Distributed DC micro-grid secondary control method based on preset time

Abstract

The invention belongs to the field of micro-grid control, and discloses a distributed direct current micro-grid secondary control method based on preset time, which comprises the following steps: acquiring a node operation original sequence in a periodic acquisition and time sequence storage mode, analyzing a state feature vector and constructing a node state set; analyzing to obtain a primary droop feature set and a primary operation deviation set; constructing a time constraint error evolution path and determining a preset time error sequence; based on the preset time error sequence and in combination with a neighborhood communication mechanism, generating a distributed secondary regulation quantity; and generating a secondary control instruction based on the distributed secondary regulation quantity in combination with sequence updating, saturation limitation and a soft start mechanism. Through a distributed sequential control means based on predetermined voltage recovery and frequency coordination, the frequency consistency of a plurality of micro-grid systems can be maintained, and significant frequency fluctuation and drift caused by load fluctuation or power supply input abrupt change can be effectively avoided.

Description

Technical Field

[0001] This invention belongs to the field of microgrid control, specifically relating to a secondary control method for a distributed DC microgrid based on a preset time. Background Technology

[0002] DC microgrids, by rapidly integrating renewable energy and distributed generation, can be transformed into an effective mode for improving energy efficiency and sustainability. Typically, the stable operation of a microgrid relies on main voltage regulation, secondary recovery, and a three-tiered optimization architecture. In this scenario, secondary control is responsible for restoring the bus voltage and system frequency to normal values ​​and ensuring the accuracy of power distribution among distributed generation.

[0003] To achieve the above objectives, a centralized secondary controller is a feasible solution; however, this solution suffers from drawbacks such as single point of failure and high communication requirements, leading to difficulties in rapid recovery and high maintenance costs. To address these issues, distributed secondary control technology has been extensively studied. By abandoning the traditional centralized controller and adopting the basic principle of proximity communication, this strategy significantly improves scalability while drastically reducing communication overhead. Typical breakthroughs in distributed strategies include: in distributed secondary schemes operating under time-varying electricity prices and internal price-rate mechanisms, the introduction of market signals and power management technologies enables adaptive operation of DC microgrids; in guiding communication networks, distributed secondary control technology based on preset time has been developed, ensuring that the microgrid can complete voltage and frequency regulation and restoration within a specified time.

[0004] Despite the significant progress made, the field of microgrid control still faces many major challenges: the primary issue is that the literature on distributed systems still lacks sufficient synchronous safeguards for voltage recovery, frequency regulation, and precise power distribution within a preset time frame; in addition, the robustness of controllers due to communication delays and model uncertainties makes it difficult to solve these problems within a unified framework; at the same time, most design schemes do not clearly define the transient error range, making safety certification and subsequent engineering deployment more complex and difficult to manage. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a secondary control method for distributed DC microgrids based on a preset time. It employs a unified frequency regulation control strategy and utilizes distributed timing control methods based on predetermined voltage recovery and frequency coordination to maintain frequency consistency across multiple microgrid systems while effectively avoiding significant frequency fluctuations and drift caused by load fluctuations or sudden changes in power input.

[0006] To achieve the above objectives, this application employs the following technical solution:

[0007] This application discloses a secondary control method for a distributed DC microgrid based on a preset time, comprising the following steps:

[0008] Step 1: Obtain the original sequence of node operation by periodically collecting and storing data in a time-series manner, parse out the state feature vector, and construct the node state set;

[0009] Step 2: Construct a set of drooping features and a set of running deviations based on the node state set built in Step 1;

[0010] Step 3: Based on the set of one-time operation deviations constructed in Step 2, construct the time-constrained error evolution path and determine the preset time error sequence;

[0011] Step 4: Based on the preset time error sequence determined in Step 3, and combined with the neighborhood communication mechanism, generate a distributed secondary adjustment quantity;

[0012] Step 5: Based on the distributed secondary adjustment quantity generated in Step 4, and combined with sequence update, saturation limit and soft start mechanism, generate secondary control instructions.

[0013] A further improvement in this application is that, in step 1, the original sequence of node operation is obtained, the state feature vector is parsed, and the node state set is constructed, which specifically includes the following steps:

[0014] Step 1.1: Based on the physical wiring structure of the distributed DC microgrid, and according to the connection points of each parallel converter and distributed generator set, determine each node in the microgrid topology:

[0015] i∈{1,2,…,N}

[0016] Each node corresponds to the operating interface of its connected distributed generator set or energy storage converter, and the line admittance Y ik for

[0017] Y ik =G ik +jB ik

[0018] Among them, G ik For the line conductance, B ik For line susceptance, if node i and node k are not connected, then Y ik =0, the set of neighboring nodes of node i is

[0019]

[0020] Each node connects to the local load.

[0021] S Li =P Li +Q Li

[0022] Among them, S Li P represents the complex power of the load connected to node i. Li Q represents the active power of the local load at node i. Li This represents the reactive power of the local load at node i;

[0023] Step 1.2, using T s The sampling period is used to periodically collect the node voltage amplitude V. i (t m Voltage phase angle δ i (t m and line susceptance B ik The original sequence of node execution is obtained by storing it in chronological order.

