Method and system for optimizing long-time scale energy efficiency of direct-current nano-grid cluster

Abstract

The invention relates to a long-time scale energy efficiency optimization method and system for a direct-current nano-grid cluster, and belongs to the technical field of energy efficiency optimization of a power system. The method comprises the following steps: establishing a cluster equivalent circuit model considering interconnection converter loss, and accurately describing a cluster power flow relationship; based on an equivalent circuit model, constructing a cluster global loss model, and quantifying three types of core losses of a line, a battery and a converter; establishing a cluster energy sharing model, classifying nano-networks according to net power characteristics of the nano-networks, and defining an energy interaction rule; based on the three types of models, an energy optimization management strategy is adopted, on the premise that safe operation is guaranteed, energy flow is optimized through a particle swarm optimization algorithm, and the self-sufficiency rate and energy efficiency of the cluster under the long-time scale are improved. According to the method, the cluster energy flow mechanism is optimized by establishing the cluster analysis model considering the characteristics of the interconnection converter, and the energy transmission efficiency and the self-sufficiency rate of the cluster under the long-time scale are remarkably improved on the premise of ensuring safe and stable operation.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power system energy efficiency optimization, and relates to a long-time-scale energy efficiency optimization method and system for a direct-current micro-grid cluster. BACKGROUND

[0002] The construction of new power systems mainly based on new energy is accelerating, and the direct-current micro-grid cluster, as an important part of the new power system, has significant advantages in distributed energy consumption and local energy utilization, and has been widely used in residential communities, commercial buildings, and power supply in remote areas. However, the source-storage-load operation range of the direct-current micro-grid cluster is wide, and there are distributed line losses, battery internal resistance losses, and interconnection converter losses in the energy transmission process, among which the interconnection converter losses account for the main part of the total losses of the cluster, resulting in low overall energy efficiency (usually less than 50%) of the cluster, which seriously limits its large-scale popularization and application.

[0003] The existing energy efficiency optimization research of the direct-current micro-grid cluster mainly focuses on the device level, and reduces the converter self-loss by improving the modulation strategy of the interconnection converter and optimizing the circuit parameters. However, the optimization at the device level can only improve the overall efficiency of the converter, and cannot solve the problem of large efficiency difference at different power working points, and it is difficult to reduce the distributed line loss and battery internal resistance loss.

[0004] At the system level, the existing energy management strategy focuses on supply-demand balance and safe and stable operation, and does not fully consider the influence of interconnection converter loss on the energy efficiency of the cluster, resulting in unreasonable energy flow mechanism of the cluster, deviation of the converter working point from the high-efficiency region, and further restriction on the improvement of the energy efficiency of the cluster. At the same time, the existing equivalent circuit model of the cluster does not consider the operation characteristics of the interconnection converter, and cannot accurately describe the power flow relationship of the cluster, which affects the optimization effect of the energy management strategy. Therefore, a long-time-scale energy efficiency optimization method considering the device characteristics and system coordination is needed to optimize the energy flow mechanism of the cluster from the system level and realize the comprehensive improvement of the energy efficiency and self-sufficiency rate of the cluster. SUMMARY

[0005] Therefore, the purpose of the present application is to provide a long-time-scale energy efficiency optimization method and system for a direct-current micro-grid cluster, which establishes a cluster analysis model considering the characteristics of the interconnection converter, optimizes the energy flow mechanism of the cluster, and significantly improves the energy transmission efficiency and self-sufficiency rate of the cluster at the long-time-scale under the premise of ensuring safe and stable operation.

[0006] To achieve the above purpose, the present application provides the following technical solutions: A direct current nanogrid cluster long time scale energy efficiency optimization method is applied to a direct current nanogrid cluster (NGC) formed by a plurality of nanogrids (NGs) interconnected through distributed lines and interconnected DC-DC converters, wherein the interconnected DC-DC converters include two-port double active bridge (DAB) converters and three-port triple active bridge (TAB) converters; the method specifically comprises the following steps: S1, an equivalent circuit model of the cluster considering the loss of the interconnected converter is established to accurately describe the power flow relationship of the cluster; S2, based on the equivalent circuit model, a global loss model of the cluster is constructed to quantify the core losses of the three types of lines, batteries and converters; S3, a cluster energy sharing model is established, the NGs are classified according to the net power characteristics of the nanogrids, and the energy interaction rules are clarified; S4, based on the above three types of models, an energy optimization management strategy (EOMS) is adopted to optimize the energy flow through the particle swarm optimization (PSO) algorithm under the premise of ensuring safe operation, and the self-sufficiency rate and energy efficiency of the cluster under the long time scale are improved.

[0007] Further, the core idea of model construction includes: Based on the topological characteristics of the "source - storage - load - interconnection unit" in the direct current nanogrid cluster (NGC), the limitation of ignoring the loss of the interconnected converter in the traditional model is broken through, and the equivalent circuit model which can accurately describe the power flow relationship is established through element equivalent mapping and loss characteristic embedding. The core logic is: each functional unit in the cluster is converted into an equivalent circuit element, and the voltage matching function and energy transmission function with accompanying loss of the interconnected DC-DC converter are mainly represented by "ideal transformer + variable current source" combination, and finally a quantifiable and calculable power flow analysis model is formed.

[0008] Further, the cluster power flow relationship is quantitatively described: Based on the above equivalent elements, the equivalent circuit model of the cluster is constructed, and the power flow relationship is quantitatively described by Kirchhoff's law (KVL, KCL), and the cluster DC bus voltage V bus，0 (stable at 380V) and the bus voltage of the first i NG satisfy:

[0009] Wherein, i NG,i is the current (positive for flowing into the NG and negative for flowing out) of the first i NG and other nodes of the cluster, R line,i is the equivalent resistance of the distributed line.

[0010] Power balance relationship, internal power balance of NG: the power balance relationship of the firsti The interaction power of each NG P NG,i, load power P load,i, converter bus power P con,i satisfies:

[0011] in, P con,i represents the power on the converter bus side, which needs to take into account converter losses, i.e.:

[0012] This formula breaks through the assumption of "equal input and output power of converter" in the traditional model and accurately reflects the impact of losses on power flow.

