and obtaining the main loop and branch loops of the closed-loop current loop based on the active power density vector; superimposing the energy accumulation nodes of the main loop to obtain the energy echo core region; and expanding the boundary of the energy echo core region based on the branch loops to obtain the energy echo zone.

[0009] In a preferred embodiment, the step of extending the boundary of the energy echo core region based on the branch extension direction to obtain the energy echo region specifically involves: extracting the current oscillation characteristics of the branch circuits and performing cluster analysis on the branch circuits based on the current oscillation characteristics to obtain oscillation clusters; extending the boundary along the oscillation cluster direction with the energy echo core region as the center to obtain a first extension zone; calculating the avoidance radius based on the first extension zone and adjusting the boundary of the energy echo core region according to the avoidance radius to obtain the energy echo region.

[0010] In a preferred embodiment, the step of cutting off the echo channel corresponding to the energy echo zone according to a preset cutoff frequency to obtain the first cutoff zone specifically involves: acquiring the echo channel within the energy echo zone, and classifying the channel according to preset harmonic components based on the current amplitude and frequency characteristics of the echo channel to obtain the classified echo channel; acquiring the harmonic resonance characteristics of the classified echo channel, and preset several cutoff frequency points based on the harmonic resonance characteristics; and sequentially cutting off each classified echo channel according to the preset cutoff frequency points to obtain the first cutoff zone.

[0011] In a preferred embodiment, the step of extracting the residual current decay characteristics of the first cut-off region and performing channel recovery on the first cut-off region based on the residual current decay characteristics to obtain the second cut-off region specifically involves: acquiring the residual current waveform and current response curve of the first cut-off region, and performing exponential fitting on the residual current waveform to obtain the residual decay time; calculating the residual energy dissipation rate based on the residual decay time, identifying recoverable channels based on the dissipation rate distribution map; adjusting the rise rate of the voltage excitation according to the slope change of the current response curve, and applying voltage excitation to the recoverable channels to obtain the second cut-off region.

[0012] In a preferred embodiment, the step of dynamically adjusting the timing path of the internal power flow of the battery cell based on the second cut-off zone specifically involves: acquiring the three-dimensional coordinates of the second cut-off zone to obtain the spatial distribution matrix of the second cut-off zone; reconstructing the preset energy allocation matrix of the matrix battery cell based on the spatial distribution of the second cut-off zone to obtain the reconstructed energy allocation matrix; allocating a variable-length time window to each battery cell based on the reconstructed energy allocation matrix and collecting voltage ripple and total harmonic distortion (THD) data of the current in real time; and dynamically calibrating the start delay and duration of the timing path according to the voltage ripple and harmonic data to obtain the optimized power flow timing path set.

[0013] The technical effects and advantages of the energy scheduling method, system, and electronic equipment for data centers disclosed in this invention are as follows:

[0014] This invention extracts energy accumulation nodes to accurately identify concentrated energy areas in space. Then, based on preset wire connection relationships, it determines the extension direction of the current loop, constructing an energy echo zone to reveal the cyclic echo effect of current in a complex wire network. Next, the channels within the energy echo zone are cut off at a preset frequency to form a first cutoff zone, interrupting high-frequency energy return. Based on this, the attenuation characteristics of the residual current are extracted, and voltage excitation is applied to recoverable channels to obtain a second cutoff zone, achieving gradual recovery and optimization of the energy loop. Finally, based on the second cutoff zone, the temporal path of energy flow within the battery cell is dynamically adjusted, enabling orderly energy distribution and temporal balance among different cells. Through these steps, not only are the energy echo and high-frequency oscillation problems caused by local accumulation effectively eliminated, but the efficiency and stability of energy transmission are also improved, achieving synergistic optimization of the energy storage network in both spatial and temporal dimensions, and significantly enhancing the intelligent scheduling and adaptive operation capabilities of data centers in complex power environments. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating an energy scheduling method for a data center according to the present invention.

[0016] Figure 2 This is a schematic diagram of the structure of an energy dispatching system for a data center according to the present invention.

[0017] Figure 3 This is a schematic diagram of the electronic device structure of an energy scheduling method for a data center according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1, Figure 1 The present invention provides an energy scheduling method for a data center, comprising the following steps:

[0020] S1, acquire the battery unit operation data in the solar power station energy storage system, and extract the energy accumulation node based on the battery unit operation data;

[0021] In this example, energy accumulation nodes are extracted based on battery cell operating data, specifically as follows:

[0022] Based on the battery cell operation data, multi-channel voltage and current sequences of the battery cells are extracted to obtain the battery operation state matrix;

[0023] Based on the battery operating state matrix, the charge accumulation rate and discharge fluctuation amplitude of each battery cell are extracted.

[0024] The energy difference per unit time is calculated based on the charge accumulation rate and discharge fluctuation amplitude, and the initial energy accumulation point is identified based on the energy difference.

[0025] The physical connection topology between battery cells is obtained, and cluster analysis is performed on the initial energy accumulation points based on the physical connection topology to obtain the energy accumulation region;

[0026] The energy density of the energy accumulation zone is assessed, and the initial energy accumulation points are screened based on the assessment results to obtain energy accumulation nodes.