[0024]

[0025] Among them, t m =mT s m is the sampling number, V i Let δ be the voltage magnitude at node i. i Let i be the voltage phase angle at node i;

[0026] Step 1.3: Calculate the active power P at the nodes based on the power balance relationship. i Reactive power Q i :

[0027]

[0028] Among them, V k Let δ be the voltage magnitude at node k. k Let B be the voltage phase angle at node k. ik For line susceptance;

[0029] Step 1.4: At each sampling time t m Organize the node variables into a vector to form the state feature vector S. i (t m ):

[0030]

[0031] Among them, deg i To determine the node topology, assign all state feature vectors S from 1 to N according to the bus number. i (t m The node state set S(t) is obtained by splicing the states together. m ):

[0032] S(t m )=[S1(tm ),…,S N (t m )]

[0033] And based on the sampling period, a node state set sequence {S(t1),S(t2),…,S(t)} is formed. m )}.

[0034] A further improvement of this application is that step 2 specifically includes the following steps:

[0035] Step 2.1: In the primary control layer, node i uses droop control, and the simplified dynamics of node i are as follows:

[0036]

[0037] Where, ω d V is the reference angular velocity given by the primary control layer. d k is the reference voltage given by the primary control layer. Pi k is the active power droop coefficient. Qi k is the reactive power droop factor. Vi This is the equivalent gain of the voltage loop. For the desired active power, For the desired reactive power, To measure the low-pass filter value of active power, The low-pass filter value is used to measure reactive power;

[0038] The filter model is

[0039]

[0040] Where, τ Pi With τ Qi These are the first-order low-pass filter time constants for the active power measurement channel and the reactive power measurement channel at node i, respectively.

[0041] Step 2.2: Based on step 2.1, combine the node state set S(t) m V in ) i Node active power P i Reactive power Q i Extract (k) Pi ,k Qi ,k Vi ,τ Pi ,τ Qi ) and rated quantity This constitutes a set of drooping features D:

[0042]

[0043] in, This is the reference value for the rated voltage of the primary control layer at the i-th node. Let be the rated active power of the i-th node. Let be the rated reactive power of the i-th node;

[0044] Step 2.3, using the system voltage reference value V ref With angular velocity reference value ω ref Based on this, we define node i at sampling time t. m voltage deviation e v,i (t m ) and node i at sampling time t m angular velocity deviation e ω,i (t m The direction of the deviation change Δe is obtained from the time difference. ω,i (t m ) and the rate of change of deviation Δe v,i (t m ):

[0045] e v,i (t m ) = V i (t m )-V ref

[0046] e ω,i (t m )=ω i (t m )-ω ref

[0047] Δe v,i (t m ) = e v,i (t m )-e v,i (t m-1 )

[0048] Δe ω,i (t m ) = e ω,i (t m )-e ω,i (t m-1 )

[0049] Where, ω i (t m ) represents node i at sampling time t m The angular velocity value, e v,i (t m-1 ) represents node i at the previous sampling time t m-1 The voltage deviation value, e ω,i (t m-1) represents node i at the previous sampling time t m-1 Angular velocity deviation value;

[0050] Step 2.4: Form the set of primary operating deviations E(t) m ):

[0051] E(t m )=[e v,1 ,…,e v,N ,e ω,1 ,…,e ω,N ] T

[0052] Among them, e v,N Let e ​​be the voltage deviation value of the Nth node at the current sampling time. ω,N Let T be the angular velocity deviation of the Nth node at the current sampling time, and T be the transpose.

[0053] Step 2.5: Compare the first-order droop feature set D constructed in Step 2.2 with the first-order running deviation set E(t) constructed in Step 2.4. m It performs periodic updates to obtain continuous input data for the time evolution path.

[0054] A further improvement in this application is that, in step 3, constructing the time-constrained error evolution path and determining the preset time error sequence specifically includes the following steps:

[0055] Step 3.1: Pre-set the error to be within a preset time T. * Converging to accuracy k, T * Define a piecewise smooth non-decreasing function γ for k > 0 and k > 0. i (t) serves as the funnel boundary generating function, satisfying

[0056] γ i (t0)=0,

[0057] And t>t0+T * γ i (t)≡1 / k, Represents a smooth, non-decreasing function γ i (t) represents the time derivative at the initial time t0, i.e., the growth rate of the preset time boundary function at the initial time. Represents a smooth, non-decreasing function γ i (t) at the preset time endpoint t0+T * The time derivative of the boundary function is the rate of increase of the boundary function at the point of convergence.

[0058] Step 3.2: Analyze the error component in the single-operation deviation, i.e., the node voltage deviation e. v,i (t) and the deviation of the nodal angular velocity eω,i (t) Constraints are constructed, and the node voltage deviation e is taken respectively. v,i (t) and the deviation of the nodal angular velocity e ω,i (t)

[0059] The initial error value |e at the preset time control start time t0 v,i (t0)|、|e ω,i (t0)|, and combined with the corresponding upper bound of error, based on the piecewise smooth non-decreasing function γ i (t) Establish the error allowable boundary 1 / γ i (t),

[0060] get

[0061]

[0062] Among them, deviation e i (t) represents the node voltage deviation e v,i (t) or nodal angular velocity deviation e ω,i (t);

[0063] The evolution of the above boundary over time is the evolution path of the time-constrained error.

