A Photovoltaic-Storage-Charging Intelligent Collaborative Control System Based on the Energy Internet

By using a multi-source sensing layer and a spectrum feature extraction module to make collaborative decisions on the energy internet cloud platform, the system solves the problems of distinguishing the spectrum features of photovoltaics and charging piles and differentiating energy storage types. This enables second-level response to disturbances in photovoltaics and charging piles, improving system stability and the lifespan of energy storage equipment.

CN123121522APending Publication Date: 2026-11-17HEBEI SHUOHUA OPTOELECTRONICS TECHNOLOGY CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202611534076.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-30
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing photovoltaic-storage-charging integrated systems face challenges such as random photovoltaic output and DC bus power disturbances caused by the pulse nature of electric vehicle charging loads. Existing control logic response is lagging, energy storage system lifespan is shortened, and the system fails to effectively distinguish between spectrum characteristics and energy storage type. SOC management lacks dynamic coupling and cannot effectively cope with second-level power disturbances.

Method used

The system employs a multi-source sensing layer to collect data in real time. Through a spectrum feature extraction module and a disturbance coupling analysis module, collaborative decision-making is carried out on the energy internet cloud platform to distinguish the spectrum characteristics of photovoltaics and charging piles. The system coordinates energy storage units in different frequency bands through an energy storage converter and combines a SOC partitioning adaptive strategy to achieve second-level spectrum feature-driven collaborative control.

Benefits of technology

It achieves a second-level response to disturbances from photovoltaics and charging piles, avoids control mismatch and shortened lifespan of the energy storage system, reduces DC bus voltage fluctuations, and improves system stability and the service life of energy storage equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121522A_ABST
    Figure CN123121522A_ABST
Patent Text Reader

Abstract

This invention discloses a photovoltaic-storage-charging intelligent collaborative control system based on the energy internet. The system includes a multi-source sensing layer, a cloud platform collaborative decision-making layer, a hierarchical collaborative control layer, an edge controller, and an energy router. A spectrum feature extraction module performs wavelet packet decomposition on the photovoltaic power change rate to obtain a first disturbance vector, and performs density clustering on the charging pile power step to obtain a second disturbance vector. A disturbance coupling analysis module performs coupling analysis under the constraint of the DC bus power balance equation and determines whether the disturbance is low-frequency slow-change, mid-frequency oscillation, or high-frequency pulse. In the case of high-frequency pulse disturbances, the energy storage converter allocates high-frequency components to power-type energy storage units and low-frequency components to energy-type energy storage batteries, limiting their switching frequency. In the case of low-frequency slow-change disturbances, the energy storage system smooths low-frequency fluctuations, and the photovoltaic inverter suppresses its own high-frequency fluctuations. This invention achieves wideband disturbance spectrum decoupling and differentiated collaborative smoothing, balancing DC bus voltage quality and energy storage battery life.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of energy internet and integrated photovoltaic-storage-charging collaborative control technology, and in particular to an intelligent collaborative control system for photovoltaic-storage-charging based on energy internet. Background Technology

[0002] The photovoltaic-storage-charging integrated system is a system that integrates photovoltaic power generation, energy storage system and electric vehicle charging facilities. Existing photovoltaic-storage-charging integrated systems face a challenge in operation: photovoltaic output has strong randomness and volatility, and electric vehicle charging load has strong pulse and uncertainty. When the two are superimposed on the DC bus, they form a wideband power disturbance, which poses a challenge to the system's power balance control, the lifespan of energy storage equipment and the DC bus voltage quality.

[0003] In the prior art, patent application CN114243802A discloses a method and system for coordinated control of photovoltaic, energy storage, and charging resources in a distribution transformer area. This method acquires data from photovoltaic, energy storage, charging piles, and distribution transformers through an energy internet gateway, identifies data anomalies, and calculates the active power demand of the distribution transformer area based on its load rate, thereby coordinating and controlling photovoltaic, energy storage, and charging resources. While this solution achieves coordinated operation of these resources to a certain extent, its core control logic is a passive, load-rate-driven coordination mechanism.

[0004] First, the time granularity of the load factor criterion is coarse, making it unable to cope with power disturbances on a second-level scale. Distribution transformer load factors are usually calculated based on average values ​​over a time scale of minutes or even longer, while the cloud shading effect on photovoltaic output can cause a sudden drop in power of 30% to 50% within seconds. The connection and disconnection of electric vehicle fast charging piles also have second-level random impact characteristics. The load factor-driven control logic lags in response to second-level disturbances, resulting in frequent fluctuations in DC bus voltage, and in severe cases, even triggering protection actions.

[0005] Second, the spectral coupling between photovoltaic power output fluctuations and charging load pulses has not been decoupled. Photovoltaic power output fluctuations are mainly characterized by low-frequency slow variations (such as fluctuations of 0.01Hz to 1Hz caused by cloud cover) and medium-frequency oscillations (such as fluctuations of several Hz caused by inverter MPPT jitter), while charging pile load fluctuations are characterized by high-frequency random pulses. When these two types of disturbances are superimposed on the DC bus, they will create a wideband power impact on the energy storage converter. Existing solutions do not distinguish, extract, and couple the spectral characteristics of these two types of disturbances, resulting in a lack of targeted control strategies for the energy storage converter, frequent charge-discharge switching of the energy storage battery, and accelerated cycle life degradation.