[0027] It's important to note that in solar energy storage scenarios, each battery cell outputs multi-dimensional operational data in real time, such as terminal voltage, current, temperature, charge / discharge status indicators, and SOC (State of Charge). To extract multi-channel voltage and current sequences, the raw operational data within a certain sampling period (e.g., once per second or once per millisecond) needs to be obtained from the battery management device. Then, the voltage and current data of each battery cell at different sampling times are recorded separately for each channel. For example, within a 10-minute monitoring period, a voltage-time sequence and a current-time sequence can be formed. After time alignment, noise filtering, and normalization, these sequences can be used to construct a multi-dimensional matrix composed of time, channel, and energy parameters. This matrix is ​​called the battery operating state matrix. It intuitively reflects the energy input and output behavior of each battery cell throughout the entire sampling period. For example, in an array of 100 battery cells, each cell corresponds to one row of time series data, and the columns of the matrix correspond to signal channels such as voltage, current, and temperature.

[0028] After obtaining the battery operating state matrix, features can be extracted from the trends of voltage and current changes over time. The charge accumulation rate can be understood as the rate of energy accumulation reflected by the increasing trend of current over time during the charging phase. Specifically, the current sequence is segmented into time windows, and the rising trend or stability of the current within each time period is calculated. This is then combined with the average level of voltage changes to reflect the charge accumulation rate of the battery cell during this phase. The discharge fluctuation amplitude corresponds to the fluctuation of current output during the discharge phase, which can be obtained by analyzing the range of fluctuations in the current curve and the rate of voltage drop during the discharge period. For example, if a battery cell experiences a stable voltage drop and small current fluctuations within 10 seconds, it indicates stable energy output; if the current fluctuations are severe, it means that its discharge characteristics are uneven and there may be local energy accumulation. Ultimately, each battery cell will obtain two characteristic parameters: one representing the energy accumulation capacity (charge accumulation rate), and the other representing the energy release stability (discharge fluctuation amplitude).

[0029] Furthermore, once the charge accumulation rate and discharge fluctuation amplitude of each battery cell are known, their energy changes per unit time can be compared. Simply put, this involves comparing the difference between the "energy absorbed" and "energy released" by the battery within the same time period. If a battery cell exhibits a high accumulation rate and low discharge stability over multiple consecutive time periods—that is, fast charging but slow discharging with significant energy retention—this cell may form a localized energy concentration phenomenon. Such cells are identified as initial energy accumulation points. For example, in an energy storage module composed of 100 batteries, if batteries numbered 45 and 46 have energy input greater than output and large discharge curve fluctuations over multiple time periods, they will be identified as initial energy accumulation points. These points are often the source of subsequent energy imbalances.

[0030] Secondly, battery cells are typically connected in series and parallel to form an overall energy storage array, so the connection relationship of each cell can be clearly defined during the design phase. To understand the spatial distribution of energy in the array, it is necessary to first extract the electrical connection relationships between the cells from the design drawings or monitoring platform, such as which cells are adjacent, which are on the same branch, and which are at the same voltage level. Then, combining these connection relationships with the initially identified energy accumulation points, clustering algorithms (such as hierarchical clustering based on adjacency relationships and energy feature similarity) are used to group multiple spatially related accumulation points with similar energy difference trends together. These aggregated regions are defined as energy accumulation zones. In layman's terms, an energy accumulation zone refers to a local area in the energy storage array where multiple battery cells simultaneously exhibit the common characteristics of energy accumulation, insufficient discharge, or overcharging.

[0031] Finally, after identifying the energy accumulation zones, it is necessary to further determine which areas have energy concentration levels reaching the threshold for dispatch intervention. This can be accomplished by calculating the energy density within that area, i.e., the degree of energy accumulation or total residual energy per unit number of batteries within the same time period. The evaluation considers comprehensive indicators of voltage, current, and temperature to reflect the compactness of the energy distribution. Areas with high energy density and rapid change rates indicate severe local energy stagnation. Subsequently, based on the evaluation results, the initial accumulation points within the area are screened, eliminating units with slight fluctuations or small energy differences, retaining only those points with the highest energy concentration and the greatest impact on the overall energy flow; these points are the energy accumulation nodes.

[0032] S2, based on the preset wire connection relationship, identify the current loop extension direction of the energy accumulation node to obtain the energy echo zone;

[0033] In this example, the current loop extension direction of the energy accumulation node is identified based on the preset wire connection relationship to obtain the energy echo region, specifically:

[0034] Obtain the topology of the conductor connection of the solar power station energy storage system, and start from the energy accumulation node, traverse the conductor connection topology in both the forward and reverse directions to obtain several current flow paths.

[0035] Calculate the impedance and current phase delay for each current flow path, and identify the closed-loop current loop based on the impedance and current phase delay.

[0036] Calculate the active power density vector between nodes of the closed-loop current loop, and obtain the main loop and branch loop of the closed-loop current loop based on the active power density vector;

[0037] The energy accumulation nodes of the main loop are superimposed to obtain the energy echo core region;

[0038] The boundary of the energy echo core region is extended by branching loops to obtain the energy echo region.

