Multi-micronet coordination regulation method and system based on contribution degree and dynamic compensation under communication packet loss

Abstract

The application discloses a kind of communication packet loss based on contribution degree and dynamic compensation multi-micromesh coordination regulation method and system, method is through building multidimensional comprehensive contribution degree index covering power supply capacity, adjustable power and power supply cost, and using subjective and objective joint empowerment, power distribution model is established considering fairness and economy;At the same time, according to the fluctuation intensity of microgrid parameters, the dynamic compensation mechanism of improved smoothing-filter fusion is adaptively switched to reconstruct missing data;The mathematical expectation of the Lagrange function containing random variable is taken, so that stochastic optimization is converted into deterministic optimization solution, the system includes comprehensive contribution degree calculation module, data acquisition and compensation module, optimization model construction and solution module and instruction issuing module.The application effectively solves the problem of low state information acquisition accuracy and imbalance of distribution fairness and economy under communication limitation, improves the operation robustness of the system under non-ideal communication environment, and has wide application prospect.

Description

Technical Field

[0001] This invention relates to the field of power system operation and control technology, specifically to a multi-microgrid coordinated control method and system based on contribution and dynamic compensation under communication packet loss. Background Technology

[0002] In recent years, with the continuous increase in the proportion of distributed energy in distribution networks, microgrids have gradually evolved into a key technological means to integrate renewable energy and enhance grid flexibility. To ensure the efficient and coordinated operation of microgrids, research on coordinated control strategies for multi-microgrid systems composed of multiple microgrids has become a hot topic in the power system field. In this system configuration, how to achieve efficient power allocation is of significant research value: on the one hand, reasonable power allocation can effectively reduce the total operating cost of the system and improve resource utilization efficiency; on the other hand, by coordinating the output of each microgrid, it is possible to provide ancillary services such as voltage support and frequency regulation for the active distribution network, significantly enhancing grid reliability.

[0003] However, achieving efficient and reliable power coordination and allocation under complex real-world operating conditions remains a significant challenge. Firstly, at the physical level, the heterogeneous characteristics of each microgrid—such as power supply capacity, adjustable power, and operating costs—make it a complex multi-objective trade-off to balance overall scheduling efficiency with the fairness of allocation among microgrids while meeting load demands. Secondly, at the communication level, the surge in multi-microgrid interaction data can easily exceed the limited communication bandwidth, leading to random packet loss, resulting in low accuracy of the state information acquired by the central controller, and even causing oscillations or instability in the control system. Furthermore, when parameters fluctuate drastically, single-dimensional compensation mechanisms often lack adaptive capabilities, leading to significant lag in state reconstruction. Existing methods mainly suffer from the following shortcomings:

[0004] 1. Most methods tend to focus on a single economic goal or technical indicator, and rarely construct a comprehensive evaluation model that covers multiple dimensions such as capacity, cost, and adjustability. This makes it difficult to ensure the fairness of distribution among the participating entities while pursuing overall economic efficiency.

[0005] 2. Existing compensation mechanisms often employ fixed prediction models or static parameters, lacking real-time perception and identification of the intensity of fluctuations in microgrid operating parameters, resulting in low compensation accuracy under complex and variable operating conditions. Furthermore, power allocation models under random packet loss constraints often involve complex random variables, and existing solution algorithms still require further improvement in terms of theoretical convergence proof.

[0006] These issues indicate that there is still significant room for improvement in the application of existing power allocation technologies in non-ideal communication environments, necessitating a new approach to enhance the autonomous operation capabilities and task execution efficiency of multi-microgrid systems in complex environments. Summary of the Invention

[0007] This invention aims to address the shortcomings of existing power allocation methods for multi-microgrids in distribution networks, particularly in balancing fairness and economy, and in handling non-ideal communication environments. It proposes a multi-microgrid coordinated control method and system based on contribution and dynamic compensation under packet loss conditions. By constructing a multi-dimensional comprehensive contribution index to balance fairness and economy, and relying on an adaptive dynamic compensation mechanism and stochastic optimization theory, it can efficiently generate optimal power allocation schemes suitable for random packet loss environments while ensuring the convergence of the algorithm.