[0006] Third, energy storage systems fail to differentiate between the functional differences between energy-type storage batteries and power-type energy storage units. High-frequency power throughput requires energy storage devices to have rapid response capabilities and high cycle life tolerance, while energy-type storage batteries, although having high energy density, experience significantly accelerated aging due to frequent charge-discharge switching. In existing solutions, energy storage systems are typically composed of a single type of battery, making it difficult to simultaneously meet the two mutually restrictive objectives of high-frequency power throughput and low switching frequency. If the battery bears the high-frequency power component, its lifespan decreases sharply; if the battery switching frequency is limited, high-frequency disturbances cannot be effectively mitigated.

[0007] Fourth, there is a lack of dynamic coupling between the energy storage SOC management strategy and the control mode. In existing solutions, the energy storage charging and discharging strategy is usually designed independently of the collaborative control strategy. The differences in the system's response to photovoltaic fluctuations and charging shocks under different SOC ranges are not considered, which increases the risk of deep charging and discharging of energy storage and results in insufficient system regulation capability under extreme SOC conditions.

[0008] Fifth, the existing scheme lacks the ability to converge and analyze regional disturbance characteristics of multiple sites and make early predictions. The disturbance information of a single photovoltaic storage and charging site is not fully utilized, and it is impossible to identify the spatiotemporal distribution pattern of high-frequency pulse-type disturbances at the regional level, making it difficult to achieve early switching and preparatory adjustment of control modes.

[0009] Therefore, in response to the problems mentioned above, this invention proposes a photovoltaic-storage-charging intelligent collaborative control system based on the energy internet. Summary of the Invention

[0010] To overcome the problem that photovoltaic-storage-charging collaborative control technology is unable to effectively cope with the superposition of broadband power disturbances formed on the DC bus due to time scale mismatch, undecoupled spectrum coupling, undifferentiated energy storage types, and uncoupled SOC management and control modes, this invention proposes a photovoltaic-storage-charging intelligent collaborative control system based on the energy internet.

[0011] The technical solution of the present invention is: a photovoltaic-storage-charging intelligent collaborative control system based on the energy internet, comprising a multi-source sensing layer, a cloud platform collaborative decision-making layer, a hierarchical collaborative control layer, an edge controller, and an energy router.

[0012] The multi-source sensing layer is used to collect real-time data on the output power of the photovoltaic array and its power change rate, the state of charge (SOC) and charging and discharging power of the energy storage system, the load power of the charging pile and its power step amplitude, the DC bus voltage and the power at the grid connection point; the collected data is uploaded to the cloud platform collaborative decision-making layer via the energy internet cloud platform.

[0013] The energy storage system includes energy-type energy storage batteries and power-type energy storage units. Energy-type energy storage batteries have high energy density characteristics and are suitable for handling low-frequency average power and long-term energy balance. Power-type energy storage units have high power density and fast response characteristics and are suitable for handling high-frequency power components and short-time pulse power. Both are connected to the DC bus through their respective converters or connected in parallel through a shared DC bus and are uniformly coordinated and controlled by the energy storage converter.

[0014] For the acquisition of charging pile load power and its power step amplitude, the multi-source sensing layer adopts a dual sampling frequency mechanism. The instantaneous value of the charging pile load power is acquired at the first sampling frequency, and the moving average value of the charging pile load power is acquired at the second sampling frequency. The first sampling frequency is greater than the second sampling frequency. The moment when the difference between the instantaneous value and the moving average value exceeds the power step judgment threshold is marked as a power step event. The amplitude and occurrence time of the event are recorded. The amplitude and occurrence time of multiple consecutive power step events constitute the power step amplitude data sequence of the charging pile load power.

[0015] The cloud platform collaborative decision-making layer runs a collaborative decision-making engine on the energy internet cloud platform. This engine includes a spectrum feature extraction module, a disturbance coupling analysis module, and a hierarchical collaborative control layer. The spectrum feature extraction module is used to perform time-frequency domain analysis on the power change rate of the photovoltaic array output power, extract the first disturbance vector characterizing the spectrum characteristics of photovoltaic power output fluctuation, and simultaneously perform pulse feature clustering on the power step amplitude of the charging pile load power to extract the second disturbance vector characterizing the pulse characteristics of the charging load.

[0016] Specifically, the time-frequency domain analysis of the power change rate of the photovoltaic array output power includes: inputting the power change rate of the photovoltaic array output power into a wavelet packet decomposer, performing three-level wavelet packet decomposition to obtain the decomposition coefficients of eight frequency bands, calculating the energy proportion of the decomposition coefficients of each frequency band, and constructing the frequency band number and energy proportion of the frequency bands whose energy proportion exceeds the first energy threshold as the first perturbation vector. This first perturbation vector can characterize the frequency band distribution characteristics of photovoltaic power output fluctuations, such as the low-frequency slowly varying components caused by cloud cover and the mid-frequency oscillation components caused by inverter MPPT jitter.

[0017] The pulse feature clustering of the power step amplitude of the charging pile load power includes: using the amplitude and occurrence time of the power step event as feature quantities, the density clustering algorithm is used to perform cluster analysis on the power step amplitude data sequence, and the amplitude and occurrence interval corresponding to the cluster center in the clustering result are used as pulse feature parameters in the second perturbation vector. The pulse feature parameters include the average pulse amplitude and the average pulse interval, thereby extracting the typical pulse amplitude and typical time interval of the charging load.