[0039] It's important to note that in the energy storage devices of solar power plants, numerous battery cells are interconnected in series and parallel via wires. These connections are typically documented in electrical topology files or wiring diagrams during the design phase. A wire connection topology diagram is a graphical representation of the interconnections of all battery cells, busbars, connecting wires, and measurement nodes in the energy storage device, abstracted into a network structure. Each node represents an electrical component (e.g., a battery cell or busbar), and each connection line represents an electrical connection of a segment of wire. This topology diagram clearly describes the possible energy transmission paths throughout the entire energy storage architecture. There are generally two ways to obtain this topology diagram: one is to directly export a digitized connection table from the engineering design drawings; the other is to identify the electrical connection paths using online monitoring equipment (e.g., voltage sampling modules and busbar monitoring points). After obtaining the topology diagram, starting from the identified energy accumulation nodes, the process expands outward layer by layer along the wire connections. On one hand, it traverses forward along the main current flow direction (i.e., the energy output path); on the other hand, it traverses backward against the flow direction (i.e., the energy return path). By traversing in both forward and reverse directions, multiple complete current flow paths can be obtained. For example, starting from the energy accumulation node numbered 45, the current flows through cells 46, 47, and 48 in series before reaching the junction point, forming a forward path; while the reverse path may return to cell 44 or 43 via parallel branches. All these traversal paths are recorded, forming the basis for subsequent current analysis.

[0040] After obtaining multiple current flow paths, it is necessary to evaluate the energy conduction characteristics of these paths. Impedance values ​​reflect the overall obstruction of current flow by the conductors and their connections, typically depending on the conductor length, material, and contact quality. Current phase delay values ​​describe the time lag or phase difference of current propagation within the path, reflecting effects such as inductance and capacitance. By acquiring voltage and current signals at different nodes, the equivalent impedance and phase delay characteristics of each path can be calculated. Next, a correlation analysis of the impedance and phase delay of all paths is performed. If several paths are found to be connected end-to-end with a periodic closed-loop phase delay, it indicates that energy circulates between these paths, forming a so-called closed-loop current circuit. In simple terms, this circuit is like a channel through which energy repeatedly circulates within a certain area. For example, in three parallel branches, one branch may form a partial circulating current due to its low impedance, causing current to flow back and forth between that branch and adjacent conductors. This phenomenon is identified as a closed-loop current circuit in the topology diagram. Such circuits are often potential areas of energy retention and heat generation.

[0041] Furthermore, after identifying the closed-loop current circuit, it is necessary to further analyze the energy distribution between each node. By comprehensively analyzing the voltage, current, and phase at each node, the direction and intensity of energy transfer between each pair of adjacent nodes can be obtained. This quantity, expressed in terms of direction and amplitude, is called the active power density vector. It reflects the actual flow trend of energy in the circuit. For example, if the power density on certain sections of a closed loop is significantly higher than that on other sections, it indicates that these sections are the main energy transmission channels. Based on these vector distributions, the closed-loop structure can be divided into two categories: one is the trunk circuit, which is the main path where energy flow is most concentrated, usually carrying the majority of the active power in the circuit, where energy flow is most stable and lasts the longest; the other is branch circuits, which are connected to the trunk circuit but have lower power density and often only participate in energy transfer under specific conditions (such as transient load changes). For example, if the energy flow is strongest along the path from node 45-46-47-48 in a closed loop, while the flow is weaker along the branch from node 47 (47-49-50), then the former is called the main loop, and the latter is called the branch loop. The main loop is responsible for the main energy transmission, while the branch loop has a greater impact on the local energy distribution balance.

[0042] Finally, after identifying the main closed-loop circuit, we can trace back to the energy accumulation nodes distributed along these main circuits. These nodes are originally locations where energy is concentrated or stagnant, and when they happen to be located on the main energy flow path, they will generate a feedback effect on the entire energy conduction. By superimposing the energy characteristics of all energy accumulation nodes on the main circuit, we can determine the core region where energy is most concentrated and energy reflection and redistribution phenomena are most significant; this region is called the energy echo core region. The energy echo core region can be understood as the spatial range in the energy conduction network where current and power signals repeatedly fluctuate and feedback within a local area due to the coupling effect of energy accumulation nodes. Here, the echo is the phenomenon of multiple reflections and repropagation of electrical energy within this region. For example, in a certain energy storage architecture, if batteries 45, 46, and 47 are all energy accumulation nodes located on the main circuit, their energy interaction will lead to enhanced local current oscillations, forming a concentrated area of ​​energy echo.

[0043] In this example, the boundary of the energy echo core region is extended based on the branch extension direction to obtain the energy echo region, specifically:

[0044] The current oscillation characteristics of the branch circuits are extracted, and cluster analysis is performed on the branch circuits based on the current oscillation characteristics to obtain oscillation clusters;

[0045] Centered on the energy echo core region, the boundary is extended along the direction of the oscillation cluster to obtain the first extension zone;

[0046] The avoidance radius is calculated based on the first extension zone, and the boundary of the energy echo core region is adjusted according to the avoidance radius to obtain the energy echo region.

[0047] It should be noted that after identifying the closed-loop current circuit, several branch circuits exist outside the main branch. These branches often play a role in local energy regulation and current distribution. To further analyze the impact of these branches on energy distribution, it is necessary to extract their current oscillation characteristics. Specifically, time-series analysis is performed on the current signals at each node in multiple branch circuits, recording characteristic parameters such as the fluctuation range of current amplitude, oscillation frequency, phase drift, and energy decay trend. These characteristics can be used to determine whether a branch circuit exhibits periodic energy back-and-forth flow. For example, if a branch shows repeated changes in current direction and significant amplitude fluctuations within a short period, it indicates strong energy oscillation behavior. Next, after vectorizing the oscillation characteristics of all branch circuits, similarity analysis or clustering algorithms are used to group branches with similar oscillation characteristics into the same group. These groups of branches with the same oscillation mode, similar frequency, and energy reflection characteristics are called oscillation clusters. An "oscillation cluster" actually represents a group of current oscillation channels exhibiting consistent energy flow behavior; they are often located in spatially adjacent or electrically coupled regions. For example, if branches 47-49-50 and 46-48-51 in a certain energy storage unit exhibit synchronous fluctuations at similar frequencies, they will be classified into the same oscillation cluster. The significance of this concept lies in its revelation of the dynamic region where energy between branches is mutually coupled and jointly influences the direction of energy propagation.