[0064] A further improvement in this application is that step 4 specifically includes the following steps:

[0065] Step 4.1: Introduce secondary inputs based on the primary control layer. Obtain extended dynamics:

[0066]

[0067] in, These are the voltage channel control quantity and angular velocity channel control quantity for the secondary adjustment of the i-th node, respectively. ω is the rate of change of angular velocity. d The equivalent angular velocity reference value is given by the primary control layer. The rate of change of voltage;

[0068] Step 4.2: Label the neighborhood node set N using the pre-established neighborhood identification list. i The data packets reported by neighboring nodes include local error values, state values, and timestamps. Time synchronization processing is performed on the data packets reported by neighboring nodes to unify the neighborhood information in the data packets to the reference time of the current control cycle, and a local consistency error is defined.

[0069]

[0070] in, and Z represents the pinning gain applied to the voltage channel and the pinning gain applied to the angular velocity channel at the i-th node, respectively, used to track and couple the node to the voltage and equivalent angular velocity reference values. v,i Let z be the voltage consistency error of the i-th node. ω,i The equivalent angular velocity consistency error of the i-th node;

[0071] Step 4.3: For any error z i Perform funnel-constrained scalar transformation:

[0072]

[0073] This transformation makes when ξ i When bounded, z i Including z v,i z ω,i , z i satisfy Corresponding preset time error constraints;

[0074] Step 4.4, each node at sampling time t m Based on local ξ i (t m ) and neighborhood ξ k (t m Calculate the distributed secondary regulation Δu i (t m ):

[0075]

[0076] Among them, B i β is the preset time convergence gain for the i-th node. i η is the preset time-convergence auxiliary gain for the i-th node. i (t m () represents the positive definite gain term, which is output in the voltage channel. Output in frequency channel And write it to the register area.

[0077] A further improvement in this application is that step 5 specifically includes the following steps:

[0078] Step 5.1: Adjust the distributed secondary adjustment amount Δu for this cycle. i (t m ) and the historical adjustment amount of the previous cycle Combine to form a continuous judgment sequence:

[0079]

[0080] Where ρ∈(0,1] are the update coefficients, Δu i (tm ) represents the current sampling period t of the i-th node. m Distributed secondary adjustment quantity;

[0081] Step 5.2: Based on the inverter's allowable drive range For continuous judgment sequences Truncation processing enables continuous judgment sequences By satisfying the allowable control output range, the saturated state of the i-th node at sampling time t can be obtained. m Secondary adjustment amount after amplitude truncation

[0082]

[0083] in, This represents the maximum allowed secondary adjustment value for the i-th node. This represents the minimum allowed secondary adjustment value for the output of the i-th node;

[0084] Step 5.3: Secondary adjustment amount after saturation obtained in Step 5.2. The amplitude is used to divide the adjustment process into several soft-start stages l = 1, ..., L, and a gradual ramp-up slope σ is preset for the l-th soft-start stage. l ∈(0,1), within each soft-start phase, the secondary adjustment is processed in segments and gradually increased according to the following formula to obtain the sampling time t of the i-th node. m The final secondary control command output:

[0085]

[0086] Among them, u i (t m-1 ) represents the secondary control command at the previous sampling time; l represents the soft start stage number;

[0087] Step 5.4: Send the final secondary control command u i (t m The voltage reference channel and equivalent angular velocity reference channel of the i-th node inverter are written respectively, corresponding to the correction of the reference voltage V given by the control layer. d and the reference angular velocity ω given by the primary control layer d Each node repeats steps 5.1-5.3 within each control cycle until the error of each node is within the preset time T. * Internal satisfaction And within the tolerance range, the consistency error z between the first droop feature set D and the equivalent angular velocity of the i-th node. ω,i The combined effect enables each node to output its nodal active power P in steady state. i According to their respective rated active power The proportions are allocated accordingly.

[0088] The beneficial effects of this application are:

[0089] This application integrates multiple core control objectives: the bus voltage can be restored within a limited time to ensure stable microgrid voltage quality; and a unified frequency regulation control strategy is adopted to maintain the frequency consistency of multiple microgrid systems and effectively avoid significant frequency fluctuations and drift caused by load fluctuations or sudden changes in power input.

[0090] This application's time-limited funnel constraint strictly controls the convergence process of voltage and frequency errors. Unlike traditional control methods that only ensure convergence without setting a convergence time, this method allows designers to pre-set the time according to actual operating requirements. By ensuring that voltage and frequency errors converge within a specified time, the system response speed and stability are guaranteed. This innovative achievement is of great significance to the safe and reliable operation of microgrids.

