Distributed photovoltaic hierarchical coordination control method based on micro-grid architecture

Abstract

The invention discloses a distributed photovoltaic hierarchical coordination control method based on a micro-grid architecture, and relates to the technical field of distributed photovoltaic control, a three-layer architecture of an equipment layer, a micro-grid layer and an area layer is taken as a core, and the equipment layer realizes MPPT control by an improved perturbation and observation method and adjusts reactive power according to a target power factor; the micro-grid layer uses model prediction control to optimize power distribution, and controls energy storage charging and discharging by means of a state machine strategy; the regional layer overall plans coordination of multiple microgrids through a multi-objective optimization function and a Lagrange dual decomposition algorithm, improves decision precision in combination with combined load prediction, realizes inter-layer information synchronization by relying on a distributed consistency algorithm, constructs a multi-level fault detection and hierarchical self-healing mechanism to guarantee operation, maintains voltage stability through hierarchical voltage control, and achieves multi-level fault detection and hierarchical self-healing. Optimal control parameters are output through economic and environment-friendly optimization, a closed loop is formed through negative feedback adjustment, robust model predictive control is matched to deal with uncertainty, and reliable operation of the system is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of distributed photovoltaic control, in particular to a distributed photovoltaic hierarchical coordination control method based on a microgrid architecture. BACKGROUND

[0002] With the rapid development and large-scale application of distributed photovoltaic power generation technology, the traditional centralized power grid control mode is facing severe challenges. The main defects in the prior art are as follows: the traditional centralized control architecture is difficult to cope with the randomness and intermittency characteristics of distributed photovoltaic power generation. Distributed photovoltaic power generation is significantly affected by weather conditions, and the power output is highly volatile. The traditional scheduling system has a slow response speed, and it is difficult to achieve real-time tracking and rapid adjustment. The existing distributed control method lacks an effective hierarchical coordination mechanism. Most researches focus on a single control layer, such as only focusing on inverter-level control or microgrid-level control, and lack global coordination optimization from the device layer to the regional layer. And the existing control method takes power balance as the core target, and lacks overall consideration of economy and environmental benefits. Some schemes only optimize economy by reducing power purchase cost, but ignore the network loss cost of microgrid power exchange and the maintenance cost of energy storage system; another research focuses on environmental indicators, but does not establish a quantitative correlation between carbon emission cost and power scheduling, resulting in phenomena such as "high-carbon power purchase" and "redundant energy storage charging and discharging" during system operation.

[0003] In addition, the lack of fault handling capability is also a key bottleneck restricting the development of distributed photovoltaic systems. Traditional fault detection relies on a single indicator, which is difficult to identify hidden faults such as device aging and communication interruption; and lacks a hierarchical self-healing mechanism, so after a fault occurs, the system is often shut down for maintenance, resulting in a long power outage time of several minutes, which seriously affects user power supply reliability.

[0004] In summary, the shortcomings of the existing technology in control architecture coordination, operation stability, economic and environmental overall consideration, and fault self-healing capability have become the core obstacles to the large-scale and high-quality development of distributed photovoltaic systems, and a new type of control method with multi-level coordination, strong robustness, and full-scenario adaptation is urgently needed. SUMMARY

[0005] The purpose of the present application is to provide a distributed photovoltaic hierarchical coordination control method based on a microgrid architecture, which realizes stable, self-healing, economic and environmentally friendly operation of the distributed photovoltaic system by using device layer, microgrid layer and regional layer as the basis, through data acquisition preprocessing, fault monitoring and diagnosis, voltage stability control, economic and environmental optimization and negative feedback regulation, combined with MPPT control, model predictive control, Lagrange dual decomposition and other strategies.

[0006] To achieve the above purpose, the present application provides the following technical solutions:

[0007] A distributed photovoltaic hierarchical coordination control method based on microgrid architecture, characterized in that it comprises the following steps:

[0008] S1: Construct a three-layer control architecture of device layer, microgrid layer and regional layer, the device layer is responsible for local control of a single distributed photovoltaic device, the microgrid layer coordinates multiple distributed power sources and loads in the microgrid, and the regional layer coordinates the coordinated operation of multiple microgrid layers;

[0009] S2: Data acquisition and preprocessing, real-time acquisition of operation data at each level, filtering processing of the acquired data, elimination of abnormal values, and data import into the three-layer control architecture for execution;

[0010] S3: Fault monitoring and diagnosis, through a multi-level fault detection mechanism and a self-healing control strategy, continuously monitor the three-layer control architecture of device layer, microgrid layer and regional layer, and when a fault is detected, start the hierarchical progressive self-healing control decision;

[0011] S4: Voltage stability judgment, for unstable voltage, adopt a hierarchical voltage control strategy, each layer coordinates to maintain voltage stability, the device layer voltage regulation adopts droop control, the microgrid layer voltage regulation adopts secondary control to eliminate the steady-state error of droop control, and the regional layer realizes voltage coordination through optimizing power exchange of each microgrid, and the voltage control strategy is selected through the fault diagnosis result of S3;

[0012] S5: Economic and environmental optimization, for stable voltage and through the target function considering economy and environmental benefit after S4 voltage control strategy, determine the target function, obtain the optimal control parameters of each level through solving the multi-objective optimization function, including MPPT control parameters of the device layer, power distribution instructions of the microgrid layer, and microgrid power exchange instructions of the regional layer, and obtain the final control instruction;

[0013] S6: Negative feedback regulation, perform effect evaluation on the final control instruction, feed back the deviation data to each layer control architecture, dynamically adjust the control parameters, form a closed-loop optimization control, and return the evaluation data to the device layer, microgrid layer and regional layer three-layer control architecture for optimization architecture.

[0014] Step S1: The device layer control system is responsible for local control of a single distributed photovoltaic device, and the state variables of the device layer controller are defined as:

[0015] S i =[V i ,I i ,T i ,H i ,θ i ]

[0016] In the formula, V i , I iThese represent voltage and current, respectively; T represents temperature; and H represents temperature. i For humidity, θ i These are equipment status parameters; status variables are used to monitor the equipment's operating status in real time, providing a data foundation for fault detection and MPPT control.

[0017] The power control equation for the device layer is:

[0018] P i (t)=η i ·A i ·G(t)·[1-β i (T i (t)-T ref )]

[0019] In the formula, P i (t) represents the output power of the i-th photovoltaic device, η i For conversion efficiency, A i Let G(t) be the area of ​​the photovoltaic panel, G(t) be the solar radiation intensity, and β be the solar radiation intensity. i T is the temperature coefficient. ref For reference temperature;

[0020] The device layer employs an improved perturbation-observation method to achieve maximum power point tracking (MPPT) control. The improved MPPT algorithm introduces an adaptive step size adjustment mechanism, the formula of which is as follows:

[0021]

[0022] In the formula, ΔV(k+1) is the voltage perturbation amount (V) at step k+1, that is, the voltage change to be applied in the next step; ΔV(k) is the voltage perturbation amount at step k, that is, the voltage change in the current step; k is the discrete time step index, representing the number of iterations of the MPPT algorithm; γ is the adjustment factor, 0<γ<1, used to control the size of the perturbation step; dP / dV is the derivative of power with respect to voltage at the current moment (W / V), representing the slope of the power-voltage curve; dP / dV(k-1) is the derivative of power with respect to voltage at the previous moment (W / V); P is the output power of the photovoltaic array (W); V is the terminal voltage of the photovoltaic array (V).