[0048] Furthermore, after obtaining the oscillation clusters, it is necessary to assess the energy interaction between these oscillation groups and the energy echo core region. The energy echo core region can be considered as the center of strongest energy reflection, while the oscillation clusters often act as bridges for energy propagation from the core to the periphery. Therefore, the energy echo core region is first designated as the center point or center plane and its position is marked in three-dimensional spatial coordinates. Then, based on the average oscillation direction, energy propagation tendency, and coupling strength with the core region of each oscillation cluster, the spatial boundary of the core region is gradually expanded along these directions. This expansion is not a simple geometric expansion, but a directional extension based on the oscillation energy propagation path. For example, if the current phase of an oscillation cluster is strongly coupled with the core region, it indicates a high probability of energy propagation in that direction, thus expanding the boundary coverage in that direction; conversely, if the oscillation amplitude is small or inconsistent with the phase of the core region, the expansion is less. After this gradual expansion, an energy influence band naturally extends around the core region along the oscillation direction. This region is called the first extension band, representing the first layer of envelope region for the outward propagation of energy from the core region, like a transitional circle where an energy wavefront diffuses outward. For example, in an energy storage array, if the core area covers units 45-47, and the branch units 49-50 connected to unit 47 belong to the same oscillation cluster, then the first extension zone will cover the area of ​​49-50, forming a strip structure for energy conduction from the core outward.

[0049] Finally, after determining the first extension zone, it is necessary to consider the overlap, interference, and safety distance of energy propagation in different directions. To this end, a parameter called the avoidance radius needs to be calculated to determine the minimum distance between different regions during energy propagation, in order to avoid oscillation superposition or excessive energy concentration. Specifically, by analyzing the energy density distribution and oscillation phase difference in each direction within the first extension zone, regions with strong energy interference can be identified, and then a reasonable spatial avoidance range can be determined based on the energy fluctuation amplitude. The size of the avoidance radius depends on the energy coupling strength and spatial distribution density. For example, in regions with severe energy fluctuations, the avoidance radius is larger; while in directions with faster energy attenuation, the avoidance radius is smaller. Next, based on this avoidance radius, the original boundary of the energy echo core region is fine-tuned or deformed so that it spatially covers the main energy propagation direction while maintaining a moderate separation from high-interference regions. The new boundary region formed in this way is called the energy echo zone. This energy echo zone is equivalent to a dynamic energy interaction spatial range, which not only includes the original energy echo core region but also encompasses the transition zone between energy propagation and attenuation. It reflects the actual spatial distribution of energy within the energy storage structure during the processes of recirculation, reflection, attenuation, and rebalancing. From a practical perspective, for example, if the core area is concentrated in units 45-47, while the first extension zone extends to units 48-50, after calculating the avoidance radius, the final energy echo zone may cover the area of ​​units 45-50. This ensures complete capture of the energy propagation direction while avoiding the superposition of interference from different oscillation directions.

[0050] S3, cut off the echo channel corresponding to the energy echo zone according to the preset cutoff frequency to obtain the first cutoff zone;

[0051] In this example, the echo channel corresponding to the energy echo zone is cut off according to a preset cutoff frequency to obtain the first cutoff zone, specifically:

[0052] The echo channels within the energy echo zone are obtained, and the channels are classified according to preset harmonic components based on the current amplitude and frequency characteristics of the echo channels to obtain the classified echo channels.

[0053] Obtain the harmonic resonance characteristics of the classified echo channels, and preset several cutoff frequency points based on the harmonic resonance characteristics;

[0054] For each classified echo channel, the channel is cut off sequentially according to the preset cutoff frequency point to obtain the first cutoff zone.

[0055] It is important to note that after identifying the energy echo zone, further analysis of the details of energy flow within this zone is necessary. Energy echo zones typically contain multiple energy propagation paths, where current reflection, fluctuations, and energy back-and-forth propagation phenomena occur. To characterize the specific trajectories of these energy flows, echo channels need to be extracted. An echo channel refers to a specific conduction path formed within the echo zone due to current reflection and harmonic interference; it manifests as the periodic back-and-forth energy transmission trajectory of current between wires, nodes, or battery cells. In other words, an echo channel is the actual path through which energy repeatedly propagates within the echo zone. Obtaining these echo channels involves two steps: first, by monitoring the current waveforms and phase changes at each node within the echo zone, current paths exhibiting significant reflection characteristics or periodic fluctuations in time are identified; second, these paths are abstracted into traceable channel data structures for subsequent analysis. After obtaining the echo channels, they need to be classified according to their current amplitude and frequency characteristics. The specific approach involves analyzing the current amplitude changes and corresponding frequency distribution of each channel over a period of time to extract its main harmonic components, such as the fundamental frequency, second harmonic, and third harmonic. Then, according to a pre-defined harmonic classification standard, the channels are divided into several groups, such as "low-frequency high amplitude," "mid-frequency mid-amplitude," and "high-frequency low amplitude." For example, in an energy echo zone, channels numbered 45-46-48 might exhibit low-frequency, large current fluctuations, while channels 47-49-50 might exhibit high-frequency, low-amplitude oscillations; thus, they would be classified into different categories. The classification results provide a basis for subsequent targeted disconnection and energy reconstruction.