[0018] The disturbance coupling analysis module is configured to perform superposition coupling analysis on the first disturbance vector and the second disturbance vector under the constraint of the DC bus power balance equation, calculate the composite power disturbance spectrum on the DC bus, and determine the current disturbance type as the dominant frequency band type based on the energy ratio of each frequency band in the composite power disturbance spectrum.

[0019] The establishment of the DC bus power balance equation includes: ; in, Photovoltaic output power; This is the sum of the charging pile load power and the local AC load power. The charging and discharging power of the energy storage system; Power exchange at the grid connection point; This is the DC bus capacitance value; This refers to the DC bus voltage. The positive direction of power injection is defined as the direction of the DC bus voltage. , , A positive value indicates that power is injected into the bus, while a negative value indicates that power is absorbed from the bus. A positive value indicates that power is absorbed from the bus.

[0020] Substitute the first and second disturbance vectors into the DC bus power balance equations mentioned above, respectively, and solve the composite power disturbance spectrum through frequency domain transformation. Calculate the energy proportion of the composite power disturbance spectrum in the low-frequency, mid-frequency, and high-frequency bands. The low-frequency band corresponds to slowly varying fluctuations such as photovoltaic cloud shading, the mid-frequency band corresponds to oscillating fluctuations such as MPPT jitter, and the high-frequency band corresponds to charging pile pulse impacts.

[0021] The principle for determining whether the current disturbance type is the dominant frequency band type is as follows: When the proportion of low-frequency energy exceeds the first proportion threshold of total energy, it is determined to be a low-frequency slow-change type; When the proportion of mid-frequency energy exceeds the second threshold of the total energy, it is determined to be a mid-frequency oscillation type; When the proportion of high-frequency energy exceeds the third threshold of total energy, it is determined to be a high-frequency pulse type; When the energy proportion of multiple frequency bands does not exceed the corresponding proportion threshold, it is determined to be a mixed disturbance type.

[0022] The beneficial effects of this invention are: 1. This invention uses a spectrum feature extraction module to perform wavelet packet decomposition on the photovoltaic power change rate and density clustering on the charging pile power step. Under the constraint of the DC bus power balance equation, it superimposes and couples the two types of disturbances for analysis and determines the dominant frequency band type. This changes the basis for generating coordinated control commands from minute-level average load rate to second-level spectrum features, eliminating the response lag caused by time scale mismatch.

[0023] 2. This invention decouples two types of disturbances in the frequency domain and determines the dominant disturbance type through a disturbance coupling analysis module, so that disturbances in different frequency bands can be differentiated and assigned to the corresponding control modes for processing, avoiding control mismatch caused by the superposition of broadband disturbances acting uniformly on the energy storage converter.

[0024] 3. This invention uses an energy storage converter to distribute high-frequency power components to power-type energy storage units and low-frequency power components to energy-type energy storage batteries under high-frequency pulse-type disturbances. This allows the high-frequency throughput to be handled by the power-type units and the charging and discharging switching frequency of the energy-type batteries to be limited to a preset threshold, thus solving the problem of battery life degradation without sacrificing smoothing capabilities.

[0025] 4. This invention uses a SOC partitioning adaptive strategy to determine whether to enable high-frequency power throughput mode based on the SOC range under high-frequency pulse-type disturbances. In the deep discharge or deep charge zone, high-frequency throughput is prohibited and the power limit of the charging pile is lowered instead, so that the energy storage protection and control mode selection are linked, avoiding the risk of deep charging and discharging under extreme SOC.

[0026] 5. This invention aggregates the disturbance vectors and disturbance type determination results of multiple sites through the energy internet cloud platform to generate a regional disturbance feature map. When the frequency of a high-frequency pulse disturbance exceeds the threshold, an early switching command is issued, so that the energy storage converters of each site can enter the preparatory state of high-frequency power throughput mode in advance before the disturbance type is switched. Attached Figure Description

[0027] Figure 1 The diagram shown is a system structure block diagram of the present invention; Figure 2 The diagram shown is a schematic of the density clustering algorithm of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 some embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Please see Figure 1 The present invention provides an embodiment: This system comprises a multi-source sensing layer, a cloud platform collaborative decision-making layer, a hierarchical collaborative control layer, an edge controller, and an energy router. Deployed at an integrated photovoltaic-storage-charging site, the system includes photovoltaic arrays, an energy storage system, electric vehicle charging piles, and local AC loads. Each unit is connected to the external power distribution network via a DC bus and a grid connection point. The energy storage system consists of energy-type storage batteries and power-type energy storage units, each connected to the DC bus via its own converter and uniformly coordinated and controlled by the energy storage converter. The energy-type storage batteries utilize high-energy-density cells, primarily used for handling low-frequency average power and long-term energy balance; the power-type energy storage units employ high-power-density, fast-response energy storage elements, primarily used for handling high-frequency power components and short-term pulse power.

[0030] The multi-source sensing layer deploys various types of sensors and acquisition terminals on the site side to collect real-time data on the output power of the photovoltaic array and its power change rate, the state of charge (SOC) and charging / discharging power of the energy storage system, the load power of the charging pile and its power step amplitude, and the DC bus voltage and grid connection point power.