[0023] Condition 1:

[0024] Physical meaning: The current power change direction is the same as the previous step.

[0025] Control strategy: Continue to perturb in the same direction, maintaining the original perturbation direction.

[0026] Mathematical expression: ΔV(k+1)=γ·ΔV(k)

[0027] Condition 2:

[0028] Physical meaning: the current power change direction is opposite to the previous step, which may have crossed the maximum power point

[0029] Control strategy: reverse disturbance, find the maximum power point

[0030] Mathematical expression: ΔV(k+1)=-γ·ΔV(k)

[0031] The voltage disturbance ΔV(k) is used to adjust the photovoltaic array operating point to achieve maximum power point tracking, and the disturbance is transmitted to the voltage control link to participate in voltage stability control.

[0032] Step S1 device layer also includes reactive power control, which automatically adjusts the power factor according to the grid voltage:

[0033] Q i (t)=P i (t)·tan(cos -1 (PF target ))

[0034] In the formula, Q i (t) is the reactive power, P i (t) is the output power of the i-th photovoltaic device, PF target is the target power factor, cos -1 (PF target ) is the inverse cosine function, which represents the angle corresponding to the target power factor, and its unit is radian; tan(cos -1 (PF target )) is the tangent function, and the input parameter is the above angle; the tangent value is equal to the ratio of reactive power to active power, that is, tanφ=Q / P. Here, the required reactive power ratio is directly deduced through the target power factor.

[0035] Step S1 microgrid layer controls and coordinates multiple distributed power sources and loads in the microgrid, so as to realize power balance and voltage stability in the microgrid, and the power balance constraint of the microgrid layer is defined as:

[0036]

[0037] In the formula, P pv,i (t) is the photovoltaic power generation, P es,j (t) is the energy storage system power, P grid (t) is the exchange power with the main grid, and P l,k (t) is the load power.

[0038] The microgrid layer uses the model predictive control (MPC) method to realize multi-time scale coordination optimization, and the prediction model is established as follows:

[0039] x(k+1) = Ax(k) + Bu(k) + Ed(k)

[0040] where x(k) is the state vector, u(k) is the control vector, d(k) is the disturbance vector, A, B, E are system matrices;

[0041] The optimization objective function of the microgrid layer is set as:

[0042]

[0043] where N p is the prediction horizon length, x(k) is the state vector, u(k) is the control vector, x T (k) is the transpose of x(k), u T (k) is the transpose of u(k), Q, R, S are weight matrices, the objective function is in the form of quadratic cost, the first term x T (k)Qx(k) represents the penalty of state deviation, the second term u T (k)Ru(k) represents the penalty of control action, and the third term x T (N p )Sx(N p ) is the terminal cost, which ensures the stability of the system at the end of the prediction horizon. The state prediction and control optimization inside a single microgrid are focused on.

[0044] Step S1 The state machine switching strategy is adopted to coordinate the charge and discharge control of the energy storage system in the microgrid layer:

[0045]

[0046] where SOC is the state of charge, P surplus is the remaining power, P deficit is the power shortage, and P smooth is the smooth power.

[0047] First case: when the state of charge of the energy storage is lower than the minimum value and there is remaining power, charge with maximum power

[0048] Second case: when the state of charge of the energy storage is higher than the maximum value and there is power shortage, discharge with maximum power

[0049] Third case: smooth power adjustment under other conditions

[0050] This is a state machine control strategy, which ensures that the energy storage system operates within a safe range and balances the power supply and demand in the microgrid.

[0051] Step S1 The regional layer coordinates the operation of multiple microgrids to achieve overall optimization in the region. The multi-objective optimization function of the regional layer is:

[0052] minJ = w1J cost +w2J loss +w3J emission

[0053] wherein J cost is the operation cost, J loss is the network loss, and J emission is the carbon emission, and w1, w2, and w3 are weight coefficients.

[0054] The coordination between multiple microgrids is focused on, and global objectives such as economy, network loss, and carbon emission are considered.

[0055] The regional layer uses a distributed optimization algorithm to solve the multi-microgrid coordination problem. The distributed algorithm framework based on Lagrange dual decomposition is:

[0056]

[0057] wherein f i (x i ) is the objective function of the ith microgrid, g(x) is the coupling constraint, and λ is the Lagrange multiplier.

[0058] Algorithm principle:

[0059] 1) Problem decomposition: The multi-microgrid coordination problem is decomposed into multiple sub-problems, and each microgrid optimizes its own objective function independently.

[0060] 2) Coupling constraint processing: The coupling constraints (such as power balance constraints) between microgrids are processed through the Lagrange multiplier.

[0061] 3) Iterative solution: Each microgrid solves its own sub-problem in parallel, and the coordination is achieved through the update of the Lagrange multiplier.

[0062] Coordination mechanism:

[0063] 1) Each microgrid perceives the "shadow price" of the overall system constraint through the Lagrange multiplier.

[0064] 2) While optimizing its own objective, it automatically considers the impact on the overall system.

[0065] 3) Ultimately, a globally optimal coordination operation state is achieved.

[0066] Dual decomposition iterative process:

[0067] x i k+1 = argminx i [fi (x i )+(λ k ) T g i (x i )]

[0068] λ k+1 =λ k +ρg(x k+1 )

[0069] where ρ is the step size parameter; x i k+1 is the decision variable of the i-th microgrid in the k+1th iteration; f i (x i ) is the local objective function of the i-th microgrid; λ k is the Lagrange multiplier vector of the kth iteration, representing the shadow price of the constraints; g i (x i ) is the constraint function related to the i-th microgrid; ρ is the step size parameter, controlling the updating speed of the Lagrange multiplier; λ k+1 is the Lagrange multiplier vector of the k+1th iteration; g(x k+1 ) is the value of the global constraint function at the k+1th iteration solution.

[0070] The regional layer also includes load forecasting, which uses a combination forecasting method to improve forecasting accuracy:

[0071]

[0072] where P is the predicted load, h is the prediction time step, and α1, α2, and α3 are combination weights. The role of load forecasting includes prediction basis: providing future load information for the multi-objective optimization of the regional layer; optimization input: the prediction result is an important input parameter of the MPC prediction model; decision support: helping the system to make power distribution and energy storage scheduling decisions in advance.