[0056] Furthermore, after the echo channels are classified, the next step is to analyze the harmonic resonance characteristics of different categories of channels. Harmonic resonance characteristics refer to the phenomenon where energy is amplified, superimposed, or delayed in dissipation within a channel at a specific frequency. This resonance usually originates from the matching relationship between conductor length, inductance, and capacitance characteristics. To obtain resonance characteristics, frequency analysis of the current waveform is required for each classified echo channel, observing the energy response intensity and phase stability of each frequency band. If the energy amplitude of a certain frequency band is significantly higher than the surrounding frequency bands and persists, it can be determined as the resonance point of that channel. Through the analysis of multiple channels, the resonance patterns of various channels in different frequency bands can be summarized. Based on these patterns, a set of cutoff frequency points is preset in engineering. These frequency points represent the frequency bands most likely to induce harmonic accumulation or energy backflow during energy conduction. The principle for selecting these points is to cover the high-risk harmonic concentration frequency bands while avoiding interference with normal energy flow. For example, in an energy storage architecture, it may be found that energy transmission is normal in the low-frequency band (such as around 50Hz), while energy resonance is obvious in the higher-frequency band (such as 250Hz, 400Hz). In this case, 250Hz and 400Hz will be used as cutoff frequencies. These frequencies will serve as control references for subsequent cutoff operations, used to orderly interrupt or isolate energy reflections in the corresponding channels, thereby weakening the energy accumulation effect in the echo zone.

[0057] Finally, after obtaining the cutoff frequency points, it is necessary to intervene in an orderly manner in various echo channels. Channel cutoff does not mean physically disconnecting the wires, but rather suppressing or interrupting energy propagation at specific frequencies by adjusting the current transmission path or control signals within the channel. The specific steps are as follows: for each type of echo channel, cutoff operations are performed sequentially from low to high frequencies according to the preset cutoff frequency points. During cutoff, electronic switches, filtering modules, or control signals are used to adjust and obstruct energy flow at the corresponding frequency. For example, in the aforementioned example, if a channel experiences harmonic resonance at 250Hz, cutoff control is triggered at that frequency, preventing energy from propagating back along this frequency path; if the same channel also resonates at 400Hz, a second cutoff is performed. After cutoff, the regions where energy propagation is restricted within the cutoff frequency range but still retains some normal low-frequency energy flow form the so-called first cutoff zone. The first cutoff zone can be understood as the region within the energy echo zone where preliminary energy channel suppression has been completed; it is equivalent to an energy band layer isolated by frequency. Here, high-frequency energy reflection is blocked, while low-frequency energy can still be transmitted normally, thus achieving an initial weakening of energy oscillations. For example, if, in paths 45-48 within the energy echo zone, after cutting off 250Hz and 400Hz, the high-frequency harmonic energy decreases significantly while the dominant current continues to transmit smoothly, then this region is designated as the first cutoff zone.

[0058] S4, extract the residual current attenuation characteristics of the first cut-off region, and perform channel recovery on the first cut-off region based on the residual current attenuation characteristics to obtain the second cut-off region;

[0059] In this example, the residual current attenuation characteristics of the first cutoff region are extracted, and channel recovery is performed on the first cutoff region based on the residual current attenuation characteristics to obtain the second cutoff region, specifically:

[0060] The residual current waveform and current response curve of the first cut-off zone are acquired, and the residual current waveform is subjected to exponential fitting to obtain the residual decay time.

[0061] The residual energy dissipation rate is calculated based on the residual decay time, and recoverable channels are identified based on the dissipation rate distribution map.

[0062] The rise rate of the voltage excitation is adjusted according to the slope change of the current response curve, and the voltage excitation is applied to the recoverable channel to obtain the second cut-off zone.

[0063] It should be noted that although the high-frequency harmonic channels are isolated within the first cutoff zone, a portion of the current signal remains that has not completely dissipated; this is called residual current. It typically manifests as a current waveform that slowly decays or oscillates over a short period, indicating that local energy has not been fully released. To analyze these residual phenomena, high-precision sampling devices are needed to collect current waveform data from each channel in real time within the first cutoff zone, while simultaneously recording the current response curve when a small excitation is applied after cutoff. The current waveform reflects the instantaneous energy decay trend, while the current response curve reflects the channel's response speed and stability to external excitation. Next, the collected residual current waveforms are subjected to time series analysis to observe their gradual decay over time. Typically, this decay is not linear but exhibits a rapid initial decay followed by a slower energy release process. To quantitatively describe this characteristic, the decay process is approximated as an exponential decay model through data fitting, thereby extracting a key time parameter representing the decay rate; this parameter is called residual decay time. Residual decay time refers to the time elapsed from the moment of cutoff until its amplitude decreases to a stable range; it reflects the rate at which energy naturally dissipates within the cutoff channel. For example, if the current in channels 45-46 rapidly approaches zero within 3 seconds after being disconnected, it indicates a short residual decay time and sufficient energy release. However, if the current in channels 47-8 does not stabilize completely after 15 seconds, the residual decay time is longer, indicating significant energy retention within the channels. This time indicator is a crucial basis for subsequently determining which channels can be restored and which still require isolation.