[0013] Cluster global power balancing, power exchange among all NGs, and distributed line losses. P l oss，line Interaction power between the cluster and external clusters P net satisfy:

[0014] in , P net represents the interaction power between the cluster and external clusters (positive for absorption, negative for output).

[0015] Preferably, in step S1, a mechanism can be provided to ensure the accuracy of the model.

[0016] Loss model adaptation: The loss values ​​of the converter's variable current source can be directly applied to the pre-established refined loss model (including conduction losses). P con, switching loss P sw, copper consumption P cu, core loss P m), ensuring that the loss calculation error is controlled within 1.2% (experimental verification). Dynamic parameter correction: Battery port voltage is corrected in real time with the SoC (based on a fitting formula), photovoltaic port voltage... v pv,i is dynamically adjusted according to the light intensity to ensure the real-time performance of the equivalent value of the current source. Constraint embedding: Embedding bus voltage fluctuations into the model ( v Constraints such as bus,i fluctuating between 361V and 399V and battery SoC (fluctuating between 0.2 and 0.8) ensure that power flow calculations meet the requirements of the safe operating domain. The equivalent circuit model constructed by the above method can accurately describe the power flow path and loss distribution of "photovoltaic - energy storage - load - converter - line" in the cluster, and provide accurate model support for the development of subsequent cluster energy optimization management strategy (EOMS). Experimental verification shows that the error of the model in describing the power flow of the cluster is less than 3%.

[0017] Preferably, in step S2, the equivalent circuit model is used to construct a cluster global loss model to quantify the line, battery, and converter core losses. In order to provide real-time power loss estimation and improve energy transmission efficiency, it is necessary to further establish a power loss model of the cluster. The loss model can reflect the functional relationship between the power transmission loss of the cluster and the power flow mode of the cluster, and is the model basis for realizing the energy efficiency optimization of the cluster. As mentioned earlier, the energy transmission efficiency of the direct current nanogrid cluster is affected by the "source", "transmission", and "storage" links. "Source" represents the energy conversion efficiency of the photovoltaic array and the utilization rate of the photovoltaic array; "transmission" represents the transmission line loss and interconnection converter loss; "storage" represents the energy conversion efficiency of the energy storage unit. The energy conversion efficiency of the photovoltaic array and the energy storage unit is mainly related to its own architecture and material selection, and the energy management strategy of the cluster has little effect on it.

[0018] Distributed line loss:

[0019] Battery internal resistance loss: ignoring the imbalance problem of single battery, the second-order RC equivalent circuit model of the battery pack composed of n batteries in series and m batteries in parallel, the equivalent resistance Rpe,q, Ro,eq and Rs,eq of the battery pack, and the open circuit voltage Voce,eq are respectively:

[0020]

[0021] In this method, the battery is NCR18650B ternary lithium battery produced by Panasonic Company. The rated capacity of the single battery is 3350mAh, and the port rated voltage is 3.6V. By performing pulse discharge experiment on the battery, the curve of the related parameters of NCR18650B changing with voltage is measured, and it can be observed from the curve that R p and R s resistance is less than 0.01 ohm, and R0 is almost six times larger than Rp and Rs, so the loss caused by R p and R s is ignored, and the loss of the battery internal resistance can be represented as:

[0022] Considering the working life and safety of the battery, the working range of the battery SoC is limited to the range of 0.2~0.8, in which R o≈0.06Ω, battery port current i bat,i is the following formula:

[0023] V ocv represents the open circuit voltage of the battery, when the SoC is in the range of 0.2~0.8, V oce is fitted according to the change of SoC as follows:

[0024] According to KVL, NG battery port voltage in i v The relationship between bat,i and open circuit voltage can be expressed as

[0025] Solving simultaneously i bat,i, v bat,i, p bat,i, SoC and loss p loss, bat.

[0026] Preferably, in step S3, the cluster energy sharing model is established, the NG is classified according to the NG net power characteristics, and the energy interaction rules are clarified.

[0027] Preferably, in step S4, the comprehensive optimization modulation scheme is calculated by the following steps: S401: The range of the cluster safe working domain is derived in turn to derive the optimization variable P NG,i , P bat,i and P pv,i the value range of; S402: The power interaction between the nets and the interaction power between the clusters will be determined based on the energy sharing principle of high self-sufficiency rate P NG,i and the value of net P ; S403: The optimal modulation scheme is obtained by minimizing the cluster energy transmission loss algorithm; Preferably, in step S401, the value range of the optimization variable P NG,i , P bat,i and P pv，i is derived.

[0028] The range of variation of

[0029] To ensure the bus power quality, the bus voltage v bus,i is usually required to be within ±5%. The corresponding P NG,i will be agreed within a certain range, where the maximum and minimum interaction power are

[0030]

[0031] P The range of variation of Taking into account the aging process, life and safety of the battery, the battery current i bat,i needs to be limited within I bat,min,i , I bat,max,i The minimum and maximum current emitted by the battery I bat,min,i and I bat,max,i are respectively:

[0032]

[0033] where and are the maximum current limits of the battery port, and are the limit currents determined by the range of variation of the battery SoC, which is generally limited to [0.2~0.8], based on the Coulomb counting method, and can be expressed as

[0034]

[0035] where Qbat is the battery capacity, in ampere. hour, is the battery energy conversion efficiency, is the sampling time or running time window.

[0036] To solve the battery port power P bat,ithe corresponding SoC range [SoC min,i , SoC max,i ] needs to be recalculated according to the battery terminal current range derived in the above equation; then the range of the battery port voltage change [ V bat,min,i , V bat,max,i ] can be solved, and the range of the battery port power P bat,i can also be represented as:

[0037]

[0038] the actual power generation of the photovoltaic unit P pv，i To improve the utilization of distributed renewable energy in the power system, we usually expect the photovoltaic port to operate in MPPT mode. However, in practice, affected by the energy storage capacity and other factors, the photovoltaic unit sometimes needs to exit the MPPT mode to reduce power generation. Then when to exit the MPPT mode and reduce how much power generation? This needs to be considered comprehensively based on the bus power quality, battery SoC state, maximum power output of the photovoltaic unit in MPPT mode, and load power consumption, etc.