[0091] This application integrates soft-start and saturation control mechanisms, effectively balancing the performance and practicality of the control system. The soft-start mechanism can effectively prevent sudden changes in the control signal during startup from causing hard adjustments to the inverter and microgrid system. At the same time, the saturation control mechanism ensures that the control output is always within the inverter's drive range, avoiding equipment damage due to overdrive. Attached Figure Description

[0092] Figure 1 This is a schematic diagram of the control method flow of this application.

[0093] Figure 2 This is a schematic diagram of the voltage output curve for testing an islanded microgrid.

[0094] Figure 3 This is a schematic diagram of the frequency output curve for testing an islanded microgrid.

[0095] Figure 4 This is a schematic diagram of the voltage error curve after secondary control of an islanded microgrid.

[0096] Figure 5 This is a schematic diagram of the frequency error curve after secondary control of an islanded microgrid.

[0097] Figure 6 This is a schematic diagram of the actual power output curve for testing an islanded microgrid.

[0098] Figure 7 This is a logic diagram of the secondary control method in this application. Detailed Implementation

[0099] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0100] like Figure 1 and Figure 7 As shown, this application discloses a secondary control method for a distributed DC microgrid based on a preset time. The method includes the following steps:

[0101] Step 1: Obtain the original sequence of node operation by periodically collecting and storing data in a time-series manner, parse out the state feature vector, and construct the node state set;

[0102] In step 1, the original sequence of node execution is obtained, the state feature vector is parsed, and the node state set is constructed. This includes the following steps:

[0103] Step 1.1: Based on the physical wiring structure of the distributed DC microgrid, and according to the connection points of each parallel converter and distributed generator set, determine each node in the microgrid topology:

[0104] i∈{1,2,…,N}

[0105] Each node corresponds to the operating interface of its connected distributed generator set or energy storage converter, and the line admittance Y ik for

[0106] Y ik =G ik +jB ik

[0107] Among them, G ik For the line conductance, B ik The line susceptance is derived from the line parameter calibration or controller parameter table. If node i and node k are not connected, then Y... ik =0, the set of neighboring nodes of node i is

[0108]

[0109] Each node connects to the local load.

[0110] S Li =P Li +Q Li

[0111] Among them, S Li P represents the complex power of the load connected to node i. LiQ represents the active power of the local load at node i. Li This represents the reactive power of the local load at node i;

[0112] Step 1.2, using T s The sampling period is used to periodically collect the node voltage amplitude V. i (t m Voltage phase angle δ i (t m and line susceptance B ik The topological association data is obtained and stored in chronological order to obtain the original sequence of node operations.

[0113]

[0114] Among them, t m =mT s m is the sampling number, V i Let δ be the voltage magnitude at node i. i Let i be the voltage phase angle at node i;

[0115] Step 1.3: Calculate the active power P at the nodes based on the power balance relationship. i Reactive power Q i :

[0116]

[0117] Among them, V k Let δ be the voltage magnitude at node k. k Let B be the voltage phase angle at node k, derived from the real-time sampling in step 1.2. ik For line susceptance;

[0118] Step 1.4: At each sampling time t m Organize the node variables into a vector to form the state feature vector S. i (t m ):

[0119]

[0120] Among them, deg i The node topology degree, the number of connected branches, is derived from the topology identification list, and all state feature vectors S are numbered from 1 to N according to the bus number. i (t m The node state set S(t) is obtained by splicing the states together. m ):

[0121] S(t m )=[S1(t m ),…,S N (t m)]

[0122] And based on the sampling period, a node state set sequence {S(t1),S(t2),…,S(t)} is formed. m )}.

[0123] Step 2: Construct a droop feature set and a running deviation set based on the node state set built in Step 1. Step 2 specifically includes the following steps:

[0124] Step 2.1: In the primary control layer, node i adopts droop control. The simplified dynamics of node i (derived from the DG primary control model) are as follows:

[0125]

[0126] Where, ω d V is the reference angular velocity given by the primary control layer. d The reference voltage given to the primary control layer is derived from the system's rated settings; k Pi k is the active power droop coefficient. Qi k is the reactive power droop factor. Vi The voltage loop equivalent gain is derived from the inverter controller parameters. For the desired active power, The desired reactive power; To measure the low-pass filter value of active power, The low-pass filter value is used to measure reactive power;

[0127] The filter model is

[0128]

[0129] Where, τ Pi With τ Qi These are the first-order low-pass filter time constants for the active power measurement channel and the reactive power measurement channel at node i, respectively.