[0008] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0009] A multi-micronetwork coordination and control method based on contribution and dynamic compensation under packet loss conditions includes the following steps:

[0010] Step S1: Construct a multi-dimensional comprehensive contribution index that considers power supply capacity, adjustable power and power supply cost. Determine the weight of each index of the multi-dimensional comprehensive contribution index through a subjective and objective weighting method, and calculate the comprehensive contribution of each microgrid.

[0011] Step S2: The central controller collects the operating parameters of each microgrid in real time, including marginal cost, comprehensive contribution, and maximum adjustable power. If normal communication is detected, the original measurement data is used directly. If packet loss occurs, an adaptive compensation mechanism is triggered. In the adaptive compensation mechanism, the fluctuation intensity of key parameters of the microgrid is calculated, and the current operating condition is judged based on a preset threshold. If the packet loss rate is less than the preset threshold, a simple strategy of directly transmitting data combined with fixed old data compensation is adopted. If the packet loss rate is equal to the preset threshold, an improved exponential smoothing method is used for compensation. If the packet loss rate is greater than the preset threshold, an improved Kalman filter method is used for compensation.

[0012] Step S3: Establish a power allocation optimization model with the objective of minimizing the total power supply cost of multiple microgrids and introducing a fairness constraint that the power undertaken by each microgrid is proportional to its comprehensive contribution; use mathematical expectation to process random variables, transforming the power allocation problem under random packet loss constraints into a deterministic optimization problem; design an iterative solution algorithm based on the Karush-Kuhn-Tucker conditions, iteratively updating the power allocation values ​​and Lagrange multipliers of each microgrid until the convergence condition is met to obtain the optimal power allocation result;

[0013] Step S4: Send the calculated optimal power allocation results to the local controllers of each microgrid for execution, thereby compensating for the active power deficit in the distribution network.

[0014] As a preferred technical solution of the present invention: in step S1, the calculation of the comprehensive contribution of each microgrid is specifically as follows:

[0015] Step S11: Select power supply capacity, adjustable power and power supply cost as evaluation indicators, where power supply capacity and adjustable power are positive indicators and power supply cost is a negative indicator.

[0016] Step S12: Standardize the three indicators separately:

[0017] For positive indicators:

[0018] ;

[0019] For negative indicators:

[0020] ;

[0021] In the formula, For micro-network The original index values ​​are for microgrids. Real-time data measured, calculated, or reported by the local control system; This is the minimum value of this indicator for all microgrids; This represents the maximum value of this indicator for all microgrids; The standardized value of this indicator;

[0022] Step S13: Calculate the objective weights of each indicator using the entropy weight method. And combined with preset subjective weights The comprehensive weight of each indicator is calculated by using the optimal combination coefficient method that minimizes the difference between the comprehensive weight and the subjective and objective weights. The details are as follows:

[0023] Substitute the standardized value of the indicator calculated in step S12 into the formula to calculate the indicator percentage:

[0024] ;

[0025] In the formula, Indicates the first Category 1 Indicators The percentage of indicators for individual micro-networks; Indicates the first Category 1 Indicators Indicators for individual microgrids; Represents a local minimum, preventing hour, This renders subsequent calculations meaningless;

[0026] Calculate the information entropy and objective weight of each indicator. :

[0027] ;

[0028] ;

[0029] In the formula, Represents the number of microgrids; It is its maximum entropy value; information entropy The smaller its value, the higher the dispersion of the indicator and the greater its information value; It is the coefficient of difference. The larger the value, the higher the weight. The sum of the objective weights of all indicators is ultimately 1.