[0039] Based on the range of the grid interaction power P NG,i bus and the range of the battery port power P bat,i solved in the previous, the actual maximum power generation of the photovoltaic port P pv,max,i can be derived here. P pv,max,i is the maximum power that the photovoltaic unit can generate under the conditions of meeting the load power demand and avoiding the over-limit of the cluster bus voltage quality and battery overcharge, which is expressed as

[0040] Combined with the P-V curve of the photovoltaic unit, the maximum power output of the photovoltaic unit P pv,max,i and its corresponding terminal voltage v pv,i can be solved. Because the maximum power output of the photovoltaic unit cannot exceed P pv,MPPT,i , when P pv,max,i ≥ P pv,MPPT,i , the photovoltaic port can work in MPPT mode. WhenP pv,max,i ≤ P pv,MPPT,i When the port voltage v pv,i deviates from the MPPT, the PV unit power generation is reduced to avoid overcharging of the battery or overvoltage of the bus. In summary, the actual power generation of the PV port P pv,i can be expressed as:

[0041] In addition, for the grid without installed PV panels, the actual power generation of the PV port P pv,i can be set to 0 in the subsequent calculation.

[0042] Preferably, in step S402, the cluster self-sufficiency rate optimization is one of the core goals of the direct current nanogrid cluster energy optimization management strategy (EOMS), and the core logic is to prioritize energy interaction between internal nanogrids (NGs) in the cluster to meet supply and demand balance, reduce energy interaction with external clusters, thereby reducing additional losses from external interaction and enhancing the cluster's energy self-sufficiency. The specific implementation steps are as follows: The cluster self-sufficiency rate K ss represents the replacement ratio of PV energy within the cluster to the load demand, reflecting the cluster's self-sufficiency ability, and the calculation formula is:

[0043] wherein : total load demand of all NGs in the cluster (unit: kW); P net : interaction power between the cluster and external clusters (positive for absorbing external energy, negative for outputting energy to the outside) Kss The value range is [0, 1], and the closer the value is to 1, the stronger the cluster's self-sufficiency and the lower the external dependence.

[0044] Optimization goal: by formulating the rule of "internal energy interaction first, external interaction only as a supplement", the K ss is maximized, ultimately achieving high self-sufficiency rate of the cluster in a long time scale (such as weekly average self-sufficiency rate ≥ 85%).

[0045] Determination of NG interaction power range Before optimizing the self-sufficiency rate, the interaction power limit of each NG needs to be determined based on the previously calculated "safe operating domain constraint" to provide boundaries for subsequent energy interaction decision-making: Derivation of NG interaction power limit: combined with the bus voltage fluctuation constraint ( v bus,i∈[380V±5%]) and Kirchhoff's voltage law (KVL), the first i interaction power range of the NG with other nodes in the cluster: P NG,i ∈[ P NG,min,i , P NG,max,i ] where, P NG,min,i NGminabsorbed power (NG-state), P NG,max,i NGmaxoutput power (NG+state), calculated by: P NG,min,i = P pv,i + P bat,min,i + P load,i - P loss,con,i P NG,max,i = P pv,i + P bat,max,i + P load,i - P loss,con,i where, P pv,i PVactual output, P bat,min,i / P bat,max,i battery charge and discharge power limit, P loss,con,i interconnection converter loss.

[0046] NG role classification: according to the positive and negative of P NG,i , the NG is divided into two categories: NG+ (energy surplus type): P NG,i > 0, PV output + battery discharge power (deducting loss) can meet its own load, and there is surplus energy to participate in internal sharing; NG- (energy shortage type): P NG,i < 0, PV output + battery discharge power (deducting loss) cannot meet its own load, and needs to absorb energy from other NGs or the outside.

[0047] Formulate energy interaction rules for different scenarios: Based on the interaction power range of all NGs in the cluster, the net power range of the entire cluster is calculated, and three types of scenarios are formulated to optimize the self-sufficiency rate: , ] and formulate energy interaction strategies for three types of scenarios to achieve self-sufficiency rate optimization: Scenario 1: Energy shortage within the cluster ( <0) Judgment basis: The sum of the maximum output power of all NGs (NG+ total surplus) is less than the sum of the minimum absorption power of all NGs (NG- total gap), and the internal supply and demand cannot be balanced, and external energy is needed. Interaction rules (minimize external dependence): NG+ maximize output: all NG+ according to their maximum output power P NG,max,i Output energy to the cluster, fully utilize internal surplus; NG- according to the minimum absorption power of all NG- P NG,min,i Absorb energy from the cluster, prioritize filling the load gap External supplement gap: the power Pnet absorbed from the outside of the cluster is only used to supplement the internal gap, and the calculation formula is Self-sufficiency rate characteristics: in this scenario K ss <1, but through internal maximum sharing, the external dependence has been minimized. Scenario 2: Energy surplus within the cluster ( >0) Judgment basis: The sum of the minimum output power of all NGs (NG+ minimum surplus) is greater than the sum of the maximum absorption power of all NGs (NG- maximum gap), and the cluster has excess energy, which needs to be output to the outside to avoid battery overcharging.

[0048] Interaction rules (preferentially consume internal photovoltaic) NG+ minimize output: all NG+ according to their minimum output power ( P NG,min,i ) output energy, avoid excessive battery discharge (protect battery life); NG- maximize absorption: all NG- according to their maximum absorption power ( P NG,max,i ) absorb internal energy, fully utilize internal surplus; External surplus consumption: the power ( P net ) output to the outside of the cluster is only the internal surplus photovoltaic energy (preferentially abandon light and minimize), and the calculation formula is:

[0049] Self-sufficiency rate characteristics: in this scenarioK ss = 1, the cluster is completely self-sufficient, and surplus energy is output to the outside, without external dependence. Scenario 3: The cluster is self-sufficient internally <0 ) Judgment basis: The total surplus energy of NG+ in the cluster can cover the total gap energy of NG-, without the need for external interaction, which is the optimal self-sufficient scenario.