[0073] Load forecasting is an auxiliary technology that supports the main control process, and its prediction result is input into the S5 economic and environmental optimization step to improve the accuracy and economy of the system predictive control, serving as a supporting technology for regional layer decision-making.

[0074] Step S1 specifically uses a distributed consistency algorithm to achieve information synchronization and coordinated control between layers of control. The basic form of the consistency algorithm is:

[0075]

[0076]

[0077] where x is the state change rate of the i-th node, and N​i N(i) is the neighbor set of node i, a ij a(i, j) is the adjacency matrix element, representing the connection strength between node i and j, a i x(i) is the current state variable of node i, including voltage, frequency, power, a j x(j) is the state variable of neighbor node j, where the node includes device layer nodes, individual distributed photovoltaic inverters, energy storage converters, etc. device controllers; microgrid layer nodes, central controllers of individual microgrids; regional layer nodes, regional dispatch centers or master controllers;

[0078] Through the consensus algorithm, the state variables x i of each control node gradually converge to a consistent value, achieving information synchronization between the device layer, microgrid layer, and regional layer, where x i represents the key state parameters of each layer. The state change rate of each node i depends on the state difference with neighbor node j, and through continuous information exchange, the state of all nodes tends to be consistent.

[0079] To improve the convergence speed, a weighted consensus algorithm is introduced:

[0080]

[0081] where w ij is the communication weight coefficient between node i and j, 0 ≤ w ij ≤ 1; u i is the external input signal of node i, including upper-level control instructions, reference values; x(i) is the state change rate of node i, representing the speed of state change over time; i is the index number of the current node, including device, microgrid, or regional controller number; j is the index number of the neighbor node; N i N(i) is the neighbor set of node i, specifically other nodes that directly communicate with node i; x i x(i) is the current state variable of node i, including voltage, frequency, power; x j x(j) is the state variable of neighbor node j.

[0082] The inter-layer information transmission adopts an event-triggered mechanism to reduce communication burden, and the event-triggering condition is designed as:

[0083] |e i (t)| ≥ α|x i (t)| + β

[0084] where e i (t) is the measurement error of the i-th node at time t; x i (t) is the actual state variable of the i-th node at time t, including voltage, power; α, β are triggering threshold parameters.

[0085] Unlike traditional periodic communication, event-triggered mechanism only communicates and updates control when "needed".

[0086] Trigger condition design:

[0087] Event-triggered mechanism sends data only when the ratio of measurement error e i (t) and state variable x i (t) exceeds a set threshold, effectively reducing unnecessary communication and reducing network burden.

[0088] Workflow:

[0089] 1) Continuous monitoring: each node continuously monitors its own state variable and estimation error

[0090] 2) Condition judgment: real-time calculation of whether the trigger condition is met

[0091] 3) Event triggering: send data or update control immediately when the condition is met

[0092] 4) State reset: reset error accumulation after triggering and start a new monitoring period

[0093] Application scenarios include:

[0094] Device layer: inverter state monitoring, only report when voltage and current change significantly

[0095] Microgrid layer: power balance monitoring, only adjust when power imbalance exceeds threshold

[0096] Regional layer: inter-microgrid coordination, only communicate when power needs to be redistributed

[0097] This mechanism is suitable for distributed photovoltaic systems because illumination and load changes have certain slowness and predictability, and do not require high-frequency continuous communication.

[0098] Step S3 When a fault is detected, the system automatically starts the corresponding self-healing control strategy to ensure continuous and stable operation of the system.

[0099] Communication fault detection uses a timestamp-based method:

[0100]

[0101] Where T fault is the communication fault flag, t current is the current time, t last is the last received data time, and T threshold is the fault judgment threshold.

[0102] Detection process:

[0103] 1) Time stamp record: record the receiving time t for each received data packet last

[0104] 2) Real-time monitoring: system clock provides the current time t current

[0105] 3) Timeout judgment: calculate the time difference and compare with the threshold T threshold

[0106] 4) Fault confirmation: if timeout, judge as communication failure

[0107] It also includes a multi-level fault detection mechanism, including device-level fault detection, microgrid-level fault detection, and regional-level fault detection.

[0108] Device-level fault detection uses statistical process control method:

[0109] T 2 = (x - μ) T S -1 (x - μ)

[0110] In the formula, μ is the mean vector of historical normal operation data, S is the covariance matrix of historical normal operation data; T 2 is the Hotelling T 2 statistic, used for multivariate statistical process control; x is the current observed state vector, including voltage, current, temperature, S-1 is the inverse matrix of the covariance matrix, is the matrix transpose symbol;

[0111] The fault detection threshold is:

[0112]

[0113] In the formula, m is the sample size, p is the variable dimension, and B is the beta function.

[0114] Microgrid-level fault detection is based on residual analysis:

[0115]

[0116] In the formula, y(k) is the actual output, is the predicted output.

[0117] Self-healing control uses a hierarchical progressive strategy, with priority levels: device-level self-healing, microgrid-level reconstruction, and regional-level coordination.

[0118] Device-level self-healing uses fault isolation and backup device input; microgrid-level reconstruction bypasses the fault area by changing the network topology; regional-level coordination achieves overall balance by redistributing the power of each microgrid. ​

[0119] Step S5 The system target function considers economy and environmental benefits comprehensively:

[0120] By minimizing the target function J, the optimal power allocation scheme and equipment operation parameters are obtained, and specific control instructions are generated, including: power reference value of each photovoltaic equipment, charging and discharging power of the energy storage system, and microgrid power exchange instruction;

[0121]

[0122] In the formula, C grid (t) is the purchase cost, C maint (t) is the maintenance cost, C emiss (t) is the carbon emission cost, R sell (t) is the electricity selling income.

[0123] The purchase cost calculation formula is:

[0124]

[0125] In the formula, C grid (t) is the purchase cost (yuan) at t time, is the purchase power, π buy (t) is the purchase price, and Δt is the time step (hour).

[0126] The carbon emission cost is calculated as:

[0127]

[0128] In the formula, is the carbon emission factor, π carbon is the carbon price, is the purchase power, and Δt is the time step (hour).

[0129] Robust model predictive control method is used to handle uncertainty throughout the whole process. Uncertainty factors run through the whole hierarchical coordinated control process, and are uniformly handled and compensated in each step by the robust control method. The uncertainty model is established as:

[0130] x(k+1) = Ax(k) + Bu(k) + w(k)

[0131] In the formula, x(k+1) is the system state vector at k+1 time, w(k) is the bounded disturbance, x(k) is the system state vector at k time, including voltage, current and power, A is the system state transition matrix, which is used to describe the system dynamic characteristics; B is the control input matrix, which is used to describe the control action; u(k) is the control input vector at k time; w(k) is the bounded disturbance vector at k time, representing various uncertainty factors, and k is the discrete time step index.