[0030] The overall weight is obtained by combining subjective and objective weights:

[0031] ;

[0032] In the formula, subjective weight It was set by humans. These are objective weighting coefficients; These are subjective weighting coefficients; and The difference between the comprehensive weight and the subjective and objective weights is minimized using the least squares criterion: To obtain the optimal combination coefficients , The final comprehensive weight is obtained. ;

[0033] Step S14: Obtain the overall contribution of each microgrid. :

[0034] ;

[0035] In the formula, Indicates the first Category 1 Indicators The standardized value of a microgrid; It is a positive number and its value range is (0,1); For the first The higher the overall contribution value of a microgrid, the better its overall performance in terms of fairness and economy, and the higher its priority should be in the allocation of active power deficit.

[0036] As a preferred technical solution of the present invention: In step S2, the central controller collects the operating parameters of each microgrid in real time, as follows:

[0037] Define the set of key parameters for the microgrid that need compensation:

[0038] ;

[0039] In the formula, (k = 1, 2, 3) respectively correspond to the first, second, third, and third... Micro-network at any time marginal cost Overall contribution With maximum adjustable power The actual value is collected directly by the central controller when communication is successful; when packet loss occurs, an estimated value is generated by a compensation algorithm to replace the actual value in the power allocation calculation.

[0040] As a preferred technical solution of the present invention: in step S2, the fluctuation intensity of the key parameters of the microgrid is calculated. The calculation formula is:

[0041] ;

[0042] In the formula, The actual value of the parameters from the last successful communication before packet loss; The actual value at the previous moment;

[0043] Based on parameter fluctuation intensity With preset threshold By comparing the scenarios, packet loss can be categorized as follows:

[0044] In scenarios with extremely low packet loss rates, the packet loss rate is... ;

[0045] Stable scenarios, ;

[0046] Fluctuation scenarios, ;

[0047] For the different scenarios mentioned above, a simple strategy of directly transmitting data combined with fixed old data compensation, an improved exponential smoothing method, and an improved Kalman filtering method are used for compensation.

[0048] As a preferred technical solution of the present invention, the core formula of the improved exponential smoothing method is:

[0049] ;

[0050] ;

[0051] ;

[0052] In the formula, For the first The first micro-network Class parameters at time The smoothed value; For the previous moment The smoothed value; For smoothing coefficients; For the first The first micro-network Class parameters at time Trend items; Historical trend value; Trend coefficient; This is the compensation value; This is a continuous packet loss counter.

[0053] As a preferred technical solution of the present invention, the improved Kalman filtering method is specifically as follows:

[0054] The improved Kalman filter introduces a state correction term. Sum of squares correction term These are used for state prediction and covariance update, respectively, to suppress error accumulation caused by long-term packet loss. The core iterative formula is:

[0055] ;

[0056] ;

[0057] In the formula, It is the first Individual micro-networks Predicted parameter values ​​at time; It is the first At the last moment, the micro-network State estimates; This is a state correction item; As a correction factor; It is a moment The prediction error covariance matrix; This is the state transition matrix; The input matrix; Input for the system; It was the previous moment The posterior error covariance matrix; It is the state transition matrix The transpose of the matrix; It is the process noise covariance matrix.

[0058] As a preferred technical solution of the present invention, the power allocation optimization model in step S3 is as follows:

[0059] The objective function of the power allocation optimization model is to minimize the total power supply cost, as shown in the formula:

[0060] ;

[0061] The fairness constraint is that the power borne by each microgrid is proportional to its overall contribution:

[0062] ;

[0063] The physical constraints are:

[0064] ;

[0065] ;

[0066] in, For micro-network The power supply cost function; For micro-network The marginal cost is the first derivative of the total power supply cost with respect to the output power. For fixed costs; and The first The micro-network and the first The active power of each microgrid; and The first The micro-network and the first The overall contribution of each microgrid; For the active power deficit of the distribution network; For micro-network Maximum adjustable power.