[0130] Step 2.2: Based on step 2.1, combine the node state set S(t) m V in ) i Node active power P i Reactive power Q i Extract (k) Pi ,k Qi ,k Vi ,τ Pi ,τ Qi ) and rated quantity This constitutes a set of drooping features D:

[0131]

[0132] in, This is the reference value for the rated voltage of the primary control layer at the i-th node. Let be the rated active power of the i-th node. Let be the rated reactive power of the i-th node;

[0133] Step 2.3, using the system voltage reference value V ref With angular velocity reference value ω ref Based on this, we define node i at sampling time t. m voltage deviation e v,i (t m ) and node i at sampling time t m angular velocity deviation e ω,i (t m The direction of the deviation change Δe is obtained from the time difference. ω,i (t m ) and the rate of change of deviation Δe v,i (t m ):

[0134] e v,i (t m ) = V i (t m )-V ref ,

[0135] e ω,i (t m )=ω i (t m )-ω ref

[0136] Δe v,i (t m ) = e v,i (t m )-e v,i (t m-1 ),

[0137] Δe ω,i (t m ) = e ω,i (t m )-e ω,i (t m-1 )

[0138] Where: ω i (t m ) represents node i at sampling time t m The angular velocity value, e v,i (t m-1 ) represents node i at the previous sampling time t m-1 The voltage deviation value, e ω,i (tm-1 ) represents node i at the previous sampling time t m-1 Angular velocity deviation value;

[0139] Step 2.4: Form the set of primary operating deviations E(t) m ):

[0140] E(t m )=[e v,1 ,…,e v,N ,e ω,1 ,…,e ω,N ] T

[0141] Among them, e v,N Let e ​​be the voltage deviation value of the Nth node at the current sampling time. ω,N Let T be the angular velocity deviation of the Nth node at the current sampling time, and T be the transpose, used to write the deviations arranged in order within the parentheses into a column vector form;

[0142] Step 2.5: Compare the first-order droop feature set D constructed in Step 2.2 with the first-order running deviation set E(t) constructed in Step 2.4. m It performs periodic updates to obtain continuous input data for the time evolution path.

[0143] Step 3: Based on the set of operational deviations constructed in Step 2, construct the time-constrained error evolution path and determine the preset time error sequence.

[0144] Step 3 involves constructing the time-constrained error evolution path and determining the preset time error sequence, specifically including the following steps:

[0145] Step 3.1: Pre-set the error to be within a preset time T. * Converging to accuracy k, T * >0, k>0 (derived from operational indicators / design requirements). Define a piecewise smooth non-decreasing function γ. i (t) serves as the funnel boundary generating function, satisfying

[0146] γ i (t0)=0,

[0147] And t>t0+T * γ i (t)≡1 / k, Represents a smooth, non-decreasing function γ i (t) represents the time derivative at the initial time t0, i.e., the growth rate of the preset time boundary function at the initial time. Represents a smooth, non-decreasing function γ i (t) at the preset time endpoint t0+T *The time derivative of the boundary function is the rate of increase of the boundary function at the point of convergence.

[0148] Step 3.2: Analyze the error component in the single-operation deviation, i.e., the node voltage deviation e. v,i (t) and the deviation of the nodal angular velocity e ω,i (t) Constraints are constructed, and the node voltage deviation e is taken respectively. v,i (t) and the deviation of the nodal angular velocity e ω,i (t)

[0149] The initial error value |e at the preset time control start time t0 v,i (t0)|、|e ω,i (t0)|, and combined with the corresponding acceptable upper bound of error, based on the piecewise smooth non-decreasing function γ i (t) Establish the error allowable boundary 1 / γ i (t),

[0150] get

[0151]

[0152] Among them, deviation e i (t) represents the node voltage deviation e v,i (t) or nodal angular velocity deviation e ω,i (t);

[0153] The evolution of the above boundary over time is the evolution path of the time-constrained error.

[0154] Step 4: Based on the preset time error sequence determined in Step 3, and in conjunction with the neighborhood communication mechanism, generate a distributed secondary adjustment quantity. Step 4 specifically includes the following steps:

[0155] Step 4.1: Introduce secondary inputs based on the primary control layer. Obtain extended dynamics:

[0156]

[0157] in, These are the voltage channel control quantity and angular velocity channel control quantity for the secondary adjustment of the i-th node, respectively. The equivalent angular velocity ω of the i-th node i The derivative with respect to time, i.e., the rate of change of angular velocity; ω d The equivalent angular velocity reference value given by the primary control layer; Let V be the voltage amplitude of the i-th node. i The derivative with respect to time, i.e., the rate of change of voltage;

[0158] Step 4.2: Label the neighborhood node set N using the pre-established neighborhood identification list.i (Sourced from communication topology configuration or online discovery), the data packets reported by neighboring nodes include neighborhood local error, neighborhood state, and timestamps. Time synchronization processing is performed on the data packets reported by neighboring nodes to unify the neighborhood information in the data packets to the reference time of the current control cycle, and a local consistency error is defined.

[0159]

[0160] in, and These are the pinning gains acting on the voltage channel and the pinning gains acting on the angular velocity channel in the i-th node, respectively, used to track and couple the node to the voltage and equivalent angular velocity reference values; z v,i Let z be the voltage consistency error of the i-th node, used to characterize the deviation of the node's voltage relative to its neighborhood and the reference value; ω,i The equivalent angular velocity consistency error of the i-th node is used to characterize the deviation of the equivalent angular velocity of the node relative to its neighborhood and reference value.