[0132] The robust control tube invariant set is defined as:

[0133]

[0134] where x is the state constraint set, U is the control constraint set, and K is the feedback gain matrix. The optimization problem under uncertainty is formulated as:

[0135] min u max w∈W J(x,u,w)

[0136] where u is the control decision variable, w is the uncertainty parameter, W is the constraint set of the uncertainty parameter; and J(x, u, w) is the objective function value under the uncertainty w.

[0137] This min-max formulation means finding the optimal control strategy in the worst case to ensure the robustness of the system.

[0138] The scenario tree method is used to handle random uncertainty:

[0139]

[0140] where ω i is the i-th scenario, is the realization of the random variable of the i-th scenario at time t.

[0141] The scenario tree method constructs multiple possible future scenarios, each corresponding to a different combination of light and load. The system optimizes all scenarios and selects the most robust control strategy.

[0142] System uncertainties include external environmental uncertainty, device parameter uncertainty, and measurement uncertainty. External environmental uncertainty includes changes in light intensity, cloud cover, and weather changes, temperature fluctuations affecting photovoltaic power generation efficiency, and load randomness, which is the unpredictability of user behavior.

[0143] Device parameter uncertainty includes photovoltaic panel aging, gradual decline in efficiency, inverter parameter drift, changes in line impedance due to temperature and load.

[0144] Measurement uncertainty includes sensor error, measurement accuracy limitations, communication delay, data transmission time uncertainty, and quantization error, A / D conversion accuracy limitations.

[0145] The distributed photovoltaic hierarchical coordination control method takes "three-layer architecture as the core, closed-loop optimization as the logic, and multi-strategy cooperation as the support". Through the hierarchical management and information interaction of the equipment layer, microgrid layer, and regional layer, the stable operation, fault self-healing, and economic and environmental optimization of the distributed photovoltaic system are realized. The specific mechanism is as follows: First, the system takes a three-layer control architecture as the basic framework: the equipment layer focuses on the local precise control of a single distributed photovoltaic device. By improving the perturb and observe method, an adaptive step adjustment mechanism is introduced to dynamically adjust the voltage perturbation according to the power-voltage curve slope to achieve maximum power point tracking (MPPT). At the same time, the reactive power is automatically adjusted according to the grid voltage, and the target power factor is used to back-propagate the reactive power demand to provide stable equipment operation data for the upper layer control. The microgrid layer takes the coordination of power and load balance within the microgrid as the goal. Based on model predictive control (MPC), a multi-time scale optimization model is constructed to optimize power distribution by weighing the state deviation and control action cost of the quadratic objective function. Relying on the state machine switching strategy, the charging and discharging of the energy storage system are controlled according to the state of charge, residual power, and other parameters to maintain voltage stability and power balance within the microgrid. The regional layer coordinates the multi-microgrid from a global perspective. Through a multi-objective optimization function, the operating cost, network loss, carbon emissions, and Lagrange dual decomposition algorithm are combined to decompose the global problem into microgrid sub-problems and solve them in parallel. Combined with the combined load forecasting results, the optimal allocation of power exchange between microgrids is realized, taking into account the system economy and environmental benefits.

[0146] Second, the system ensures reliability through full-process data interaction and closed-loop control: real-time acquisition of operation data at each level, after filtering and outlier removal preprocessing, input into the three-layer architecture; relying on the distributed consistency algorithm, including weighted improvement and event triggering mechanism, only when the state error exceeds the threshold, the information between layers is synchronized to reduce the network communication burden; at the same time, a multi-level fault detection mechanism is constructed at the device level, microgrid level, and regional level. The device level monitors voltage, current, and other variables through the Holt-Lin T 2 statistic, the microgrid level identifies abnormalities based on residual analysis of actual and predicted output, and communication faults are determined by time stamp timeout. Once a fault is detected, a hierarchical progressive self-healing strategy is immediately started, prioritizing device fault isolation and backup investment, followed by microgrid topology reconstruction, and finally regional layer power redistribution to ensure continuous system operation.

[0147] Finally, the system realizes dynamic optimization on the basis of stable operation: first, the voltage stability is judged, and the unstable voltage is eliminated by hierarchical control, including droop control at the equipment layer, secondary control at the microgrid layer to eliminate steady-state error, and power exchange optimization at the regional layer to restore stability; the stable voltage enters the economic and environmental optimization link, and the optimal control parameters such as MPPT parameters at the equipment layer and power distribution instructions at the microgrid layer are output by minimizing the objective function, comprehensively considering the purchase cost, maintenance cost, carbon emission cost and electricity sales revenue, and solving the multi-objective optimization problem by quantitatively calculating the purchase cost and carbon emission cost; at the same time, the system introduces a negative feedback regulation mechanism to evaluate the execution effect of the control instructions, monitor indicators such as voltage deviation and power tracking error, and feedback the deviation data to each layer architecture in reverse, dynamically adjust the control parameters, form a closed-loop optimization, and through the robust model predictive control, construct a tubular invariant set, use the scene tree method to process uncertainties such as light fluctuation, equipment aging and measurement error, ensure the robustness of the control strategy, and finally realize the efficient, stable and economic operation of the distributed photovoltaic system.

[0148] Compared with the prior art, the beneficial effects of the present application are:

[0149] 1. A complete three-layer control architecture is constructed, realizing all-around coordinated control from the equipment level to the regional level, effectively solving the problem of control complexity caused by large-scale access of distributed photovoltaic, and through hierarchical distributed control, the system can realize global coordination while ensuring local optimization, significantly improving the control effect.

[0150] 2. An efficient information transmission and coordination mechanism is designed, which uses a distributed consistency algorithm and an event triggering mechanism to significantly reduce the communication burden while ensuring control performance, significantly reduces the dependence of the system on the communication network, and improves the autonomy and reliability of the system.

[0151] 3. A multi-objective optimization model is established, which considers multiple objectives such as economy, stability and environmental protection, and realizes the comprehensive optimal operation of the system. Through intelligent coordinated control, the system operation cost is reduced by 18.6%, and carbon emissions are reduced by 25.2%, with significant economic and environmental benefits.

[0152] 4. Robust control method is used to effectively handle various uncertainties, significantly improving the anti-interference ability and adaptability of the system. Under uncertain conditions such as light change and load fluctuation, the system can still maintain stable operation, with a voltage deviation controlled within 1.3%.

[0153] 5. A perfect fault detection and self-healing mechanism is designed, and the system has the ability of rapid fault detection, isolation and recovery. In various fault scenarios, the system can complete fault handling within seconds, ensuring the continuous and stable operation of the system. BRIEF DESCRIPTION OF DRAWINGS

[0154] Figure 1 A flow chart of a distributed photovoltaic hierarchical coordination control method based on a microgrid architecture according to the present application;

[0155] Figure 2 A comparison result graph of different control methods of a distributed photovoltaic hierarchical coordination control method based on a microgrid architecture according to the present application on various performance indicators;

[0156] Figure 3 A recovery performance graph of a system under different fault scenarios of a distributed photovoltaic hierarchical coordination control method based on a microgrid architecture according to the present application. DETAILED DESCRIPTION

[0157] The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application.