[0067] As a preferred technical solution of the present invention: in step S3, the Lagrangian function considering random packet loss is constructed as follows:

[0068] ;

[0069] In the formula, For micro-network The power supply cost function; , and All are Lagrange multipliers; This is the marginal cost compensation value in the event of packet loss; This is the maximum adjustable power compensation value during packet loss; This is a fairness penalty coefficient; and For micro-network Hewei.com The overall contribution compensation value;

[0070] By taking the mathematical expectation of the Lagrange function The power allocation problem under random packet loss constraints is transformed into a deterministic optimization problem, as shown in the formula:

[0071] ;

[0072] ;

[0073] ;

[0074] ;

[0075] in:

[0076] Equation (20) is the stationarity condition in the sense of expectation;

[0077] Equation (21) represents the feasibility conditions in the expected sense;

[0078] Equation (22) is the nonnegativity condition of the multipliers in the sense of expectation;

[0079] Equation (23) represents the complementary relaxation condition in the desired sense.

[0080] As a preferred technical solution of the present invention: In step S3, the iterative solution algorithm based on the KKT conditions is specifically as follows:

[0081] S31. Initialize the power allocation to First, initial power is allocated according to the proportion of overall contribution, taking fairness into account; second, the Lagrange multipliers are initialized, and the power balance multipliers are adjusted. To minimize marginal cost and avoid over-incentivizing, use the upper limit multiplier. lower limit multiplier There is no initial penalty; finally, the compensation parameters are initialized. , , By default, there is no packet loss, and the actual values ​​are used to ensure that the initial state is unbiased.

[0082] S32. Power and multiplier updates are performed based on KKT conditions, and the fair mean is updated as follows: The formula for updating the ideal power without projection is:

[0083] ;

[0084] In the formula, The updated value is the ideal power value before projection. For micro-network The overall contribution compensation value; For the first At the current moment, the micro-network The actual output value of active power; The difference between the marginal value of power balance and the marginal cost of the microgrid is used to incentivize the microgrid to output power. The power boundary is adjusted by penalizing the difference between the upper and lower limit constraints. For fairness penalty coefficient, The larger the value, the stronger the effect of fairness on power adjustment;

[0085] S33. Project the ideal power onto the feasible region to avoid power exceeding the upper limit or becoming negative, ensuring that physical constraints are met:

[0086] ;

[0087] if Less than 0, below the lower limit, truncated to 0; if Greater than If it exceeds the upper limit, it will be truncated to... ;

[0088] S34. Update the Lagrange multipliers:

[0089] ;

[0090] Equation (26) is for power balance multiplier updates, where the penalty increases when the total power is insufficient; upper limit multiplier updates, where the penalty increases when the upper limit is exceeded; and lower limit multiplier updates, where the penalty increases when the power is negative, with a step size of [missing information]. Attenuation method is used;

[0091] S35. Perform a convergence test:

[0092] ;

[0093] Since the allocated power has been projected into the non-negative interval, there is no need to determine the lower bound multiplier as the convergence threshold.

[0094] The iteration terminates when equation (27) is satisfied, and the optimal solution is output.

[0095] A multi-microgrid coordination and control system based on contribution and dynamic compensation under packet loss conditions includes:

[0096] Comprehensive contribution calculation module: Based on the collected microgrid data, a multi-dimensional comprehensive contribution index is constructed. The weight of each index is determined by a subjective and objective weighting method, and the comprehensive contribution of each microgrid is calculated.

[0097] Data acquisition and compensation module: Real-time acquisition of operating parameters of each micronet, monitoring of communication status, and dynamic reconstruction of missing information by quantifying the intensity of parameter fluctuations and adaptively switching compensation strategies based on preset thresholds when packet loss occurs.

[0098] The optimization model construction and solution module aims to minimize the total power supply cost and introduces a fairness constraint that the allocated power is proportional to the overall contribution to establish a power allocation optimization model. By taking the mathematical expectation of the Lagrangian function containing random variables, the stochastic optimization problem is transformed into a deterministic optimization problem, and iterative solutions are performed based on the KKT conditions.