[0050] Interaction rules (close external interaction, optimize internal distribution): 1. Prohibit external interaction: set P net = 0, all energy requirements are met through internal interaction of NG+ and NG-; 2. Internal power optimization distribution: adjust the interaction power of each NG through particle swarm optimization (PSO) algorithm to minimize the total loss of the cluster P NG,i (need to meet P NG,i ∈[ P NG,min,i , P NG,max,i ), so that the surplus energy of NG+ accurately matches the gap energy of NG-; 3. Priority constraint: the energy absorbed by NG- is only used to fill the load gap and cannot be used for additional battery charging (to avoid energy conversion loss); the energy output by NG+ is prioritized for photovoltaic surplus, followed by available battery discharge power (SoC≥0.2).

[0051] Self-sufficiency rate characteristics: under this scenario K ss = 1, the cluster is completely self-sufficient, and the internal energy flow loss is minimized, which is the optimal balance state of energy efficiency and self-sufficiency rate.

[0052] Key constraints: ensure the safety of self-sufficiency rate optimization In the above scenario decision, the following constraints must always be met to avoid equipment damage or excessive power quality due to the pursuit of high self-sufficiency rate: Battery SoC constraint: NG+ battery SoC participating in discharge ≥0.2 (to avoid over-discharge), and NG battery SoC participating in charging ≤0.8 (to avoid overcharge); Bus voltage constraint: the bus voltage of all NGs v bus,i must be stable at 380V±5%, to avoid voltage out-of-limit due to energy interaction; Photovoltaic output constraint: the photovoltaic output of NG+ P pv,i ≤ P pv,MPPT,i (maximum MPPT output), to avoid photovoltaic unit overload. Preferably, the step S403 is solved by minimizing the cluster energy transmission loss algorithm to obtain the optimal modulation scheme. The optimization objective function is a high-order non-convex non-differentiable function.

[0053] Optimization effect verification: Based on the cluster parameters (5 NGs, line resistance 3Ω / 3Ω / Ω / 4Ω / 1Ω, battery SoC initial 0.2~0.5), simulation is carried out with one week as the time span, and the effects of "lossless optimization" and "the optimization strategy of the present section" are compared: 1. Loss reduction: after adopting the strategy of the present section, the total loss of the cluster per week is reduced from 45.4kWh (without optimization) to 26.3kWh, wherein: the distributed line loss is reduced by 65.2% (from 13.8kWh to 4.8kWh); the battery internal resistance loss is reduced by 56% (from 5.2kWh to 2.3kWh); the interconnection converter loss is reduced by 28% (from 26.5kWh to 19.2kWh); 2. Energy efficiency improvement: the energy transmission efficiency of the cluster is improved from 86.4% (without optimization) to 91.7%, which is about 6% higher than the system level without optimization; 3. Converter efficient operation: after optimization, the working point of the interconnection converter converges to the high-efficiency area (such as the DAB converter efficiency is maintained above 95%, and the TAB converter efficiency is maintained above 96%), which further verifies the effectiveness of the loss optimization.

[0054] The present application has the following advantages: In the practice of DC nanogrid cluster energy efficiency optimization, by constructing a cluster equivalent circuit model considering the operating characteristics of interconnected converters, distributed lines are equivalent to precise line resistances, generalized loads are equivalent to constant current sources, and battery and photovoltaic ports are equivalent to variable current sources. The interconnected DC-DC converter is innovatively equivalent to a combination of an "ideal transformer + variable current source" (the ideal transformer matches different port voltage levels to achieve voltage adaptation, and the variable current source accurately represents the conduction loss, switching loss, and magnetic element loss during energy transmission). This breakthrough overcomes the limitations of traditional models that ignore converter losses, leading to distorted power flow descriptions. It can clearly quantify the complete power flow path and power distribution relationship of each link in the "photovoltaic-energy storage-load-converter-line" within the cluster, providing high-precision, calculable model support for the development of subsequent energy efficiency optimization strategies, ensuring that optimization decisions are always based on real cluster operating conditions. On this basis, by integrating device loss characteristics and system energy flow optimization, the interconnected converter loss (accounting for more than 60% of the total cluster loss), distributed line loss, and battery internal resistance loss are included in a unified optimization framework. Through collaborative optimization under multiple constraints (both meeting basic safety constraints such as bus voltage ±5% fluctuation limit, battery SoC 0.2~0.8 safety range, and photovoltaic output not exceeding MPPT limit, and dynamically adjusting the interactive power between nanogrids through the particle swarm optimization algorithm), not only does the interconnected converter operating point converge to the high-efficiency region (such as the optimized DAB converter efficiency maintaining above 95% and the TAB converter efficiency maintaining above 96%), effectively reducing the converter's own loss, but also reduces the joule heat loss of the distributed line by optimizing the energy flow path (line loss reduced by 65.2% compared to the non-optimized scheme), while avoiding the surge in internal resistance loss caused by excessive battery charging and discharging (battery internal resistance loss reduced by 56% compared to the non-optimized scheme). Ultimately, the cluster energy transmission efficiency is significantly improved from 86.4% to 91.7%. In addition, the strategy always follows the "internal energy interaction first" principle, accurately determines the cluster net power range (if the internal NG+surplus can cover the NG-gap, external interaction is closed, and only when there is an internal energy shortage or surplus does external interaction start, and the interaction power is strictly controlled to only supplement the gap or absorb surplus photovoltaic power), significantly reducing the additional line loss and converter loss caused by energy interaction between clusters, increasing the cluster's average self-sufficiency rate from 34.9% to 85.1%, and significantly enhancing the cluster's energy self-sufficiency ability and reducing its dependence on external energy.The optimization scheme is also particularly suitable for long-time scale energy efficiency scheduling. Real-time data such as photovoltaic output, load demand, battery SoC and the like are dynamically collected at a time interval of 30 minutes, and the random fluctuations of photovoltaic and load (such as sudden change of light intensity, sudden increase and decrease of load) are quickly responded to through offline optimization + online table lookup, so that stable optimization effect can be maintained under different working conditions, energy consumption increase caused by frequent adjustment of short-time scale scheduling is avoided, and high efficiency and reliability of long-term operation of the cluster are ensured, thereby providing complete technical support for large-scale efficient operation of various direct current power grid clusters such as communities, industrial parks and remote areas, and assisting in efficient consumption and local energy utilization of distributed energy in the new power system.