[0158] See Figure 1 , the specific implementation includes the following steps:

[0159] A three-layer control architecture of a device layer, a microgrid layer and a regional layer is constructed, the device layer is responsible for local control of a single distributed photovoltaic device, the microgrid layer coordinates multiple distributed power sources and loads in a microgrid, and the regional layer coordinates the coordinated operation of multiple microgrid layers;

[0160] The device layer control system is responsible for local control of a single distributed photovoltaic device, and the state variable of the device layer controller is defined as:

[0161] S i =[V i ,I i ,T i ,H i ,θ i ]

[0162] In the formula, V i , I i are voltage and current, T is temperature, H i is humidity, and θ i is a device state parameter;

[0163] The power control equation of the device layer is:

[0164] P i (t)=η i ·A i ·G(t)·[1-β i (T i (t)-T ref )]

[0165] In the formula, P i (t) is the output power of the i-th photovoltaic device, η iFor conversion efficiency, A i For photovoltaic panel area, G(t) is solar radiation intensity, β i For temperature coefficient, T ref For reference temperature;

[0166] The equipment layer realizes maximum power point tracking (MPPT) control by using an improved perturbation and observation method. The improved MPPT algorithm introduces an adaptive step adjustment mechanism, and the formula is as follows:

[0167]

[0168] In the formula, ΔV(k+1) is the voltage perturbation quantity (V) of the k+1 step, that is, the voltage change quantity to be applied next time; ΔV(k) is the voltage perturbation quantity of the k step, that is, the voltage change quantity of the current step; k is a discrete time step index, indicating the iteration number of the MPPT algorithm; γ is an adjustment factor, 0<γ<1, used to control the size of the perturbation step; dP / dV is the derivative of power with respect to voltage (W / V) at the current time, representing the slope of the power-voltage curve; dP / dV(k-1) is the derivative of power with respect to voltage (W / V) at the previous time; P is the output power of the photovoltaic array (W); and V is the terminal voltage of the photovoltaic array (V).

[0169] The equipment layer also includes reactive power control, which automatically adjusts the power factor according to the grid voltage:

[0170] Q i (t)=P i (t)·tan(cos -1 (PF target ))

[0171] In the formula, Q i (t) is the reactive power, P i (t) is the output power of the i-th photovoltaic equipment, PF target is the target power factor, cos -1 (PF target ) is the inverse cosine function, representing the angle corresponding to the target power factor, and its unit is radian; tan(cos -1 (PF target )) is the tangent function, and the input parameter is the above angle; the tangent value is equal to the ratio of the reactive power to the active power, that is, tanφ=Q / P.

[0172] The microgrid layer controls and coordinates multiple distributed power sources and loads in the microgrid, so as to realize power balance and voltage stability in the microgrid. The power balance constraint of the microgrid layer is defined as:

[0173]

[0174] In the formula, Ppv,i (t) is the photovoltaic power generation power, P es,j (t) is the energy storage system power, P grid (t) is the power exchanged with the main grid, P l,k (t) is the load power;

[0175] The microgrid layer adopts a model predictive control (MPC) method to realize coordinated optimization in multiple time scales, and the prediction model is established as follows:

[0176] x(k+1) = Ax(k) + Bu(k) + Ed(k)

[0177] In the formula, x(k) is a state vector, u(k) is a control vector, d(k) is a disturbance vector, and A, B and E are system matrices;

[0178] The optimization objective function of the microgrid layer is set as:

[0179]

[0180] In the formula, N p is the length of the prediction time domain, x(k) is a state vector, u(k) is a control vector, x T (k) is the transpose of x(k), u T (k) is the transpose of u(k), Q, R and S are weight matrices, the objective function is in the form of a quadratic cost, the first term x T (k)Qx(k) represents the penalty of state deviation, the second term u T (k)Ru(k) represents the penalty of control action, and the third term x T (N p )Sx(N p ) is the terminal cost, which ensures the stability of the system at the end of the prediction time domain.

[0181] The state machine switching strategy is adopted for the charge and discharge control of the energy storage system in the microgrid layer:

[0182]

[0183] In the formula, SOC is the state of charge, P surplus is the remaining power, P deficit is the power shortage, and P smooth is the smooth power.

[0184] The regional layer coordinates the coordinated operation of multiple microgrids to realize overall optimization in the region, and the multi-objective optimization function of the regional layer is:

[0185] min J = w1J cost + w2J loss + w3Jemission

[0186] In the formula, J cost For operating costs, J loss For network loss, J emission For carbon emissions, w1, w2, and w3 are weighting coefficients;

[0187] The regional layer employs a distributed optimization algorithm to solve the multi-micronet coordination problem. The distributed algorithm framework based on Lagrange dual decomposition is as follows:

[0188]

[0189] In the formula, f i (x i Let g(x) be the objective function of the i-th micronet, g(x) be the coupling constraint, and λ be the Lagrange multiplier.

[0190] The dual decomposition iterative process is as follows:

[0191] x i k+1 =argminx i [f i (x i )+(λ k ) T g i (x i )]

[0192] λ k+1 =λ k +ρg(x k+1 )

[0193] In the formula, ρ is the step size parameter; x i k+1 f is the decision variable for the i-th microgrid in the (k+1)-th iteration; i (x i Let λ be the local objective function of the i-th microgrid; k Let g be the Lagrange multiplier vector for the k-th iteration, representing the shadow price of the constraint; i (x i ) represents the constraint function associated with the i-th microgrid; ρ is the step size parameter, controlling the update speed of the Lagrange multipliers; λ k+1 Let g(x) be the Lagrange multiplier vector for the (k+1)th iteration; k+1 () represents the value of the global constraint function at the solution of the (k+1)th iteration;

[0194] The regional layer also includes load forecasting, employing combined forecasting methods to improve forecast accuracy:

[0195]

[0196] wherein, is the predicted load, h is the prediction time step, and a1, a2, a3 are combination weights.

[0197] Distributed consensus algorithm is used for information synchronization and coordinated control between layers. The basic form of the consensus algorithm is:

[0198]

[0199] wherein, is the state change rate of the ith node, N i is the neighbor set of node i, a ij is the adjacency matrix element, indicating the connection strength between nodes i and j, and is 0 or 1; j is the index number of the neighbor node; x i is the current state variable of node i, including voltage, frequency, and power; x j is the state variable of neighbor node j, wherein the nodes include device layer nodes, such as individual distributed photovoltaic inverters and energy storage converters; microgrid layer nodes, such as central controllers of individual microgrids; and regional layer nodes, such as regional dispatch centers or master controllers;

[0200] To improve the convergence speed, a weighted consensus algorithm is introduced:

[0201]

[0202] wherein, w ij is the communication weight coefficient between nodes i and j, and 0≤w ij ≤1; u i is the external input signal of node i, including upper control instructions and reference values; is the state change rate of node i, indicating the change speed of the state over time; i is the index number of the current node, including device, microgrid, or regional controller number; j is the index number of the neighbor node; N i is the neighbor set of node i, specifically other nodes directly communicating with node i; x i is the current state variable of node i, including voltage, frequency, and power; x j is the state variable of neighbor node j;

[0203] Event-triggered mechanism is used for inter-layer information transmission to reduce communication burden. The event-triggered condition is designed as:

[0204] |e i (t)|≥α|x i (t)|+β

[0205] wherein, e i (t) is the measurement error of the ith node at time t; x i(t) is the actual state variable of the i-th node at time t, including voltage, power; α, β are trigger threshold parameters.