[0031] This embodiment employs a dual-sampling-frequency mechanism for collecting the charging pile load power and its power step amplitude. The instantaneous value of the charging pile load power is collected at a first sampling frequency (preferably 1-10kHz), set to a higher frequency to ensure the capture of rapid changes such as charging pile switching and power step fluctuations. The moving average value of the charging pile load power is collected at a second sampling frequency (preferably 10-100kHz), set to a lower frequency. The moving average window (preferably 10 power frequency cycles) covers several power frequency cycles to reflect the steady-state level of the charging load. The first sampling frequency is greater than the second sampling frequency. A power step event is marked when the difference between the instantaneous value and the moving average value exceeds the power step judgment threshold (preferably 5% to 15% of the charging pile's rated power), and the amplitude and occurrence time of this power step event are recorded. The amplitudes and occurrence times of multiple consecutive power step events constitute a power step amplitude data sequence of the charging pile load power. This data sequence is then uploaded to the energy internet cloud platform for use by the cloud platform's collaborative decision-making layer.

[0032] All data collected by the multi-source sensing layer is uploaded through energy routers on the energy internet cloud platform. Each energy router assigns a unique energy internet protocol address to the photovoltaic inverter, energy storage converter, and charging pile, achieving device-level addressing and data isolation. Simultaneously, based on the disturbance type determined by the cloud platform's collaborative decision-making layer, the energy router assigns different transmission priorities to different instruction categories in the collaborative control instruction set. When the disturbance type is a high-frequency pulse, the transmission priority of the energy storage converter's charging and discharging power reference value and its response time constant is higher than the adjustment amount of the photovoltaic inverter's active power reference value. This ensures that critical control instructions can reach the executing device with minimal delay under second-level high-frequency disturbance conditions.

[0033] The cloud platform collaborative decision-making layer runs a collaborative decision-making engine on the energy internet cloud platform. This engine includes a spectrum feature extraction module, a disturbance coupling analysis module, and a hierarchical collaborative control layer.

[0034] The spectrum feature extraction module performs time-frequency domain analysis on the power change rate of the photovoltaic array output power and extracts the first disturbance vector characterizing the spectrum features of photovoltaic power output fluctuation.

[0035] Specifically, the power change rate of the photovoltaic array output power is input into a wavelet packet decomposer for three-level wavelet packet decomposition, yielding decomposition coefficients for eight frequency bands. The energy percentage of each frequency band's decomposition coefficients is calculated, and the frequency band number and energy percentage corresponding to the frequency band with an energy percentage exceeding a first energy threshold (preferably 5-15%) constitute the first perturbation vector. Through three-level wavelet packet decomposition, photovoltaic power output fluctuations can be divided into eight frequency bands from low to high frequency, effectively distinguishing low-frequency slowly varying components caused by cloud cover, mid-frequency oscillation components caused by inverter MPPT jitter, and other high-frequency components. The first perturbation vector not only contains the frequency band number with a large energy percentage but also the specific energy percentage of each frequency band, providing a quantitative basis for subsequent coupling analysis.

[0036] The spectrum feature extraction module simultaneously performs pulse feature clustering on the power step amplitude of the charging pile load power to extract the second disturbance vector characterizing the pulse features of the charging load.

[0037] Specifically, using the amplitude and occurrence time of power step events as features, a density clustering algorithm is employed to perform cluster analysis on the power step amplitude data sequence. The density clustering algorithm automatically identifies high-density regions in the data sequence, grouping power step events with similar characteristics into the same cluster. The amplitude and occurrence interval corresponding to the cluster centers in the clustering results are used as pulse feature parameters in the second perturbation vector. The pulse feature parameters in the second perturbation vector include the mean pulse amplitude and the mean pulse interval, representing the typical intensity and typical time interval of the charging load pulse, respectively.

[0038] Please see Figure 2 Specifically, let's explain the density clustering algorithm: S1, extract the amplitude of each power step event from the power step amplitude data sequence. and the time of occurrence To eliminate the dimensional difference between amplitude and time, the amplitude is expressed in terms of the charging pile's rated power. Normalization is performed on the baseline to obtain The relative occurrence time is calculated based on the occurrence time of the first power step event, and then a moving average window is used. Normalization is performed on the baseline to obtain Among them, the moving average window Ten power frequency cycles are preferred. Through the above steps, each power step event is represented as a two-dimensional feature sample. .

[0039] S2, neighborhood radius In the normalized feature space, a value of 0.05 is used, corresponding to 5% of the rated power of the charging pile; the minimum number of neighborhood points, MinPts, is set to 5. The above values ​​fall within the range of 3-10% of the rated power of the charging pile for the preferred neighborhood radius and 3 to 8 for the preferred minimum number of neighborhood points.

[0040] S3, for any two samples and The distance between the two in the normalized feature space is calculated using Euclidean distance: ; like ≤ Then the sample is called Belongs to the sample of Neighborhood.

[0041] S4, for each sample Statistics The number of samples within the neighborhood, this number includes samples It itself. If the sample of If the number of samples in the neighborhood is greater than or equal to MinPts, then the sample... Mark it as a core point; otherwise, mark it as a non-core point.

[0042] S5 starts from any unvisited core point and... All samples within the neighborhood are added to the current cluster. For newly added core points within the current cluster, continue to add them... Samples within the neighborhood are added to the current cluster until no new core points can be expanded. This process is repeated until all core points have been visited. Through these steps, all samples connected by density reachability are grouped into the same cluster; samples not belonging to any cluster are marked as noise points and do not participate in subsequent impulse feature parameter calculations.

[0043] S6. For each cluster, calculate the arithmetic mean of all samples in the amplitude dimension, which is used as the amplitude component of the cluster center. This amplitude component is the pulse amplitude mean of the cluster. All power step events within the cluster are sorted by occurrence time. Sort and calculate the time interval between adjacent events. The average value of the pulse interval is taken as the time interval component corresponding to the cluster center of the cluster. This time interval component is the average pulse interval of the cluster.