[0055] Other advantages, objects, and features of the application will be in part apparent and in part pointed out hereinafter. The objects and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and claims hereinafter as well as the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to make the objects, technical solutions and advantages of the present application clearer, the preferred embodiments of the present application will be described in detail below with reference to the accompanying drawings, in which: Figure 1 The logic block diagram of the comprehensive optimization modulation method in the embodiment is shown in the figure; Figure 2 The simulation parameters of the direct current power grid cluster are shown in the table; Figure 3 The load demand and photovoltaic power generation of the NGs in a week are shown in the table; Figure 4 The main differences of the four schemes are shown in the table; Figure 5(a) is the battery charging and discharging power Pbat,i and SoC, and figure 5(b) is the power grid interaction power P NG,i and bus voltage fluctuation v bus,i ; Figure 6 The total photovoltaic maximum / actual power generation and photovoltaic utilization rate K pv ; Figure 7 The cluster interaction power P net and cluster self-sufficiency rate K ss ; Figure 8 The cluster self-sufficiency rate of scheme 1 and scheme 2; Figure 9 The cluster energy transmission loss of the four schemes. Detailed Implementation

[0057] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0058] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures, and should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0059] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0060] In this invention, the objective function is a high-order non-convex, non-differentiable function. Considering the complexity of the optimization problem, the PSO optimization method is used for offline optimization.

[0061] The specific implementation steps are as follows: I. Core Objectives and Optimization Targets Optimization Objective: With minimizing total cluster loss as the core objective, and based on the "cluster loss model," an objective function is established. By adjusting the interaction power PNG,i between NGs, the total sum of various losses during cluster energy transfer is reduced to the minimum. Optimization target: The optimization variable is the interaction power between each NG and other nodes within the cluster. P NG,i (i=1,2,...,n), where n is the number of NGs in the cluster, and the following safety constraints must be met: P NG,i ∈ [P NG,min,i , P NG,max,i Power balance constraint: =0 (self-sufficient within the cluster, no external energy interaction, P net =0).

[0062] II. Key premise: clearly define the total cluster loss composition Total cluster loss P loss,NGC is composed of three parts, and all are directly related to the optimization variables P NG,i , providing the basis for the construction of the objective function: III. Core step: build optimization problem and solve Establish an optimization mathematical model (1) Objective function Aggregate the three types of losses to build a function with the goal of "minimum total cluster loss"

[0063] where, i NG,i , i bat,i , P loss,con,i are functions of P NG,i , the objective function presents high-order nonlinearity and non-convexity, which cannot be solved by traditional analytical methods, and intelligent optimization algorithms need to be used.

[0064] (2) Constraint conditions Inequality constraints: each NG interaction power needs to be within the safe range, i.e. P NG,min,i ≤ P NG,i ≤ P NG,max,i (i=1,2,...,n); equality constraints: Power balance within the cluster. Solve using the particle swarm optimization (PSO) algorithm Due to the non-linear characteristics of the objective function, this section uses the PSO algorithm for offline optimization, through the iterative search of particles within the feasible region, to find the optimal P NG,i combination, the specific steps are as follows: (1) Initialize the population - Particle definition: Each particle represents a set of candidate optimization solutions, and the position vector is X j = [ .... ](j=1,2,...,N), (N) is the number of particles, usually take 50~100), the velocity vector is V j = [ V j1 , V j2 ,..., V jn ] (the speed range is set to [-1,1], to avoid the particle search too fast out of the feasible region); initial position generation: based on the safety range of P NG,i , randomly generate initial position: P j NG,i = P NG,min,i +( P NG,max,i - P NG,min,i ). rand(0,1) is a [0,1] random number); initial fitness value: for each particle, put into the objective function to calculate the total loss of cluster , as the initial fitness value of the particle (the smaller the fitness value, the better the solution), and initialize the individual optimal fitness value Pfbest j = P j loss,NGC , the individual optimal position Pxbest J = X j , the global optimal fitness value Gfbest = min( Pfbest j ), the global optimal position Gxbest = X j )( X j ) is the particle position corresponding to Gfbest ). (2) iterative optimization by updating the velocity and position of the particle, gradually converging to the global optimal solution, the iterative formula is:

[0065] wherein: k is the number of iterations; is the inertia weight (take 0.7, balance the global search and local search ability); c 1、 c 2 is the learning factor (both take 2.0, respectively control the trend of the particle to the individual optimal, global optimal position); r 1、 r2 is a [0, 1] random number, increasing search randomness. (3) Constraint check and fitness value update Constraint check: check the position of the particle after iteration , whether it meets P NG,i ∈ [ P NG,min,i , P NG,max,i ] and , if not, re-generate the position; Fitness value update: for particles that meet the constraints, calculate the new fitness value , if < Pfbest j , update ; Global optimal update: compare the fitness values of all particles Pfbest j , if there is Pfbest j < Gfbest , update Gfbest = Pfbest j , Gxbest = Pxbest j .

[0066] (4) Convergence judgment When the global optimal fitness value Gfbest changes by Gfbest k+1 - Gfbest k | <0.001 (or the number of iterations reaches the maximum threshold, usually set to 1000 times), iteration is terminated, and the global optimal position Gxbest , i.e. the optimal NG interaction power .

[0067] A long-time-scale energy efficiency optimization method for a DC nanogrid cluster is disclosed in this embodiment.

[0068] As shown in Figure 1 , the long-time-scale energy efficiency optimization method for a DC nanogrid cluster comprises the following steps: S1: Establish a cluster equivalent circuit model considering interconnection converter loss to accurately describe the power flow relationship of the cluster; S2: Based on the equivalent circuit model, build a cluster global loss model to quantify the core losses of three types of lines, batteries and converters; S3: Establish a cluster energy sharing model, classify NGs according to the NG net power characteristics, and clarify the energy interaction rules; S4: Based on the above three types of models, an energy optimization management strategy (EOMS) is proposed, which optimizes energy flow through particle swarm optimization (PSO) algorithm under the premise of ensuring safe operation, and improves the self-sufficiency rate and energy efficiency of the cluster in a long time scale.