[0206] Real-time acquisition of operation data of each level, filtering of collected data, elimination of abnormal values, and data import into the three-layer control architecture for execution;

[0207] Through the multi-level fault detection mechanism and self-healing control strategy, the three-layer control architecture of the equipment layer, the microgrid layer and the regional layer is continuously monitored, and when a fault is detected, the hierarchical progressive self-healing control decision is started;

[0208] When a fault is detected, the system automatically starts the corresponding self-healing control strategy to ensure continuous and stable operation of the system;

[0209] The communication fault detection adopts a timestamp-based method:

[0210]

[0211] In the formula, T fault is the communication fault flag, t current is the current time, t last is the last received data time, T threshold is the fault judgment threshold;

[0212] It also includes a multi-level fault detection mechanism, including device-level fault detection, microgrid-level fault detection and regional-level fault detection;

[0213] The device-level fault detection adopts a statistical process control method:

[0214] T 2 = (x-μ) T S -1 (x-μ)

[0215] In the formula, μ is the mean vector of historical normal operation data, and S is the covariance matrix of historical normal operation data; T 2 is the Hotelling T 2 statistic, which is used for multivariate statistical process control; x is the current observed state vector, including voltage, current, temperature, S-1 is the inverse matrix of the covariance matrix, is the matrix transpose symbol;

[0216] The fault detection threshold is:

[0217]

[0218] In the formula, m is the sample size, p is the variable dimension, and B is the beta function;

[0219] The microgrid-level fault detection is based on residual analysis:

[0220]

[0221] where y(k) is the actual output, is the predicted output;

[0222] The self-healing control adopts a hierarchical progressive strategy, with the priority levels being: device-level self-healing, microgrid-level reconstruction, and regional-level coordination.

[0223] The device-level self-healing adopts fault isolation and standby device input; the microgrid-level reconstruction bypasses the fault area by changing the network topology structure; and the regional-level coordination realizes overall balance by reallocating the power of each microgrid.

[0224] For unstable voltage, a hierarchical voltage control strategy is adopted, with each layer coordinating to maintain voltage stability. The device layer voltage regulation adopts droop control, the microgrid layer voltage regulation adopts secondary control to eliminate the steady-state error of droop control, and the regional layer realizes voltage coordination by optimizing the power exchange of each microgrid. The voltage control strategy is selected through the S3 fault diagnosis result;

[0225] For stable voltage, the S4 voltage control strategy is adopted to determine the target function by comprehensively considering economic efficiency and environmental benefits. By solving the multi-objective optimization function, the optimal control parameters of each level are obtained, including the MPPT control parameters of the device layer, the power distribution instructions of the microgrid layer, and the inter-microgrid power exchange instructions of the regional layer, to obtain the final control instructions. The system target function comprehensively considers economic efficiency and environmental benefits:

[0226] By minimizing the target function J, the optimal power distribution scheme and device operation parameters are obtained to generate specific control instructions, including the power reference value of each photovoltaic device, the charging and discharging power of the energy storage system, and the inter-microgrid power exchange instructions. The minimized target function is as follows:

[0227]

[0228] where C grid (t) is the purchase cost, C maint (t) is the maintenance cost, C emiss (t) is the carbon emission cost, R sell (t) is the electricity sales revenue;

[0229] The purchase cost calculation formula is:

[0230]

[0231] where C grid (t) is the purchase cost (yuan) at time t, is the purchase power, π buy(t) is the electricity purchase price, and Δt is the time step (hour) ;

[0232] where the carbon emission cost is calculated as:

[0233]

[0234] where, is the carbon emission factor, π carbon is the carbon price, is the electricity purchase power, and Δt is the time step (hour).

[0235] The execution effect of the final control instruction is evaluated, the deviation data is fed back to the control architecture of each layer, the control parameters are dynamically adjusted, a closed-loop optimization control is formed, and the evaluation data is returned to the optimization architecture in front of the device layer, the microgrid layer, and the regional layer three-layer control architecture.

[0236] The entire process adopts a robust model predictive control method to process uncertainty, and uncertainty factors run through the entire hierarchical coordination control process, and are uniformly processed and compensated in each step through the robust control method, and the uncertainty model is established as:

[0237] x(k+1)=Ax(k)+Bu(k)+w(k)

[0238] In the formula, x(k+1) is a system state vector at k+1 time, x(k) is a system state vector at k time, including voltage, current, and power, A is a system state transition matrix, used to describe the dynamic characteristics of the system; B is a control input matrix, used to describe the control action; u(k) is a control input vector at k time; w(k) is a bounded disturbance vector at k time, representing various uncertainty factors, and k is a discrete time step index;

[0239] The tubular invariant set of the robust control is defined as:

[0240]

[0241] In the formula, x is a state constraint set, U is a control constraint set, and K is a feedback gain matrix;

[0242] The optimization problem under uncertainty is expressed as:

[0243] min u max w∈W J(x,u,w)

[0244] In the formula, u is a control decision variable, w is an uncertainty parameter, W is a constraint set of the uncertainty parameter; and J(x,u,w) is a target function value under the uncertainty w;

[0245] This min-max formulation means to find the optimal control strategy in the worst case to guarantee the robustness of the system.

[0246] The scenario tree method is used to handle the stochastic uncertainty:

[0247]

[0248] where ω i is the i-th scenario, is the realization of the random variable of the i-th scenario at time t;

[0249] The scenario tree method builds multiple possible future scenarios, each corresponding to a different combination of irradiance and load, and the system optimizes over all scenarios to select the most robust control strategy.

[0250] The system uncertainty includes external environment uncertainty, device parameter uncertainty, and measurement uncertainty. The external environment uncertainty includes irradiance variation, cloud cover, weather changes, temperature fluctuations, and load randomness. The device parameter uncertainty includes photovoltaic panel aging, inverter parameter drift, and line impedance variation. The measurement uncertainty includes sensor error, communication delay, and quantization error.

[0251] The system uncertainty includes external environment uncertainty, device parameter uncertainty, and measurement uncertainty. The external environment uncertainty includes irradiance variation, cloud cover, weather changes, temperature fluctuations, and load randomness. The device parameter uncertainty includes photovoltaic panel aging, inverter parameter drift, and line impedance variation. The measurement uncertainty includes sensor error, communication delay, and quantization error.