[0064] Furthermore, once the residual decay time of each channel is determined, the dissipation rate of residual energy can be calculated. The dissipation rate represents the speed at which residual energy is released or converted per unit time; it reflects the energy recovery potential of a channel in a disconnected state. A shorter residual decay time indicates faster energy release and makes it easier for the channel to return to normal conduction. Conversely, a longer decay time indicates energy retention and makes recovery temporarily difficult. To present these characteristics more intuitively, the dissipation rates of all channels are mapped onto a dissipation rate distribution map using spatial coordinates. This map shows the distribution of energy decay rates at different locations. For example, if a high dissipation rate region appears between battery cells 45-47, while the 48-50 section is a low-rate region, it means that the former has released energy more fully, while the latter still has local accumulation. Then, through clustering or hierarchical analysis of the distribution map, a group of channels with sufficient energy dissipation, stable oscillations, and low recovery risk can be identified; these are called recoverable channels. In layman's terms, recoverable channels are those that have safely released energy after being disconnected and are capable of re-establishing electrical conduction. For example, out of 50 channels, if 20 of them exhibit stable decay over a short period and their response curves show no abnormal fluctuations, these 20 channels can be marked as recoverable channels, providing target objects for subsequent recovery operations.

[0065] Finally, after identifying the recoverable channels, the next step is to gradually restore their energy flow state. During the restoration process, it is necessary to control the application of voltage excitation to avoid sudden changes that could cause new energy surges. Specifically, the slope of the current response curve for each recoverable channel is analyzed. The slope represents the channel's sensitivity to voltage excitation; a larger slope indicates a rapid response and higher risk during energy transfer, while a smaller slope indicates a more stable channel. Based on these slope characteristics, the rise rate of the voltage excitation is adjusted. For channels with sensitive responses, a slow-rising voltage excitation is used to gradually restore conduction; for channels with a smooth response, the voltage rise rate can be appropriately accelerated. This differentiated control effectively prevents new echoes or overshoots during the restoration phase. After voltage excitation of the recoverable channels is completed, these channels gradually transition from a completely off state to a controlled conduction state, and energy re-establishes a stable flow path between them. At this point, the first off-state, after channel restoration, evolves into a new working area, called the second off-state. The main difference between the second and first off-states is as follows:

[0066] The first cutoff zone is the energy isolation stage, whose main goal is to block high-frequency harmonics and energy backflow, causing energy to temporarily stagnate and dissipate. The second cutoff zone is the energy recovery stage, whose main goal is to re-establish a balanced electrical energy path through precise voltage excitation, allowing local energy to participate in overall energy scheduling again. For example, assuming that channels 45-47 in the first cutoff zone are determined to be recoverable after dissipation analysis, and after applying a slow voltage excitation, the current gradually returns to a stable waveform, and the energy flow in this region is re-established, then region 45-47 becomes part of the second cutoff zone.

[0067] S5, dynamically adjusts the timing path of electrical energy flow inside the battery cell based on the second cut-off zone.

[0068] In this example, the timing path of the internal electrical energy flow of the battery cell is dynamically adjusted based on the second cut-off zone, specifically as follows:

[0069] Obtain the three-dimensional coordinates of the second cut zone to obtain the spatial distribution matrix of the second cut zone;

[0070] The preset energy distribution matrix of the matrix battery cell is reconstructed based on the spatial distribution of the second cut-off region to obtain the reconstructed energy distribution matrix.

[0071] Based on the reconstructed energy allocation matrix, a variable-length time window is allocated to each battery cell and voltage ripple and total harmonic distortion data of current are collected in real time.

[0072] The start-up delay and duration of the timing path are dynamically calibrated based on voltage ripple and harmonic data to obtain an optimized set of power flow timing paths.

[0073] It should be noted that after the voltage excitation of the recoverable channels is completed, local energy flow paths have been re-established within the second cutoff zone. At this point, in order to further optimize overall energy dispatch, it is necessary to clarify the spatial distribution and interrelationships of these channels. Obtaining the three-dimensional coordinates of the second cutoff zone involves collecting and mapping the positional data (such as X, Y, and Z coordinates) of each recovery channel, node, and battery cell in the actual physical layout. Specifically, this is done by identifying which channels belong to the second cutoff zone using the installation coordinates of each battery cell and wire in the energy storage architecture, and combining their coordinates to form a three-dimensional spatial grid. This process is equivalent to transforming the channels, which were originally defined only from an electrical perspective, into a geometric structure that can be expressed in space.

[0074] Based on this, a spatial distribution matrix can be constructed. This matrix is ​​a data structure used to describe the relative distribution and connection relationships of each energy pathway within the second cutoff zone in three-dimensional space. Each element represents the spatial location, energy flow direction, and adjacency relationship of a specific channel or node. For example, in a 10×10×3 energy storage layout, if the second cutoff zone is mainly concentrated in the bottom 4×4 area, then the spatial distribution matrix can clearly show this local energy flow concentration area, presenting the three-dimensional shape of the energy recovery hotspot. In layman's terms, the spatial distribution matrix is ​​a three-dimensional mapping table used to reflect which channels have been restored, how they are connected, and how energy flow extends in space.