[0069] In the present application, by constructing a cluster equivalent circuit model considering the operating characteristics of interconnected converters, a global loss model integrating device loss characteristics, and an energy sharing model based on nanometer network net power classification, and combining particle swarm optimization (PSO) algorithm to generate energy optimization management strategy (EOMS), the energy transmission efficiency and self-sufficiency rate of the cluster in a long time scale are taken as double targets, the bus voltage fluctuation, battery SoC range, and photovoltaic output limit are taken as constraints for optimization scheduling, thereby obtaining lower cluster total loss and higher energy self-sufficiency, i.e. ensuring safe and stable operation of the cluster and improving overall energy efficiency, which can further optimize the operating point of the interconnected converter to reduce converter loss under the premise of reducing distributed line loss and battery internal resistance loss, thereby improving the energy utilization efficiency and power supply autonomy of the cluster, thereby ensuring efficient operation and long-term reliability of the DC nanogrid cluster. At the same time, by formulating energy interaction rules (internal sharing first, external supplement only for backup) in different scenarios, energy flow decision only needs to be made for different cluster net power scenarios, thereby simplifying the subsequent optimization calculation process and improving the efficiency of strategy implementation. In addition, the particle swarm optimization algorithm is used to solve the nonlinear loss optimization problem under multiple constraints, which is more suitable for high-order non-convex objective functions compared to traditional analytical methods, and is an efficient method for handling multi-variable coupled optimization of clusters, so the optimization based on this algorithm can more effectively find a global optimal energy flow scheme that takes into account energy efficiency and self-sufficiency rate.

[0070] In the specific implementation process, combined with Figure 2 as shown, Figure 2The specific parameters of the DC nanogrid cluster are given. In order to verify the effectiveness of the cluster energy optimization management strategy EOMS proposed in the application, the DC nanogrid cluster composed of five autonomous nanogrids is taken as an example, and the cluster energy scheduling mechanism of EOMS is analyzed in detail through MATLAB simulation. Among the five NGs, NG3 is not configured with photovoltaic panels, and the battery in NG3 is connected to the DC bus through the DAB converter; in NG1, NG2, NG4 and NG5, the battery and the photovoltaic panel are connected to the DC bus through the TAB converter. The time span of the simulation is set to one week, and the time interval is set to 30 minutes. In each time interval, the cluster coordination controller monitors the real-time state of each module in the cluster by communicating with the controllers inside the NGs; through the cluster communication link, the data including SoC, photovoltaic power generation, load demand collected from each NG; based on these data, the cluster controller runs the optimization algorithm and issues commands to the nanogrid controller; the nanogrid controller executes the upper-layer instructions, and finally realizes the optimization management of the cluster energy flow.

[0071] Figure 3 shows the real-time load demand, actual photovoltaic power generation and maximum power generation of the MPPT mode of the five NGs in the DC nanogrid cluster within a week. As shown in the figure, the total load demand of the DC nanogrid cluster within a week is 290.45 kWh / week, and the total maximum photovoltaic power generation under the MPPT mode within a week is 314.57 kWh / week.

[0072] In order to respectively illustrate the optimization modulation and parameter optimization of the equipment (converter) level and the energy management of the system level on the energy efficiency of the cluster, as shown in Figure 4 As shown in the table, scheme 1 optimizes the cluster from both the equipment and system levels, the system level NGC adopts the EOMS proposed in this chapter for energy management, the circuit parameters of the interconnection converter at the equipment level adopt the optimal parameter combination, and the modulation mode adopts the near-all-ZVS optimization modulation strategy. Scheme 2 only optimizes the cluster from the equipment level, the system level NGC does not perform energy optimization management (i.e., only ensures that the cluster operates in the safe operating domain, and does not optimize the self-sufficiency rate and energy transmission efficiency of the cluster), and the optimization at the equipment level is the same as scheme 1. Scheme 3 only optimizes the cluster from the system level, the system level NGC adopts the EOMS proposed in this chapter for energy management, and the circuit parameters and modulation mode of the interconnection converter at the equipment level are not optimized, the circuit parameters are selected as the typical combination 1 in the fourth chapter, and the modulation mode adopts the traditional modulation strategy. Scheme 4 does not optimize the NGC from both the equipment and system levels.

[0073] In order to quantitatively compare the performance of different schemes, the self-sufficiency rate of the clusterK ss The photovoltaic utilization rate and the energy transmission efficiency are defined as follows: The cluster self-sufficiency rate Kss represents the percentage of load demand replaced by photovoltaic energy in total load demand. The larger Kss is, the less energy interaction with other clusters, and the higher the cluster self-sufficiency rate is.

[0074]

[0075] Photovoltaic utilization rate K pv In order to quantify the utilization of photovoltaic energy in the cluster, the index K pv , K pv is introduced, which represents the percentage of actual power generation of photovoltaic to its maximum power generation.

[0076]

[0077] Energy transmission efficiency η et : η et The larger the energy transmission efficiency is, the less the loss in the energy transmission process, and the more the energy transmitted to the demand side. η et is defined as

[0078] In the implementation process, taking scheme 1 as an example, the real-time simulation results of scheme 1 are generated as shown in FIG. 5. FIG. 5(a) is the charging and discharging power of the five batteries and the corresponding SoC values. FIG. 5(b) is the interaction power between the five NGs, and the bus voltage fluctuation accompanying the power interaction process. It is observed that under the EOMS scheduling, the five NGs constantly change the NG+ and NG- roles in the energy interaction process, realize the balance of their own supply and demand through energy interaction and sharing with other NGs. At the same time, the EOMS successfully limits the battery state of charge to the range of 0.2~0.8 and stabilizes the bus voltage fluctuation within ±5% by constraining the charging and discharging power of the battery and the nanonet interaction power, realizing the safe and stable operation of the cluster.