[0252] The system uncertainty includes external environment uncertainty, device parameter uncertainty, and measurement uncertainty. The external environment uncertainty includes irradiance variation, cloud cover, weather changes, temperature fluctuations, and load randomness. The device parameter uncertainty includes photovoltaic panel aging, inverter parameter drift, and line impedance variation. The measurement uncertainty includes sensor error, communication delay, and quantization error.

[0253] This embodiment establishes a regional power grid simulation model containing three microgrids, each containing distributed photovoltaic, energy storage systems, and various types of loads. The simulation platform is built based on MATLAB / Simulink, with a simulation time step of 0.01 seconds and a total simulation time of 24 hours.

[0254] The simulation scenarios include normal operation, irradiance mutation, load surge, and communication failure. By comparing the system performance under different control strategies, the effectiveness of the proposed method is verified.

[0255] See Figure 2 The comparison results of different control methods on various performance indicators are shown:

[0256] See Figure 3 The recovery performance of the system under different fault scenarios is shown:

[0257] The simulation results show that the hierarchical coordination control method proposed in this patent is significantly better than the traditional method in terms of voltage stability, economy and environmental protection. The system voltage deviation is reduced to 1.3%, the operating cost is reduced by 18.6%, and the carbon emission is reduced by 25.2%. In terms of fault handling, the system has the ability of fast detection and self-healing, and can restore normal operation in a short time.

Claims

1. A distributed photovoltaic hierarchical coordination control method based on microgrid architecture, characterized in that, Comprise the following steps: S1: Construct the device layer, micro-grid layer, regional layer three layer control architecture, the device layer is responsible for the local control of single distributed photovoltaic device, the micro-grid layer coordinates multiple distributed power and load in the micro-grid, the regional layer coordinates the coordinated operation of multiple micro-grid layers; S2: data acquisition and preprocessing, real-time acquisition of operation data of each level, filtering processing to the collected data, eliminating outliers, importing the data into the three layer control architecture for execution; S3: fault monitoring and diagnosis, through multi-level fault detection mechanism and self-healing control strategy, continuously monitor the device layer, micro-grid layer, regional layer three layer control architecture, when detecting fault, start hierarchical progressive self-healing control decision; S4: voltage stability judgment, for unstable voltage, adopt hierarchical voltage control strategy, each layer coordinates to maintain voltage stability, device layer voltage regulation adopts droop control, micro-grid layer voltage regulation adopts secondary control, eliminate the steady-state error of droop control, regional layer realizes voltage coordination through optimizing power exchange of each micro-grid, through S3 fault diagnosis result selects voltage control strategy selection; S5: economic and environmental optimization, for stable voltage and through S4 voltage control strategy through the target function, comprehensive consideration of economy and environmental benefit, determine the target function, through solving multi-objective optimization function, get the optimal control parameters of each level, including device layer MPPT control parameter, micro-grid layer power distribution instruction, regional layer micro-grid power exchange instruction, get the final control instruction; S6: negative feedback regulation, evaluate the execution effect of the final control instruction, feedback the deviation data to each layer control architecture, dynamically adjust the control parameters, form closed loop optimization control, and return the evaluation data to the device layer, micro-grid layer, regional layer three layer control architecture for optimization architecture.

2. The method according to claim 1, wherein, Step S1 device layer control system is responsible for the local control of single distributed photovoltaic device, the state variable of its device layer controller is defined as: S i = [V i , I i , T i , H i , θ i ] wherein V i , I i are voltage and current, respectively, T is temperature, H i is humidity, and Θ i is a device state parameter. The power control equation of the device layer is: P i (t) = η i · A i · G(t) · [1 - β i (T i (t) - T ref )] where P i (t) is the output power of the i-th photovoltaic device, η i is the conversion efficiency, A i is the photovoltaic panel area, G(t) is the solar irradiance, β i is the temperature coefficient, T ref is the reference temperature; The device layer adopts improved perturbation and observation method to realize maximum power point tracking MPPT control, the improved maximum power point tracking MPPT algorithm introduces adaptive step adjustment mechanism, its formula is as follows: In the formula, ΔV(k+1) is the voltage disturbance of k+1 step (V), that is, the voltage change amount to be applied next time; ΔV(k) is the voltage disturbance of k step, that is, the voltage change amount of the current step; K is the discrete time step index, indicating the iteration number of MPPT algorithm; Gamma is an adjustment factor, 0<gamma<1, used to control the size of the disturbance step; dP / dV is the derivative of power to voltage (W / V) at the current time, representing the slope of the power-voltage curve; dP / dV(k-1) is the derivative of power to voltage (W / V) at the previous time; P is the output power of photovoltaic array (W); V is the terminal voltage of photovoltaic array (V).

3. The method of claim 1, wherein, Step S1 device layer also includes reactive power control, which automatically adjusts the power factor according to the grid voltage: Q i (t) = P i (t) tan(cos -1 (PF target )) where Q i (t) is the reactive power, P i (t) is the output power of the i th photovoltaic device, PF target is the target power factor, cos -1 (PF target ) is the inverse cosine function, representing the angle corresponding to the target power factor, with the unit of radian; tan(cos -1 (PF target )) is the tangent function, with the input parameter being the above-mentioned angle; the tangent value is equal to the ratio of the reactive power to the active power, i.e. tanφ = Q / P.

4. The method of claim 1, wherein, Step S1 micro-grid layer control coordinates multiple distributed power and load in the micro-grid, so as to realize power balance and voltage stability in the micro-grid, the power balance constraint of the micro-grid layer is defined as: where P pv,i (t) is the photovoltaic power generation power, P es,j (t) is the energy storage system power, P grid (t) is the power exchanged with the main grid, P l,k (t) is the load power; The micro-grid layer adopts a model predictive control (MPC) method to realize coordinated optimization in multiple time scales, and the prediction model is established as follows: x(k+1)=Ax(k)+Bu(k)+Ed(k) In the formula, x(k) is a state vector, u(k) is a control vector, d(k) is a disturbance vector, and A, B, and E are system matrices. The optimization objective function of the micro-grid layer is set as: where N p is the prediction horizon, x(k) is the state vector, u(k) is the control vector, x T (k) is the transpose of x(k), u T (k) is the transpose of u(k), Q, R, S are weight matrices, the objective function is of the quadratic cost form, the first term x T (k)Qx(k) represents the penalty of state deviation, the second term u T (k)Ru(k) represents the penalty of control action, and the third term x T (N p )Sx(N p ) is the terminal cost, which ensures the stability of the system at the end of the prediction horizon.

5. The method of claim 4, wherein, Step S1 The micro-grid layer controls and coordinates the charging and discharging control of the energy storage system in the micro-grid, and adopts a state machine switching strategy: where SOC is the state of charge, P surplus is the remaining power, P deficit is the power deficit, P smooth is the smoothed power.