[0044] If only one cluster is obtained, the average pulse amplitude and average pulse interval of that cluster are directly used as the pulse characteristic parameters in the second perturbation vector. If multiple clusters are obtained, the clusters are sorted from largest to smallest in terms of the number of samples within each cluster, and the average pulse amplitude and average pulse interval of the top-ranked cluster are used as the pulse characteristic parameters in the second perturbation vector. If the top-ranked clusters have the same number of samples, the arithmetic mean of the average pulse amplitude and average pulse interval of all clusters is used as the pulse characteristic parameters in the second perturbation vector. The resulting average pulse amplitude and average pulse interval represent the typical intensity and typical time interval of the charging load pulse, respectively, and are used as the second perturbation vector in the superposition coupling analysis of the subsequent perturbation coupling analysis module.

[0045] In the above manner, the random pulse characteristics of the charging pile load are quantified into pulse characteristic parameters that can participate in the coupling analysis.

[0046] The disturbance coupling analysis module performs superposition coupling analysis on the first disturbance vector and the second disturbance vector under the constraint of the DC bus power balance equation, calculates the composite power disturbance spectrum on the DC bus, and determines the current disturbance type as the dominant frequency band type based on the energy ratio of each frequency band in the composite power disturbance spectrum.

[0047] The DC bus power balance equation is established as follows: ; in, Photovoltaic output power; This is the sum of the charging pile load power and the local AC load power. The charging and discharging power of the energy storage system; Power exchange at the grid connection point; This is the DC bus capacitance value; This refers to the DC bus voltage. The positive direction of power injection is defined as the direction of the DC bus voltage. , , A positive value indicates that power is injected into the bus, while a negative value indicates that power is absorbed from the bus. A positive value indicates that power is absorbed from the bus.

[0048] Substitute the first and second disturbance vectors into the DC bus power balance equations described above, and solve for the composite power disturbance spectrum through frequency domain transformation.

[0049] Specifically, the energy proportions of each frequency band contained in the first disturbance vector reflect the frequency domain distribution of photovoltaic power output fluctuations, while the average pulse amplitude and average pulse interval contained in the second disturbance vector reflect the time domain characteristics of charging load pulses. Substituting these two values ​​into the power balance equation, a frequency domain transformation can convert the time domain power disturbance into a frequency domain power disturbance spectrum, thus obtaining the composite power disturbance spectrum. Then, the energy proportions of the composite power disturbance spectrum in the low-frequency, mid-frequency, and high-frequency bands are calculated. The low-frequency band (preferably 0.01-1Hz) corresponds to slowly varying fluctuations such as photovoltaic cloud shading, the mid-frequency band (preferably 1-10Hz) corresponds to oscillating fluctuations such as MPPT jitter, and the high-frequency band (preferably above 10Hz) corresponds to charging pile pulse impacts.

[0050] The principles for determining whether the current disturbance type is the dominant frequency band type include the following four points: (1) When the proportion of low-frequency energy exceeds the first proportion threshold of total energy (preferably 40-60%), it is determined to be a low-frequency slow-change type; (2) When the proportion of mid-frequency energy exceeds the second proportion threshold of the total energy (preferably 30-50%), it is determined to be a mid-frequency oscillation type; (3) When the proportion of high-frequency energy exceeds the third proportion threshold of the total energy (preferably 30-50%), it is determined to be a high-frequency pulse type; (4) When the energy proportion of multiple frequency bands does not exceed the corresponding proportion threshold, it is determined to be a mixed disturbance type.

[0051] The first, second, and third proportional thresholds can be set based on the actual operating experience of the site and the equipment's tolerance; they can be the same or different.

[0052] The hierarchical collaborative control layer generates a corresponding collaborative control instruction set based on the disturbance type determined by the disturbance coupling analysis module, and then distributes the collaborative control instruction set to the edge controller for execution via the energy internet cloud platform. The collaborative control instruction set includes the adjustment amount of the active power reference value of the photovoltaic inverter, the charging and discharging power reference value of the energy storage converter and its response time constant, and the power limit of the charging pile.

[0053] When the disturbance type is determined to be high-frequency pulse type, the cooperative control command set activates the high-frequency power throughput mode of the energy storage converter. In this mode, the energy storage converter allocates the high-frequency power command component to the power-type energy storage unit, which rapidly processes high-frequency power to smooth out the pulse impact of the charging pile; the low-frequency power command component is allocated to the energy-type energy storage battery, which only bears the low-frequency average power, and its charge-discharge switching frequency is limited to a preset threshold (preferably 8-15 times / hour). Through the above power frequency division and allocation, both rapid smoothing of high-frequency pulse disturbances and accelerated aging of the energy-type energy storage battery due to frequent charge-discharge switching are achieved, thus resolving the inherent contradiction between high-frequency power throughput and low battery switching frequency.

[0054] When the disturbance type is determined to be low-frequency slow-varying, the cooperative control command set is configured to enable the low-frequency power smoothing mode of the energy storage converter. In this mode, the low-frequency fluctuations in photovoltaic output are smoothed by using the charging and discharging power reference value of the energy storage converter. Since the low-frequency slow-varying fluctuations have large energy and change slowly, it is more suitable for energy storage batteries to undertake the main smoothing task. At the same time, the photovoltaic inverter performs its own high-frequency fluctuation suppression, that is, it suppresses high-frequency components such as MPPT jitter through the local control loop of the photovoltaic inverter to prevent them from being conducted to the DC bus.