[0099] Command distribution module: Distributes the optimal power allocation results to each microgrid for execution.

[0100] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0101] The method of this invention effectively solves the key problem in the prior art that it is difficult to balance fairness and economy in power allocation by constructing a multi-dimensional comprehensive contribution index that covers power supply capacity, adjustable power and power supply cost.

[0102] Meanwhile, by proposing a dual-modal dynamic compensation mechanism that combines adaptive switching improved exponential smoothing and improved Kalman filtering, the data reconstruction accuracy and robustness of the system under random packet loss environment are significantly improved.

[0103] This invention transforms the optimization problem under random packet loss constraints into a deterministic equivalent problem by taking the mathematical expectation of the Lagrange function containing random variables. Furthermore, it rigorously proves the convergence of the algorithm using the Lyapunov method, providing a solid theoretical guarantee for the reliable operation of the system in unstable communication environments and has broad application prospects. Attached Figure Description

[0104] Figure 1 The diagram shown is an architecture diagram of the centralized coordinated control scenario of multi-microgrid access to the distribution network according to the present invention.

[0105] Figure 2 The diagram shown is a flowchart illustrating the overall execution process of the coordination and control strategy of this invention.

[0106] Figure 3 The diagram shown is a schematic representation of the overall coordination and control strategy of this invention.

[0107] Figure 4 The diagram shown is of the IEEE 33-node topology model used in the simulation verification of this invention. Detailed Implementation

[0108] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0109] like Figure 1 The diagram shown illustrates the architecture of a centralized coordinated control scenario for multi-microgrid access in a distribution network, as described in this invention. The central controller of the distribution network and the local controllers of each microgrid interact bidirectionally via a communication network. Based on this architecture, as... Figure 2 The diagram shown is a flowchart of the overall execution process of the coordination and control strategy of this invention. This invention discloses a multi-micronetwork coordination and control method based on contribution and dynamic compensation under packet loss conditions, comprising the following steps:

[0110] Step S1: Construct a multi-dimensional comprehensive contribution index that considers power supply capacity, adjustable power and power supply cost. Determine the weight of each index of the multi-dimensional comprehensive contribution index through a subjective and objective weighting method, and calculate the comprehensive contribution of each microgrid.

[0111] The calculation of the overall contribution of each microgrid is as follows:

[0112] Step S11: Select power supply capacity, adjustable power and power supply cost as evaluation indicators, where power supply capacity and adjustable power are positive indicators and power supply cost is a negative indicator.

[0113] Step S12: Standardize the three indicators separately:

[0114] For positive indicators:

[0115] ;

[0116] For negative indicators:

[0117] ;

[0118] In the formula, For micro-network The original index values ​​are for microgrids. Real-time data measured, calculated, or reported by the local control system; This is the minimum value of this indicator for all microgrids; This represents the maximum value of this indicator for all microgrids; The standardized value of this indicator;

[0119] Step S13: To avoid interference from subjective human bias, the entropy weight method is used to calculate the objective weights of each indicator. And combined with preset subjective weights The comprehensive weight of each indicator is calculated by using the optimal combination coefficient method that minimizes the difference between the comprehensive weight and the subjective and objective weights. The details are as follows:

[0120] Substitute the standardized value of the indicator calculated in step S12 into the formula to calculate the indicator percentage:

[0121] ;

[0122] In the formula, Indicates the first Category 1 Indicators The percentage of indicators for individual micro-networks; Indicates the first Category 1 Indicators Indicators for individual microgrids; Represents a local minimum, preventing hour, This renders subsequent calculations meaningless;

[0123] Calculate the information entropy and objective weight of each indicator. :

[0124] ;

[0125] ;

[0126] In the formula, Represents the number of microgrids; It is its maximum entropy value; information entropy The smaller its value, the higher the dispersion of the indicator and the greater its information value; It is the coefficient of difference. The larger the value, the higher the weight. The sum of the objective weights of all indicators is ultimately 1.