6. The method of claim 1, wherein, Step S1 The regional layer coordinates the operation of multiple micro-grids to realize overall optimization in the region, and the multi-objective optimization function of the regional layer is: minJ = w1J cost +w2J loss +w3J emission In the formula, J cost is the operating cost, J loss is the network loss, J emission is the carbon emission, w1, w2, w3 are weight coefficients; The regional layer adopts a distributed optimization algorithm to solve the multi-micro-grid coordination problem, and the distributed algorithm framework based on Lagrange dual decomposition is: where f i (x i ) is the objective function of the ith microgrid, g(x) is the coupling constraint, and λ is the Lagrange multiplier. The dual decomposition iteration process is: x i k+1 = argminx i [f i (x i )+(λ k ) T g i (x i )] λ k+1 = λ k + ρg(x k+1 ) where ρ is the step size parameter; x i k+1 is the decision variable of the ith microgrid in the k+1th iteration; f i (x i ) is the local objective function of the ith microgrid; λ k is the Lagrange multiplier vector of the kth iteration, representing the shadow price of the constraints; g i (x i ) is the constraint function related to the ith microgrid; ρ is the step size parameter, controlling the updating speed of the Lagrange multiplier; λ k+1 is the Lagrange multiplier vector of the k+1th iteration; g(x k+1 ) is the value of the global constraint function at the k+1th iteration solution; The regional layer also includes load prediction, and adopts a combination prediction method to improve prediction accuracy: In the formula, for the prediction load, h is the prediction time step, and a1, a2, a3 are combination weights.

7. The method of claim 1, wherein, Step S1 Specifically, a distributed consistency algorithm is adopted between the control of each layer to realize information synchronization and coordinated control, and the basic form of the consistency algorithm is: wherein, is the state change rate of the ith node, N i is the neighbor set of node i, a ij is the adjacency matrix element, representing the connection strength between node i and j, is 0 or 1; j is the index number of the neighbor node; x i is the current state variable of node i, including voltage, frequency, power; x j is the state variable of neighbor node j, wherein the node includes device layer nodes, individual distributed photovoltaic inverters, energy storage converters, and other device controllers; microgrid layer nodes, central controllers of individual microgrids; regional layer nodes, regional dispatch centers or master controllers; To improve the convergence speed, a weighted consistency algorithm is introduced: where w ij is the communication weight coefficient between nodes i and j, 0≤w ij ≤1; u i is the external input signal of node i, including the upper control instruction, reference value; is the state change rate of node i, indicating the change speed of state over time; i is the index number of the current node, including the device, microgrid or regional controller number; j is the index number of the neighbor node; N i is the neighbor set of node i, specifically other nodes directly communicating with node i; x i is the current state variable of node i, including voltage, frequency, power; x j is the state variable of neighbor node j; The inter-layer information transmission adopts an event-triggered mechanism to reduce the communication burden, and the event-triggered condition is designed as: |e i (t)|≥α|x i (t)|+β In the formula, e i (t) is the measurement error of the ith node at time t; x i (t) is the actual state variable of the ith node at time t, including voltage and power; and α and β are trigger threshold parameters.

8. The method of claim 1, wherein, Step S3 When a fault is detected, the system automatically starts the corresponding self-healing control strategy to ensure continuous and stable operation of the system; The communication fault detection adopts a time-stamp-based method: In the formula, T fault is a communication failure flag, t current is a current time, t last is a last received data time, T threshold is a failure judgment threshold value; It also includes a multi-level fault detection mechanism, including device-level fault detection, micro-grid-level fault detection, and regional-level fault detection; The device-level fault detection adopts a statistical process control method: T 2 = (x - μ) T S -1 (x - μ) where μ is the mean vector of historical normal operation data, S is the covariance matrix of historical normal operation data; T 2 is the Hotelling T 2 statistic for multivariate statistical process control; x is the current observed state vector, including voltage, current, temperature, S-1 is the inverse matrix of the covariance matrix, is the matrix transpose symbol; The fault detection threshold is: In the formula, m is the sample number, p is the variable dimension, and B is the beta function; The micro-grid-level fault detection is based on residual analysis: where y(k) is the actual output, is the predicted output; The self-healing control adopts a hierarchical progressive strategy, and the priority levels are: device-level self-healing, micro-grid-level reconstruction, and regional-level coordination. Among them, the device-level self-healing adopts fault isolation and standby device input; the micro-grid-level reconstruction bypasses the fault area by changing the network topology structure; and the regional-level coordination realizes overall balance by reallocating the power of each micro-grid.

9. The method of claim 1, wherein, Step S5 The system objective function comprehensively considers the economy and environmental benefits: By minimizing the objective function J, the optimal power distribution scheme and device operation parameters are obtained to generate specific control instructions, including the power reference value of each photovoltaic device, the charging and discharging power of the energy storage system, and the power exchange instruction between micro-grids, and the minimized objective function is as follows: In the formula, C grid (t) is the purchase cost of electricity, C maint (t) is the maintenance cost, C emiss (t) is the carbon emission cost, R sell (t) is the electricity sale revenue; The formula for calculating the electricity purchase cost is: In the formula, C grid (t) is the purchase cost of electricity (yuan) at time t, is the purchase power, π buy (t) is the purchase price of electricity, and Δt is the time step (hour). The carbon emission cost is calculated as: wherein is the carbon emission factor, π carbon is the carbon price, is the electricity purchase power, Δt is the time step (hours).

10. The method of claim 1, wherein, A robust model predictive control method is adopted to handle uncertainties, and uncertainties exist throughout the entire hierarchical coordinated control process. Through the robust control method, uncertainties are uniformly handled and compensated in each step, and the uncertainty model is established as: x(k+1)=Ax(k)+Bu(k)+w(k) where x(k+1) is the system state vector at time k+1, x(k) is the system state vector at time k, including voltage, current, power, A is the system state transition matrix, which is used to describe the dynamic characteristics of the system; B is the control input matrix, which is used to describe the control effect; u(k) is the control input vector at time k; w(k) is the bounded disturbance vector at time k, representing various uncertain factors, and k is the discrete time step index; The robust control tube invariant set is defined as: where x is the state constraint set, U is the control constraint set, and K is the feedback gain matrix; The optimization problem under uncertainty is expressed as: min u max w∈W J(x,u,w) where u is a control decision variable, w is an uncertainty parameter, W is a set of constraints for the uncertainty parameter; J(x, u, w) is an objective function value under the uncertainty w; This min-max expression means finding the optimal control strategy in the worst case to ensure the robustness of the system; The scenario tree method is used to handle random uncertainty: where ω i is the i-th scene, is the realization of the random variable of the i-th scene at time t. The scenario tree method constructs multiple possible future scenarios, each corresponding to a different combination of light and load. The system optimizes all scenarios and selects the most robust control strategy.

Images