[0055] When the disturbance type is determined to be mid-frequency oscillation, the coordinated control instruction set is used to adjust the response time constant of the energy storage converter and the droop coefficient of the photovoltaic inverter, so that the energy storage converter and the photovoltaic inverter form a damping coordination in the mid-frequency band to suppress the mid-frequency oscillation of the DC bus.

[0056] When the disturbance type is determined to be a mixed disturbance, the coordinated control instruction set is used to allocate the high-frequency component to the power-type energy storage unit and the low-frequency component to the energy-type energy storage battery according to the energy ratio of each frequency band, and coordinate the power limits of the photovoltaic inverter and the charging pile to achieve comprehensive suppression of multi-frequency band disturbances.

[0057] In this embodiment, the generation of the collaborative control instruction set in the hierarchical collaborative control layer also includes an energy storage SOC partitioning adaptive strategy. This strategy divides the SOC of the energy storage battery into three zones: a deep discharge zone (preferably SOC below 20-30%), a normal operation zone (preferably SOC between 25-75%), and a deep charge zone (preferably SOC above 70-80%).

[0058] When the disturbance type is determined to be high-frequency pulse type and the SOC is in the normal operating range, the high-frequency power throughput mode of the energy storage converter is activated, and the high-frequency power component is undertaken by the power type energy storage unit, while the energy type energy storage battery does not participate in the high-frequency charge and discharge switching.

[0059] When the disturbance type is determined to be high-frequency pulse type and the SOC is in the deep discharge or deep charge zone, the high-frequency power throughput mode of the energy storage converter is prohibited. Instead, the charging pile power limit reduction strategy is activated to reduce the upper limit of the charging pile load power to a level that matches the energy storage power throughput capacity allowed by the current SOC. This prioritizes system safety under extreme SOC conditions and avoids deep charging and discharging of energy storage batteries.

[0060] The edge controller is used to execute the first-level converter local control layer and the second-level edge coordination layer.

[0061] The first layer is the converter local control layer, used to deploy droop control loops in energy storage converters and photovoltaic inverters. The droop control loop takes the DC bus voltage deviation as input and generates the first increment for converter power regulation. This layer has a fast response speed and a short control cycle (preferably 100μs to 1ms), enabling it to respond quickly to DC bus voltage fluctuations on a millisecond to second timescale.

[0062] The second layer is the edge coordination layer, which receives the collaborative control command set issued by the energy internet cloud platform and converts it into a droop coefficient correction amount that can be executed by the local control layer of the converter. The droop coefficient correction amount is added to the first increment to generate the final value of the converter power reference value. This layer has a slower response speed than the first layer and a longer control cycle (preferably 10ms to 100ms), and is responsible for realizing global optimization and disturbance type adaptation.

[0063] The response speed of the first layer is greater than that of the second layer, and the control cycle of the first layer is less than that of the second layer.

[0064] The energy internet cloud platform aggregates and analyzes the first and second disturbance vectors and disturbance type determination results of multiple photovoltaic, energy storage, and charging sites to generate a regional disturbance feature map. The regional disturbance feature map uses time as the horizontal axis and the site as the vertical axis to mark the disturbance type and frequency band energy proportion of each site at different times, thus intuitively reflecting the spatiotemporal distribution pattern of high-frequency pulse-type disturbances within the region.

[0065] When the frequency of high-frequency pulse-type disturbances in the regional disturbance feature map exceeds a preset frequency threshold (preferably 8-15 times / hour), the energy internet cloud platform issues an advance switching command to the collaborative decision engine of multiple photovoltaic-storage-charging sites. This advance switching command instructs the collaborative decision engine to adjust the control parameters of the energy storage converter to a preparatory state for high-frequency power throughput mode when the current disturbance type is low-frequency slowly varying or medium-frequency oscillating. Through this method, the system can identify the spatiotemporal distribution pattern of high-frequency pulse-type disturbances at the regional level, realize advance switching and preparatory adjustment of control modes, and further improve the system's response speed and suppression capability to high-frequency pulse disturbances.

[0066] To further verify the effectiveness of this invention, the following experiment was conducted. This experiment was carried out at the same integrated photovoltaic-storage-charging site, comparing the present invention with an existing load-rate-driven coordinated control scheme (which judges and calculates the active power demand of the distribution transformer load based on the transformer load rate, and coordinates the output of photovoltaic, energy storage, and charging piles accordingly; its control criterion is based on minute-level average load rate, without spectral decoupling of photovoltaic fluctuations and charging pulses, and without distinguishing between energy-type and power-type energy storage units). The following operating conditions were set for this experiment: The photovoltaic array encountered multiple cloud cover events during operation, and each cloud cover event caused the photovoltaic output to drop by a certain percentage within a few seconds. Electric vehicle charging stations experience multiple switching events during operation, with each switching causing a power jump.

[0067] The experiment was repeated multiple times, with each experiment lasting the same duration. Indicators such as DC bus voltage fluctuation, number of battery charge / discharge switching cycles, and DC bus voltage recovery time were recorded. The experimental results are shown in the table below: Table 1 Experimental results of existing load rate driven solutions ; Table 2 Experimental results of the present invention ; As shown in Table 1, compared with the existing load rate drive scheme, the present invention significantly reduces the maximum deviation of DC bus voltage, greatly reduces the number of charging and discharging switching of energy storage battery, and significantly shortens the DC bus voltage recovery time.