[0161] Step 4.3: For any error z i Perform funnel-constrained scalar transformation:

[0162]

[0163] This transformation makes when ξ i When bounded, z i Including z v,i z ω,i , z i satisfy Corresponding preset time error constraints;

[0164] Step 4.4, each node at sampling time t m Based on local ξ i (t m ) and neighborhood ξ k (t m Calculate the distributed secondary regulation Δu i (t m ):

[0165]

[0166] Among them, B i β is the preset time convergence gain for the i-th node. i η is the preset time-convergence auxiliary gain for the i-th node. i (t m () represents the positive definite gain term, which is output in the voltage channel. Output in frequency channel And write it to the register area.

[0167] Step 5: Based on the distributed secondary adjustment quantity generated in Step 4, and combined with sequence update, saturation limit, and soft start mechanism, generate secondary control instructions. Step 5 specifically includes the following steps:

[0168] Step 5.1: Adjust the distributed secondary adjustment amount Δu for this cycle. i (t m ) and the historical adjustment amount of the previous cycle Combine to form a continuous judgment sequence:

[0169]

[0170] Where ρ∈(0,1] are the update coefficients, Δu i (t m ) represents the current sampling period t of the i-th node. m Distributed secondary adjustment quantity;

[0171] Step 5.2: Based on the inverter's allowable drive range Derived from equipment rated / safety constraints, for continuous judgment sequences Truncation processing enables continuous judgment sequences By satisfying the allowable control output range, the saturated state of the i-th node at sampling time t can be obtained. m Secondary adjustment amount after amplitude truncation

[0172]

[0173] in, This represents the maximum allowed secondary adjustment value for the i-th node. This represents the minimum allowed secondary adjustment value for the output of the i-th node;

[0174] Step 5.3: Secondary adjustment amount after saturation obtained in Step 5.2. The amplitude is used to divide the adjustment process into several soft-start stages l = 1, ..., L, and a gradual ramp-up slope σ is preset for the l-th soft-start stage. l ∈(0,1), during each soft-start phase, the secondary adjustment is processed in segments and gradually increased according to the following formula to obtain the sampling time t of the i-th node. m The final secondary control command output:

[0175]

[0176] Among them, u i (t m-1 ) represents the secondary control command at the previous sampling time; l represents the soft start stage number;

[0177] Through the above-described soft-start segmented gradual escalation process, the secondary control command is changed from u i (t m-1 Smooth transition to Avoid sudden changes in control parameters that could impact the equipment;

[0178] Step 5.4: Send the final secondary control command u i (t m The voltage reference channel and equivalent angular velocity reference channel of the i-th node inverter are written respectively, corresponding to the correction of the reference voltage V given by the control layer. d and the reference angular velocity ω given by the primary control layer d Each node repeats steps 5.1-5.3 within each control cycle until the error of each node is within the preset time T. * Internal satisfaction And within the tolerance range, the consistency error z between the first droop feature set D and the equivalent angular velocity of the i-th node. ω,i The combined effect enables each node to output its nodal active power P in steady state. i According to their respective rated active power The proportion is allocated. Among them, equivalent angular velocity consistency refers to the local consistency error z in step 4. ω,i Converging to zero, making the equivalent angular velocity ω at each node zero. i They converge to the same reference value, thereby achieving the distribution of active power according to the rated capacity ratio under the action of the primary control layer.

[0179] The secondary control commands are updated in real time in a discrete time series manner, which not only satisfies the preset time convergence, but also ensures that the active power is distributed according to the capacity ratio (which is jointly guaranteed by the primary droop relationship and frequency consistency).

[0180] like Figure 2 and Figure 3 As shown, in the first stage, due to the droop control, the voltage and frequency outputs are constrained within a relatively stable range: voltage from 310.2V to 310.4V, and frequency from 50.005Hz to 50.035Hz. It is noticeable that there are still slight deviations between the actual output value and the expected value (V = 310V and ω = 50Hz), requiring secondary control for compensation. At t = 5 seconds (secondary side activated), the secondary controller is activated, and it can be observed that the actual output and target value continuously decrease, eventually approaching a more stable value: voltage from 310V to 310.05V, and frequency from 49.990Hz to 49.995Hz. Simultaneously, it is noted that both voltage and frequency converge within the specified time (voltage convergence time is 1.5 seconds, frequency convergence time is 0.5 seconds).

[0181] like Figure 4 and Figure 5 As shown, this is to visually demonstrate whether the error is within the set upper and lower limits. Figure 4 and Figure 5 The upper and lower limits are marked, and only the errors of the first and second stages are displayed. It can be seen that before the secondary controller starts, all errors exceed the set range. However, after starting, even under continuous load fluctuations, the errors can still accurately converge to the target range and remain stable within the specified time. In terms of active power output, the preset power allocation ratio in the model is: 0.00020:0.00022:0.00024:0.00026≈1:1.1:1.2:1.3.

[0182] like Figure 6 As shown, in actual operation, the power distribution during the startup phase will fluctuate significantly, but it will quickly stabilize, and the fluctuation time is almost negligible; the final actual distribution ratio is 1:8:11:12:15; under continuous load fluctuation conditions, the distribution ratio of distributed power sources is basically close to the target value.

[0183] As can also be seen from the attached figures, all data conversion processes are relatively smooth; this is thanks to the soft-start mechanism adopted in this application, which ensures a smooth transition of the controller and avoids sudden changes or drastic fluctuations, thus making it easier for us to observe the results and adjust the parameters.