[0075] Furthermore, in energy scheduling, each battery cell originally has a preset energy allocation matrix, which defines the proportion, priority, and time allocation rules of energy inflow and outflow between different cells. However, when the second cutoff zone is formed, the energy flow structure undergoes local changes—some channels resume conduction, while others remain isolated. To adapt to this new change, the original energy allocation matrix needs to be reconstructed. The specific process is as follows: First, based on the spatial distribution matrix obtained in the previous step, the concentrated energy flow region where the second cutoff zone is located is identified, and the spatial adjacency relationships and energy channel weights of the battery cells within this region are extracted. Then, these new connections are compared with the original energy allocation matrix, and the cells whose pathway states have changed are updated. For example, suppose that in the original allocation matrix, cell A only has energy conduction relationships with cells B and C, but after the formation of the second cutoff zone, energy flow between cells A and D is restored. Then, the new matrix needs to increase the allocation ratio of A and D and rebalance the weights of A, B, and C. The matrix obtained after this update is the reconstructed energy allocation matrix. It reflects the current actual energy flow state better than the original matrix and is the basis for real-time scheduling and timing control.

[0076] Secondly, after the energy flow is redistributed, the load state and energy exchange rate of different battery cells will differ. To ensure stable overall operation, a variable-length time window needs to be set for each cell to control the time range of its energy input or output. The length of the time window is dynamically adjusted according to the reconstructed energy distribution matrix: cells with frequent energy flow and large load changes are allocated shorter time windows for more frequent calibration; cells with stable energy flow and light loads are allocated longer windows to reduce adjustment frequency. For example, in an energy storage architecture consisting of 100 cells, cells numbered 1-20, if located in the energy interaction dense zone, may have a time window of only 2 seconds, while cells numbered 80-100, located in the low interaction zone, may have a time window extended to 10 seconds. Within each time window, two key indicators are collected in real time: voltage ripple and total harmonic distortion (THD). Voltage ripple reflects the degree of fluctuation in the energy balance within the cell, while THD reflects the stability of energy flow and the presence of harmonic interference. By continuously collecting these data, it is possible to monitor in real time whether the energy conduction process is smooth and whether oscillations or backflow occur.

[0077] Finally, after acquiring real-time data, the energy flow timing of each battery cell needs to be dynamically calibrated. The timing path refers to the sequence and duration of energy transfer between different cells. The start-up delay represents the time a cell needs to wait before the energy flow begins, while the duration represents the length of time the cell maintains the energy transfer state. The calibration process is dynamically performed based on the changing trends of voltage ripple and harmonic distortion data: if the voltage ripple of a cell is too large, it indicates that the energy input is changing too rapidly, and its start-up delay should be extended so that it participates in energy transfer later; if the harmonic distortion is too high, it indicates interference or resonance in the energy flow, and its duration should be shortened to avoid accumulation. For example, in an energy transfer path ABC, if a significant increase in ripple is detected at node B, the start-up delay of AB may be extended by 0.5 seconds, while the duration of BC is shortened to balance the flow rhythm. After multiple rounds of real-time calibration, the timing paths of all cells will tend to be coordinated, and the energy flow will gradually transition from a disordered or fluctuating state to a stable and orderly rhythm, ultimately forming an optimized set of energy flow timing paths.

[0078] Example 2, Figure 2 The present invention provides an energy scheduling method system for a data center, comprising a data acquisition module, a loop identification module, a channel disconnection module, a channel restoration module, and a battery adjustment module.

[0079] The data acquisition module is used to acquire the operating data of the battery cells in the solar power plant energy storage system and extract the energy accumulation nodes based on the operating data of the battery cells.

[0080] The loop identification module is used to identify the current loop extension direction of the energy accumulation node based on the preset wire connection relationship, and to obtain the energy echo zone;

[0081] The channel cut-off module is used to cut off the echo channel corresponding to the energy echo zone according to a preset cut-off frequency to obtain the first cut-off zone;

[0082] The channel recovery module is used to extract the residual current attenuation characteristics of the first cut-off zone and perform channel recovery on the first cut-off zone based on the residual current attenuation characteristics to obtain the second cut-off zone.

[0083] The battery adjustment module is used to dynamically adjust the timing path of electrical energy flow inside the battery cell based on the second cut-off zone.

[0084] like Figure 3 The diagram shown is a schematic diagram of the electronic device structure of an energy scheduling method for a data center according to the present invention.

[0085] The electronic device may include a processor, a memory, a communication bus, and a communication interface, and may also include a computer program stored in the memory and executable on the processor, such as an energy scheduling program for a data center.

[0086] The processor is the control unit of the electronic device. It connects to various components of the electronic device through various interfaces and lines. It performs various functions of the electronic device and processes data by running or executing programs or modules stored in the memory and calling data stored in the memory.

[0087] The memory includes at least one type of readable storage medium. In some embodiments, the memory may be an internal storage unit of an electronic device, such as a portable hard drive. The memory can be used to store application software installed on the electronic device and various types of data.

[0088] The communication bus is configured to enable communication between the memory and at least one processor.

[0089] The communication interface is used for communication between the aforementioned electronic device and other devices, including a network interface and a user interface.

[0090] The figure only shows an electronic device with components. Those skilled in the art will understand that the structure shown in the figure does not constitute a limitation on the electronic device and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0091] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0092] The memory in the electronic device stores a data center energy scheduling program, which is a combination of multiple instructions. When run in the processor, it can implement the steps in the aforementioned data center energy scheduling method.

[0093] Specifically, the processor's specific implementation system of the above instructions can be found in the description of the relevant steps in the corresponding embodiments of the accompanying drawings, which will not be repeated here.