[0068] To further verify the effectiveness of this invention under high-frequency pulse-type disturbance conditions, the following experiment was conducted. The operating conditions for this experiment were set as follows: The photovoltaic output remained relatively stable, while the charging pile experienced multiple power jumps in a short period of time, forming a high-frequency pulse-type disturbance.

[0069] The present invention was compared with existing load-rate driven schemes, and the power allocation between the power-type energy storage unit and the energy-type energy storage battery, as well as the charge-discharge switching frequency of the energy-type energy storage battery, were recorded. The experimental results are shown in the table below: Table 3. Performance of Existing Load Rate Driven Solutions under High-Frequency Pulse Disturbance Conditions ; Table 4. Effects of the present invention under high-frequency pulse-type disturbance conditions. ; As shown in the table above, under high-frequency pulse-type disturbance conditions, the present invention allocates most of the high-frequency power components to the power-type energy storage unit through the high-frequency power throughput mode of the energy storage converter, while the energy-type energy storage battery only undertakes a small amount of low-frequency power, and its charging and discharging switching frequency is limited to a low level.

[0070] To further verify the effectiveness of the SOC partitioning adaptive strategy of this invention, the following experiment was conducted. The operating conditions of this experiment were set as follows: The energy storage battery was positioned within its normal operating range, deep discharge range, and deep charge range. A high-frequency pulsed disturbance was applied to the charging pile, and the system's response was recorded at different SOC intervals. The experimental results are shown in the table below: Table 5 Validation results of the SOC partitioning adaptive strategy ; As shown in the table above, when the SOC is in the normal operating range, the system uses a high-frequency power throughput mode, with the power-type energy storage unit handling the high-frequency power component, and the energy-type energy storage battery has a low charging and discharging switching frequency. When the SOC is in the deep discharge or deep charge range, the system prohibits the use of the high-frequency power throughput mode and instead uses the charging pile power limit reduction strategy, with the energy-type energy storage battery charging and discharging switching frequency being zero, effectively avoiding deep charging and discharging of the energy-type energy storage battery.

[0071] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A photovoltaic-storage-charging intelligent collaborative control system based on the energy internet, characterized in that, include: The multi-source sensing layer is used to collect real-time data on the output power of the photovoltaic array and its power change rate, the state of charge (SOC) and charging / discharging power of the energy storage system, the load power of the charging pile and its power step amplitude, the DC bus voltage and the power at the grid connection point, and upload the collected data to the energy internet cloud platform. The energy storage system includes energy-type energy storage batteries and power-type energy storage units. The cloud platform collaborative decision-making layer is used to run the collaborative decision-making engine on the energy internet cloud platform. This collaborative decision-making engine includes: The spectrum feature extraction module is used to perform time-frequency domain analysis on the power change rate of the photovoltaic array output power, extract the first disturbance vector characterizing the spectrum characteristics of photovoltaic power output fluctuation, and simultaneously perform pulse feature clustering on the power step amplitude of the charging pile load power to extract the second disturbance vector characterizing the pulse characteristics of the charging load. The disturbance coupling analysis module is used to perform superposition coupling analysis on the first disturbance vector and the second disturbance vector under the constraint of the DC bus power balance equation, calculate the composite power disturbance spectrum on the DC bus, and determine the current disturbance type as the dominant frequency band type based on the energy ratio of each frequency band in the composite power disturbance spectrum. The disturbance type includes low-frequency slow-change type, mid-frequency oscillation type and high-frequency pulse type. The hierarchical collaborative control layer is used to generate corresponding collaborative control instruction sets according to the type of disturbance, and to send the collaborative control instruction sets to the edge controllers for execution via the energy internet cloud platform. The collaborative control instruction sets include the adjustment of the active power reference value of the photovoltaic inverter, the charging and discharging power reference value of the energy storage converter and its response time constant, and the power limit of the charging pile. Specifically, when the disturbance type is determined to be high-frequency pulse type, the cooperative control instruction set is used to enable the high-frequency power throughput mode of the energy storage converter. At this time, the energy storage converter allocates the high-frequency power command component to the power type energy storage unit and the low-frequency power command component to the energy type energy storage battery. The charging and discharging switching frequency of the energy type energy storage battery is limited to within a preset threshold. When the disturbance type is determined to be low-frequency slow change type, the cooperative control instruction set is used to enable the low-frequency power smoothing mode of the energy storage converter. At this time, the low-frequency fluctuation of photovoltaic output is smoothed by the charging and discharging power reference value of the energy storage converter, and the photovoltaic inverter performs its own high-frequency fluctuation suppression.

2. The photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 1, characterized in that, The time-frequency domain analysis of the power change rate of the photovoltaic array output power in the spectrum feature extraction module includes: inputting the power change rate of the photovoltaic array output power into a wavelet packet decomposer, performing three-level wavelet packet decomposition to obtain the decomposition coefficients of eight frequency bands, calculating the energy proportion of the decomposition coefficients of each frequency band, and constructing the first perturbation vector with the frequency band number corresponding to the frequency band whose energy proportion exceeds the first energy threshold and the energy proportion.