[0184] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0185] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0186] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0187] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0188] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0189] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A secondary control method for a distributed DC microgrid based on a preset time, characterized in that: The secondary control method for the distributed DC microgrid includes the following steps: Step 1: Obtain the original sequence of node operation by periodically collecting and storing data in a time-series manner, parse out the state feature vector, and construct the node state set; Step 2: Construct a set of drooping features and a set of running deviations based on the node state set built in Step 1; Step 3: Based on the set of one-time operation deviations constructed in Step 2, construct the time-constrained error evolution path and determine the preset time error sequence; Step 4: Based on the preset time error sequence determined in Step 3, and combined with the neighborhood communication mechanism, generate a distributed secondary adjustment quantity; Step 5: Based on the distributed secondary adjustment quantity generated in Step 4, and combined with sequence update, saturation limit and soft start mechanism, generate secondary control instructions.

2. The secondary control method for distributed DC microgrids based on a preset time according to claim 1, characterized in that: In step 1, the original sequence of node execution is obtained, the state feature vector is parsed, and the node state set is constructed. This includes the following steps: Step 1.1: Based on the physical wiring structure of the distributed DC microgrid, and according to the connection points of each parallel converter and distributed generator set, determine each node in the microgrid topology: i∈{1,2,…,N} Each node corresponds to the operating interface of its connected distributed generator set or energy storage converter, and the line admittance Y ik for Y ik =G ik +jB ik Among them, G ik For the line conductance, B ik For line susceptance, if node i and node k are not connected, then Y ik =0, the set of neighboring nodes of node i is Each node connects to the local load. S Li =P Li +Q Li Among them, S Li P represents the complex power of the load connected to node i. Li Q represents the active power of the local load at node i. Li This represents the reactive power of the local load at node i; Step 1.2, using T s The sampling period is used to periodically collect the node voltage amplitude V. i (t m Voltage phase angle δ i (t m and line susceptance B ik The original sequence of node execution is obtained by storing it in chronological order. Among them, t m =mT s m is the sampling number, V i Let δ be the voltage magnitude at node i. i Let i be the voltage phase angle at node i; Step 1.3: Calculate the active power P at the nodes based on the power balance relationship. i Reactive power Q i : Among them, V k Let δ be the voltage magnitude at node k. k Let B be the voltage phase angle at node k. ik For line susceptance; Step 1.4, at each sampling time t m Organize the node variables into a vector to form the state feature vector S. i (t m ): Among them, deg i To determine the node topology, assign all state feature vectors S from 1 to N according to the bus number. i (t m The node state set S(t) is obtained by splicing the states together. m ): S(t m )=[S1(t m ),…,S N (t m )] And based on the sampling period, a node state set sequence {S(t1),S(t2),…,S(t)} is formed. m )}.

3. The secondary control method for distributed DC microgrids based on a preset time according to claim 2, characterized in that: Step 2 specifically includes the following steps: Step 2.1: In the primary control layer, node i uses droop control, and the simplified dynamics of node i are as follows: Where, ω d V is the reference angular velocity given by the primary control layer. d k is the reference voltage given by the primary control layer. Pi k is the active power droop coefficient. Qi k is the reactive power droop factor. Vi This is the equivalent gain of the voltage loop. For the desired active power, For the desired reactive power, To measure the low-pass filter value of active power, The low-pass filter value is used to measure reactive power; The filter model is Where, τ Pi With τ Qi These are the first-order low-pass filter time constants for the active power measurement channel and the reactive power measurement channel at node i, respectively. Step 2.2: Based on step 2.1, combine the node state set S(t) m V in ) i Node active power P i Reactive power Q i Extract (k) Pi ,k Qi ,k Vi ,τ Pi ,τ Qi ) and rated quantity This constitutes a set of drooping features D: in, This is the reference value for the rated voltage of the primary control layer at the i-th node. Let be the rated active power of the i-th node. Let be the rated reactive power of the i-th node; Step 2.3, using the system voltage reference value V ref With angular velocity reference value ω ref Based on this, we define node i at sampling time t. m voltage deviation e v,i (t m ) and node i at sampling time t m angular velocity deviation e ω,i (t m The direction of the deviation change Δe is obtained from the time difference. ω,i (t m ) and the rate of change of deviation Δe v,i (t m ): e v,i (t m )=V i (t m )-V ref e ω,i (t m )=ω i (t m )-oh ref Δe v,i (t m )=e v,i (t m )-e v,i (t m-1 ) Δe ω,i (t m )=e ω,i (t m )-e ω,i (t m-1 ) Where, ω i (t m ) represents node i at sampling time t m The angular velocity value, e v,i (t m-1 ) represents node i at the previous sampling time t m-1 The voltage deviation value, e ω,i (t m-1 ) represents node i at the previous sampling time t m-1 Angular velocity deviation value; Step 2.4: Form the set of initial operating deviations E(t) m ): E(t m )=[and v,1 ,…,And v,N ,And ω,1 ,…,And ω,N ] T Among them, e v,N Let e ​​be the voltage deviation value of the Nth node at the current sampling time. ω,N Let T be the angular velocity deviation of the Nth node at the current sampling time, and T be the transpose. Step 2.5: Compare the first-order droop feature set D constructed in Step 2.2 with the first-order running deviation set E(t) constructed in Step 2.