[0079] Figure 6 The actual total power generation of photovoltaic at each moment (orange dotted line) and the total power generation of MPPT (blue solid line) are given, and the real-time change of the photovoltaic utilization rate of NGC K pv is drawn. It can be observed that most of the time the photovoltaic unit operates in MPPT mode, Kpv It is 100%. At very rare moments, such as t3, to prevent the bus voltage from exceeding the limit, the power output of NG needs to be reduced, so at that moment... K pv There was a slight decrease. However, overall, under the optimized management of EOMS, the cluster's weekly average photovoltaic utilization rate can reach 97%. Figure 7 It also demonstrated the cluster's interaction power. P net Its corresponding cluster self-sufficiency rate K ss The changes are shown in the figure. P net A value less than 0 indicates that NGC is self-sufficient, but to prevent oversupply within the cluster, it needs to release some energy to other clusters. In this case, the cluster's... K ss =100%. P net A value greater than 0 means that NGC cannot be self-sufficient and needs to absorb energy from other clusters to meet its own load requirements. In this case, the cluster's... K ss <100%. Overall, under the optimized management of EOMS, the cluster's weekly average self-sufficiency rate is approximately 85.1%.

[0080] In the specific implementation process, in order to quantitatively verify the improvement of cluster self-sufficiency rate by EOMS, Figure 8 Simulation results for Scheme 1 and Scheme 2 were compared. The figure shows the interaction power of the clusters under the two schemes. P net and cluster self-sufficiency K ss The changes were observed. It was found that because Scheme 2 did not optimize the energy flow management of the cluster, there was less energy interaction between NGs within the cluster. To ensure the supply and demand balance of NGs, the cluster had to frequently interact with other clusters for energy. The weekly average self-sufficiency rate of Scheme 2 was only 34.9%, nearly half that of Scheme 1 (85.1%) based on EOMS. This indicates that the proposed EOMS can indeed effectively promote energy sharing among NGs within the cluster, significantly enhance the cluster's autonomy, and improve its self-sufficiency rate.

[0081] To quantitatively illustrate the significance of equipment and system-level optimizations for improving cluster energy efficiency, Figure 9The loss distribution of the four schemes is given respectively, including the distributed line loss, the battery internal resistance loss and the interconnection converter loss. It is observed that the interconnection converter loss accounts for the largest proportion of the total loss of the cluster in any of the schemes. This also shows that the key to improving the energy transmission efficiency of the cluster lies in reducing the loss of the interconnection converter.

[0082] Firstly, to quantitatively verify the influence of the optimization at the converter level on the energy transmission efficiency of the cluster, the following compares scheme 1 and scheme 3. As shown in Figure 9 Since both scheme 1 and scheme 3 optimize the management of the energy flow of the cluster based on EOMS, the distributed line loss and the battery internal resistance loss of the cluster are minimized. However, since scheme 3 does not optimize the modulation mode and the circuit parameters of the interconnection converter, the power loss of the interconnection converter is very large, which is almost 4.5 times that of scheme 1. In terms of total loss, compared with scheme 3, scheme 1 reduces the total loss of the cluster by 64.9 kWh / week, and improves the energy transmission efficiency of the cluster from 76.1% to 91.7%. This result quantitatively proves the importance of parameter and modulation optimization at the converter level to improve the energy transmission efficiency of the cluster.

[0083] To improve the energy efficiency of the cluster, the present specification takes a direct current nanogrid cluster composed of TAB and DAB converters as the research object, takes improving the energy transmission efficiency and the self-sufficiency rate of the cluster at a long time scale as the goal, and proposes an energy optimization management strategy considering the loss of the interconnection converter. The present specification first considers the loss problem of the converter in the energy conversion process in combination with the working characteristics of the interconnection converter, establishes an equivalent circuit model of the cluster considering the operating characteristics of the interconnection converter. The model introduces an ideal transformer and a variable current source to represent the voltage matching function and the energy transmission function with accompanying loss of the converter on the basis of the traditional model, and more accurately clarifies the power flow relationship of the cluster. Then, in combination with the cluster analysis model, the present specification proposes an EOMS energy optimization management strategy, which can significantly improve the self-sufficiency rate and the energy transmission efficiency of the NGC by optimizing the energy flow among the NGs and the energy flow among the clusters on the premise of guaranteeing the bus voltage quality, the battery SoC and the photovoltaic utilization rate. Finally, four schemes are simulated within a week. It is found that: (1) The EOMS proposed can constrain the bus voltage fluctuation, the battery SoC, the photovoltaic power generation and the like within the required range to ensure the safe and stable operation of the cluster. (2) The optimization at the system level has significant significance to improve the energy transmission efficiency and the self-sufficiency rate of the cluster. Compared with scheme 2 which does not perform optimization at the system level, the optimized scheme (scheme 1) can improve the self-sufficiency rate of the cluster by nearly 50%, and improve the energy transmission efficiency of the cluster by nearly 6%. (3) The optimization at the converter level has a great impact on the improvement of the cluster energy transmission efficiency. Compared with the scheme 3 without the optimization at the device level, the optimized scheme (scheme 1) can improve the cluster energy transmission efficiency by nearly 15.6%. (4) The simulation results of the four schemes show that the interconnection converter loss is the main source of the total cluster loss. Therefore, the key to improving the cluster energy transmission loss lies in reducing the interconnection converter loss. The optimization at the converter level can improve the conversion efficiency of the converter in the entire working domain; the optimization at the system level can make the converter operate at a higher efficiency point.

[0084] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be covered in the scope of the claims of the present application.

Claims

1. A method for long-term energy efficiency optimization of DC nanogrid clusters, characterized in that, This method is applied to a DC nanogrid cluster consisting of multiple nanogrids (NG) interconnected by distributed lines and interconnected DC-DC converters, wherein the interconnected DC-DC converters include two-port dual active bridge converters and three-port triple active bridge converters; the method specifically includes the following steps: S1. Establish an equivalent circuit model of the cluster that takes into account the losses of interconnect converters to accurately describe the power flow relationship of the cluster. S2. Based on the equivalent circuit model, construct a cluster global loss model to quantify the three core losses of lines, batteries, and converters. S3. Establish a cluster energy sharing model, classify NGs according to the net power characteristics of NGs in the grid, and clarify the energy interaction rules; S4. Based on the above three types of models, the Energy Optimization Management Strategy (EOMS) is adopted. Under the premise of ensuring safe operation, the energy flow is optimized through the particle swarm optimization algorithm to improve the self-sufficiency and energy efficiency of the cluster over a long time scale.

2. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 1, characterized in that, In step S1, the establishment of the cluster equivalent circuit model taking into account the interconnect converter losses and accurately describing the cluster power flow relationship specifically includes: the equivalent form of each component is highly matched with its function in the cluster—the distributed line uses resistance to represent transmission loss, the load uses constant current source to represent the stability of power demand, the battery and photovoltaic use variable current source to represent power adjustability, and the interconnect converter realizes voltage matching function and loss characterization respectively through "ideal transformer + variable current source", taking into account both functionality and practicality.

3. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 2, characterized in that, In step S2, the construction of a cluster global loss model and the quantification of the three core losses—line, battery, and converter—include: The cluster global loss model takes real-time power loss estimation as its core objective and establishing a functional relationship between cluster power transmission loss and power flow as its core task, providing model support for cluster energy efficiency optimization. Before constructing the model, the influencing factors of DC nanogrid cluster energy transmission efficiency are first identified: "Source", namely the energy conversion efficiency of photovoltaic array and photovoltaic utilization rate; "Transmission", namely distributed line loss and interconnect converter loss; "Storage", namely the energy conversion efficiency of energy storage unit. Among them, the energy conversion efficiency of photovoltaic array and energy storage unit is mainly determined by their own architecture and material selection, and the impact of cluster energy management strategy on them is negligible. Therefore, it focuses on quantifying three core losses: distributed line loss, battery internal resistance loss, and interconnect converter loss. All three types of losses are based on the cluster equivalent circuit model that takes into account interconnect converter loss, ensuring the consistency between loss calculation and power flow description.

4. The long-term energy efficiency optimization method for DC nanogrid clusters according to claim 3, characterized in that, In step S2, the three components of loss—distributed line loss, battery internal resistance loss, and interconnect converter loss—are expressed by the following formula: Distributed line loss: Battery internal resistance loss: Ignoring the imbalance of individual cells, the second-order RC equivalent circuit model of a battery pack consisting of n cells connected in series and m cells connected in parallel is given. The equivalent resistances Rpe,q, Ro,eq and Rs,eq of the battery pack, and the open-circuit voltage Voce,eq are respectively: The loss due to the internal resistance of the battery is expressed as: Considering battery lifespan and safety, the operating range of the battery SoC is limited to 0.2~0.

8. R o≈0.06Ω, battery port current i bat,i is the following formula: V OCV represents the battery open-circuit voltage. When the SoC is in the range of 0.2~0.8, V The following formula fits the variation of oce with SoC: According to KVL, NG Battery port voltage in i v The relationship between bat,i and the open-circuit voltage is expressed as follows: Solve simultaneously i bat, i, v bat, i, p bat, i, SoC and losses p loss, bat; The interconnect converter loss adopts the near-all-ZVS optimized modulation method, and the circuit parameters adopt the loss expression of the optimal circuit parameters as follows: Based on the above equation, the total loss of the DC nanogrid cluster system is as follows: 。 5. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 4, characterized in that, The energy sharing model can intuitively describe the energy interaction mechanism between clusters, between NGs within a cluster, and between functional units within an NG. The Energy Management Strategy (EOMS) integrates three objectives: Based on the equivalent circuit model of the cluster, the bus, cells and photovoltaics are ensured to operate in the safe operating domain, so as to guarantee the supply and demand balance and safe and stable operation of the cluster. Based on the energy sharing model of the cluster, under the premise of ensuring supply and demand balance, the energy self-sufficiency rate of the cluster is improved over a long time scale, and the self-sufficiency capacity of the cluster is strengthened. Based on the global loss model of the cluster, the average energy transfer efficiency of the cluster is improved over a long time scale.

6. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 5, characterized in that, In step S4, the cluster energy optimization management strategy is implemented through the following steps: S41. Determine the secure working domain of the cluster; S42. Calculate the self-sufficiency rate of the cluster while ensuring the safe working domain of the cluster; S43. Determine the optimal interaction power with the goal of minimizing cluster energy transmission loss. P NG,i .

7. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 6, characterized in that, Determined by certain constraints P NG,i , P bat,i and P pv,i The range of values ​​for the safe operating domain of the cluster includes: bus voltage fluctuations not exceeding limits to ensure power quality; battery SoC not exceeding limits to avoid battery overcharging and over-discharging; battery charging and discharging power not exceeding limits to ensure battery charging and discharging safety; and photovoltaic power generation not exceeding limits to avoid system supply and demand imbalance.

8. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 7, characterized in that, Power interaction among grids is determined based on the principle of energy sharing with high self-sufficiency. P NG,i Interaction power between the cluster P net The specific principles for determining the value of are as follows: (1) Energy interaction is prioritized between NGs within the cluster; if energy interaction between NGs within the cluster ensures that all NGs achieve supply and demand balance, then energy interaction will occur between clusters. (2) Regarding the energy interaction between NGs, the energy absorbed by NG- can only be used for load demand and cannot be used for battery charging. The first choice for NG+ energy output is the remaining photovoltaic energy, and the second choice is the available energy in the battery. (3) Regarding energy interaction between clusters, the absorbed energy can only be used for the load demand in the cluster and cannot be used for battery charging. The emitted energy can only be the remaining photovoltaic energy.

9. The method for long-term energy efficiency optimization of DC nanogrid clusters according to claim 7, characterized in that, Determine the optimal interaction power with the goal of minimizing cluster energy transmission loss. P NG,i The cluster transmission loss consists of three parts: distributed line loss, battery internal resistance loss, and interconnect converter loss. The mathematical formula for the corresponding optimization problem is as follows: 。 10. A long-term energy efficiency optimization system for a DC nanogrid cluster, characterized in that, The system employs the method described in any one of claims 1 to 9.

Images