3. The photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 1, characterized in that, The perturbation coupling analysis module calculates the combined power perturbation spectrum on the DC bus, including: Establish the DC bus power balance equation: in For photovoltaic output power, It is the sum of the charging pile load power and the local AC load power. For the charging and discharging power of the energy storage system, For power exchange at the grid connection point, This is the DC bus capacitance value. This refers to the DC bus voltage; the positive direction of power injection is defined as the DC bus voltage. , , A positive value indicates that power is injected into the bus, while a negative value indicates that power is absorbed from the bus. A positive value indicates that power is absorbed from the busbar; Substitute the first and second disturbance vectors into the DC bus power balance equation, respectively, solve the composite power disturbance spectrum through frequency domain transformation, and calculate the energy proportion of the composite power disturbance spectrum in the low-frequency, mid-frequency and high-frequency bands.

4. The photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 3, characterized in that, The determination that the current disturbance type is the dominant frequency band type includes: When the proportion of low-frequency energy exceeds the first proportion threshold of total energy, it is determined to be a low-frequency slow-change type; When the proportion of mid-frequency energy exceeds the second threshold of the total energy, it is determined to be a mid-frequency oscillation type; When the proportion of high-frequency energy exceeds the third threshold of total energy, it is determined to be a high-frequency pulse type; When the energy proportion of multiple frequency bands does not exceed the corresponding proportion threshold, it is determined to be a mixed disturbance type.

5. The intelligent collaborative control system for photovoltaic, energy storage, and charging based on the energy internet according to claim 1, characterized in that, The generation of the collaborative control instruction set in the hierarchical collaborative control layer also includes an energy storage SOC partitioning adaptive strategy, which includes: The SOC of energy storage batteries is divided into three regions: deep discharge region, normal operation region, and deep charge region. When the disturbance type is determined to be high-frequency pulse type and the SOC is in the normal operating range, the high-frequency power throughput mode of the energy storage converter is activated, and the power type energy storage unit undertakes the high-frequency power component, while the energy type energy storage battery does not participate in the high-frequency charge and discharge switching. When the disturbance type is determined to be high-frequency pulse type and the SOC is in the deep discharge or deep charge zone, the high-frequency power throughput mode of the energy storage converter is prohibited. Instead, the charging pile power limit reduction strategy is activated to reduce the upper limit of the charging pile load power to a level that matches the energy storage power throughput capacity allowed by the current SOC.

6. The intelligent collaborative control system for photovoltaic, energy storage, and charging based on the energy internet according to claim 1, characterized in that, The edge controller is used to execute the following hierarchical collaborative control strategy: The first layer, the converter local control layer, is used to deploy droop control loops in energy storage converters and photovoltaic inverters. These droop control loops take the DC bus voltage deviation as input to generate the first increment of converter power regulation. The second layer, the edge coordination layer, is used to receive the collaborative control instruction set issued by the energy internet cloud platform, and convert the collaborative control instruction set into a droop coefficient correction amount that can be executed by the local control layer of the converter. This droop coefficient correction amount is superimposed on the first increment to generate the final value of the converter power reference value. The response speed of the first layer is greater than that of the second layer, and the control cycle of the first layer is less than that of the second layer.

7. The intelligent collaborative control system for photovoltaic, energy storage, and charging based on the energy internet according to claim 1, characterized in that, The data upload and collaborative control command set distribution of the multi-source sensing layer are both executed through the energy router on the energy internet cloud platform. This energy router is used for: Each photovoltaic inverter, energy storage converter, and charging pile is assigned a unique Energy Internet Protocol address; Based on the determination of the disturbance type, different transmission priorities are assigned to different instruction categories in the coordinated control instruction set. Among them, when the disturbance type is high-frequency pulse type, the transmission priority of the energy storage converter charging and discharging power reference value and its response time constant is higher than the adjustment amount of the photovoltaic inverter active power reference value.

8. The photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 1, characterized in that, The multi-source sensing layer collects the charging pile load power and its power step amplitude, including: The instantaneous value of the charging pile load power is collected at a first sampling frequency, and the sliding average value of the charging pile load power is collected at a second sampling frequency, wherein the first sampling frequency is greater than the second sampling frequency; The moment when the difference between the instantaneous value and the moving average exceeds the power step judgment threshold is marked as a power step event. The amplitude and occurrence time of the power step event are recorded. The amplitudes and occurrence times of multiple consecutive power step events are used to form a power step amplitude data sequence of the charging pile load power.

9. A photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 8, characterized in that, The pulse feature clustering includes: using the amplitude and occurrence time of the power step event as feature quantities, performing cluster analysis on the power step amplitude data sequence using a density clustering algorithm, and taking the amplitude and occurrence interval corresponding to the cluster center in the clustering result as pulse feature parameters in the second perturbation vector. The pulse feature parameters include the average pulse amplitude and the average pulse interval.

10. A photovoltaic-storage-charging intelligent collaborative control system based on the energy internet according to claim 1, characterized in that: The energy internet cloud platform is used to aggregate and analyze the first disturbance vector, the second disturbance vector, and the disturbance type determination results of multiple photovoltaic, energy storage, and charging sites to generate a regional disturbance feature map. When the frequency of high-frequency pulse-type disturbances in the time dimension of the regional disturbance feature map exceeds the preset frequency threshold, an advance switching instruction is issued to the collaborative decision engine of multiple photovoltaic-storage-charging sites. This advance switching instruction is used to instruct the collaborative decision engine to adjust the control parameters of the energy storage converter to the preparatory state of high-frequency power throughput mode when the current disturbance type is low-frequency slow-change type or medium-frequency oscillation type.

Citation Information

Patent Citations

  • Optical storage and charging cooperative control method and system for transformer area

    CN114243802A