4. m It performs periodic updates to obtain continuous input data for the time evolution path.

4. The secondary control method for distributed DC microgrids based on a preset time according to claim 3, characterized in that: Step 3 involves constructing the time-constrained error evolution path and determining the preset time error sequence, specifically including the following steps: Step 3.1: Pre-set the error to be within a preset time T. * Converging to accuracy k, T * Define a piecewise smooth non-decreasing function γ for k > 0 and k > 0. i (t) serves as the funnel boundary generating function, satisfying γ i (t0)=0, And t>t0+T * γ i (t)≡1 / k, Represents a smooth, non-decreasing function γ i (t) represents the time derivative at the initial time t0, i.e., the growth rate of the preset time boundary function at the initial time. Represents a smooth, non-decreasing function γ i (t) at the preset time endpoint t0+T * The time derivative of the boundary function is the rate of increase of the boundary function at the point of convergence. Step 3.2: Analyze the error component in the single-operation deviation, i.e., the node voltage deviation e. v,i (t) and the deviation of the nodal angular velocity e ω,i (t) Constraints are constructed, and the node voltage deviation e is taken respectively. v,i (t) and the deviation of the nodal angular velocity e ω,i (t) The initial error value |e at the preset time control start time t0 v,i (t0)|、|e ω,i (t0)|, and combined with the corresponding upper bound of error, based on the piecewise smooth non-decreasing function γ i (t) Establish the error allowable boundary 1 / γ i (t), get Among them, deviation e i (t) represents the node voltage deviation e v,i (t) or nodal angular velocity deviation e ω,i (t); The evolution of the above boundary over time is the evolution path of the time-constrained error.

5. The secondary control method for a distributed DC microgrid based on a preset time according to claim 1, characterized in that: Step 4 specifically includes the following steps: Step 4.1: Introduce secondary inputs based on the primary control layer. Obtain extended dynamics: in, These are the voltage channel control quantity and angular velocity channel control quantity for the secondary adjustment of the i-th node, respectively. ω is the rate of change of angular velocity. d The equivalent angular velocity reference value is given by the primary control layer. The rate of change of voltage; Step 4.2: Label the neighborhood node set N using the pre-established neighborhood identification list. i The data packets reported by neighboring nodes include local error values, state values, and timestamps. Time synchronization processing is performed on the data packets reported by neighboring nodes to unify the neighborhood information in the data packets to the reference time of the current control cycle, and a local consistency error is defined. in, and Z represents the pinning gain applied to the voltage channel and the pinning gain applied to the angular velocity channel at the i-th node, respectively, used to track and couple the node to the voltage and equivalent angular velocity reference values. v,i Let z be the voltage consistency error of the i-th node. ω,i The equivalent angular velocity consistency error of the i-th node; Step 4.3: For any error z i Perform funnel-constrained scalar transformation: This transformation makes when ξ i When bounded, z i Including z v,i z ω,i , z i satisfy Corresponding preset time error constraints; Step 4.4, each node at sampling time t m Based on local ξ i (t m ) and neighborhood ξ k (t m Calculate the distributed secondary regulation Δu i (t m ): Among them, B i β is the preset time convergence gain for the i-th node. i η is the preset time-convergence auxiliary gain for the i-th node. i (t m () represents the positive definite gain term, which is output in the voltage channel. Output in frequency channel And write it to the register area.

6. The secondary control method for a distributed DC microgrid based on a preset time, as described in claim 5, is characterized in that: Step 5 specifically includes the following steps: Step 5.1: Adjust the distributed secondary adjustment amount Δu for this cycle. i (t m ) and the historical adjustment amount of the previous cycle Combine to form a continuous judgment sequence: Where ρ∈(0,1] are the update coefficients, Δu i (t m ) represents the current sampling period t of the i-th node. m Distributed secondary adjustment quantity; Step 5.2: Based on the inverter's allowable drive range For continuous judgment sequences Truncation processing enables continuous judgment sequences By satisfying the allowable control output range, the saturated state of the i-th node at sampling time t can be obtained. m Secondary adjustment amount after amplitude truncation in, This represents the maximum allowed secondary adjustment value for the i-th node. This represents the minimum allowed secondary adjustment value for the output of the i-th node; Step 5.3: Secondary adjustment amount after saturation obtained in Step 5.

2. The amplitude is used to divide the adjustment process into several soft-start stages l = 1, ..., L, and a gradual ramp-up slope σ is preset for the l-th soft-start stage. l ∈(0,1), within each soft-start phase, the secondary adjustment is processed in segments and gradually increased according to the following formula to obtain the sampling time t of the i-th node. m The final secondary control command output: Among them, u i (t m-1 ) represents the secondary control command at the previous sampling time; l represents the soft start stage number.

Images