[0094] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0095] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0096] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0097] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0098] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An energy scheduling method for a data center, characterized in that, Includes the following steps: Acquire the operating data of battery units in the solar power plant energy storage system, and extract energy accumulation nodes based on the battery unit operating data; Based on the pre-defined wire connection relationships, the current loop extension direction of the energy accumulation node is identified, and the energy echo region is obtained, specifically: Obtain the topology of the conductor connection of the solar power station energy storage system, and start from the energy accumulation node, traverse the conductor connection topology in both the forward and reverse directions to obtain several current flow paths. Calculate the impedance and current phase delay values ​​for each current flow path, and identify the closed-loop current loop based on the impedance and current phase delay values; calculate the active power density vector between the nodes of the closed-loop current loop, and obtain the main loop and branch loop of the closed-loop current loop based on the active power density vector; superimpose the energy accumulation nodes of the main loop to obtain the energy echo core region. The boundary of the energy echo core region is extended by branching loops to obtain the energy echo region. The echo channel corresponding to the energy echo zone is cut off at a preset cutoff frequency to obtain the first cutoff zone; Extract the residual current decay characteristics of the first cut-off region, and perform channel recovery on the first cut-off region based on the residual current decay characteristics to obtain the second cut-off region; The timing path of electrical energy flow inside the battery cell is dynamically adjusted based on the second cut-off zone.

2. The energy scheduling method for a data center according to claim 1, characterized in that, The extraction of energy accumulation nodes based on battery cell operating data specifically involves: Based on the battery cell operation data, multi-channel voltage and current sequences of the battery cells are extracted to obtain the battery operation state matrix; Based on the battery operating state matrix, the charge accumulation rate and discharge fluctuation amplitude of each battery cell are extracted. The energy difference per unit time is calculated based on the charge accumulation rate and discharge fluctuation amplitude, and the initial energy accumulation point is identified based on the energy difference. The physical connection topology between battery cells is obtained, and cluster analysis is performed on the initial energy accumulation points based on the physical connection topology to obtain the energy accumulation region; The energy density of the energy accumulation zone is assessed, and the initial energy accumulation points are screened based on the assessment results to obtain energy accumulation nodes.

3. The energy scheduling method for a data center according to claim 2, characterized in that, The boundary of the energy echo core region is extended based on the branch loop to obtain the energy echo region, specifically as follows: The current oscillation characteristics of the branch circuits are extracted, and cluster analysis is performed on the branch circuits based on the current oscillation characteristics to obtain oscillation clusters; Centered on the energy echo core region, the boundary is extended along the direction of the oscillation cluster to obtain the first extension zone; The avoidance radius is calculated based on the first extension zone, and the boundary of the energy echo core region is adjusted according to the avoidance radius to obtain the energy echo region.

4. The energy scheduling method for a data center according to claim 1, characterized in that, The first cut-off region is obtained by cutting off the echo channel corresponding to the energy echo region according to a preset cut-off frequency, specifically as follows: The echo channels within the energy echo zone are obtained, and the channels are classified according to preset harmonic components based on the current amplitude and frequency characteristics of the echo channels to obtain the classified echo channels. Obtain the harmonic resonance characteristics of the classified echo channels, and preset several cutoff frequency points based on the harmonic resonance characteristics; For each classified echo channel, the channel is cut off sequentially according to the preset cutoff frequency point to obtain the first cutoff zone.

5. The energy scheduling method for a data center according to claim 1, characterized in that, The process of extracting the residual current attenuation characteristics of the first cut-off region and performing channel recovery on the first cut-off region based on these characteristics to obtain the second cut-off region is as follows: The residual current waveform and current response curve of the first cut-off zone are acquired, and the residual current waveform is subjected to exponential fitting to obtain the residual decay time. The residual energy dissipation rate is calculated based on the residual decay time, and recoverable channels are identified based on the dissipation rate distribution map. The rise rate of the voltage excitation is adjusted according to the slope change of the current response curve, and the voltage excitation is applied to the recoverable channel to obtain the second cut-off zone.

6. The energy scheduling method for a data center according to claim 1, characterized in that, The timing path for dynamically adjusting the internal energy flow of the battery cell based on the second cutoff zone is specifically as follows: Obtain the three-dimensional coordinates of the second cut zone to obtain the spatial distribution matrix of the second cut zone; The preset energy distribution matrix of the matrix battery cell is reconstructed based on the spatial distribution matrix of the second cut-off region to obtain the reconstructed energy distribution matrix. Based on the reconstructed energy allocation matrix, a variable-length time window is allocated to each battery cell and voltage ripple and total harmonic distortion data of current are collected in real time. The start-up delay and duration of the timing path are dynamically calibrated based on voltage ripple and harmonic data to obtain an optimized set of power flow timing paths.

7. A system using the data center energy scheduling method as described in any one of claims 1-6, characterized in that, include: The data acquisition module is used to acquire the operating data of the battery cells in the solar power plant energy storage system and extract the energy accumulation nodes based on the operating data of the battery cells. The loop identification module is used to identify the current loop extension direction of the energy accumulation node based on the preset wire connection relationship, and to obtain the energy echo zone; The channel cut-off module is used to cut off the echo channel corresponding to the energy echo zone according to a preset cut-off frequency to obtain the first cut-off zone; The channel recovery module is used to extract the residual current attenuation characteristics of the first cut-off zone and perform channel recovery on the first cut-off zone based on the residual current attenuation characteristics to obtain the second cut-off zone. The battery adjustment module is used to dynamically adjust the timing path of electrical energy flow inside the battery cell based on the second cut-off zone.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the data center energy scheduling method as described in any one of claims 1 to 6.

Images