[0127] The overall weight is obtained by combining subjective and objective weights:

[0128] ;

[0129] In the formula, subjective weight It was set by humans. These are objective weighting coefficients; These are subjective weighting coefficients; and The difference between the comprehensive weight and the subjective and objective weights is minimized using the least squares criterion: To obtain the optimal combination coefficients , The final comprehensive weight is obtained. ;

[0130] Step S14: Obtain the overall contribution of each microgrid. :

[0131] ;

[0132] In the formula, Indicates the first Category 1 Indicators The standardized value of a microgrid; It is a positive number and its value range is (0,1), therefore... It is a very small positive number that approaches 0; For the first The higher the overall contribution value of a microgrid, the better its overall performance in terms of fairness and economy, and the higher its priority should be in the allocation of active power deficit.

[0133] Step S2: The central controller collects the operating parameters of each microgrid in real time, including marginal cost, comprehensive contribution, and maximum adjustable power. If normal communication is detected, the original measurement data is used directly. If packet loss occurs, an adaptive compensation mechanism is triggered. In the adaptive compensation mechanism, the fluctuation intensity of key parameters of the microgrid is calculated, and the current operating condition is judged based on a preset threshold. If the packet loss rate is less than the preset threshold, a simple strategy of directly transmitting data combined with fixed old data compensation is adopted. If the packet loss rate is equal to the preset threshold, an improved exponential smoothing method is used for compensation. If the packet loss rate is greater than the preset threshold, an improved Kalman filter method is used for compensation.

[0134] The central controller collects the operating parameters of each microgrid in real time, as follows:

[0135] Define the set of key parameters for the microgrid that need compensation:

[0136] ;

[0137] In the formula, Corresponding to the first Micro-network at any time marginal cost Overall contribution Compared with the actual value of maximum adjustable power ;

[0138] When communication is successful, the central controller directly collects the actual value; when packet loss occurs, an estimated value is generated through a compensation algorithm to replace the actual value in the power allocation calculation.

[0139] Furthermore, considering that multidimensional information is usually encapsulated in the same data packet for transmission in actual communication, this invention sets the parameters of the microgrid to share the same Bernoulli distribution process and defines binary random variables. Indicates the first Micro-network at any time Communication status:

[0140] ;

[0141] In the formula, the packet loss probability in a single communication is: ,satisfy Introduce a continuous packet loss counter. When communication is successful Update to when packet loss occurs .

[0142] Calculate the fluctuation intensity of key parameters of microgrid The calculation formula is:

[0143] ;

[0144] In the formula, The actual value of the parameters from the last successful communication before packet loss; The actual value at the previous moment;

[0145] If packet loss occurs on the first communication, the system faces a typical cold start problem due to the lack of historical observation data. Therefore, this invention presets the fluctuation intensity at the time of the first packet loss to 3%.

[0146] Based on parameter fluctuation intensity With preset threshold By comparing the scenarios, packet loss can be categorized as follows:

[0147] In scenarios with extremely low packet loss rates, the packet loss rate is... A simple strategy of directly transmitting data combined with compensating for old data can meet the requirements for stable system operation.

[0148] Stable scenarios, The parameters change gradually, the difference between adjacent time points is small, and the exponential smooth trend extrapolation can meet the accuracy requirements without the need for complex filtering calculations.

[0149] Fluctuation scenarios, The parameters are affected by factors such as fluctuations in new energy output and sudden load changes, and they change drastically. Kalman filtering is needed to track the dynamic trend and avoid the compensation value from diverging.

[0150] For the different scenarios mentioned above, compensation can be achieved by using a simple strategy of directly transmitting data and compensating with fixed old data, an improved exponential smoothing method, and an improved Kalman filtering method, respectively.

[0151] By introducing a trend term optimization, the compensation value is dynamically extrapolated according to historical trends. The core formula of the improved exponential smoothing method is as follows:

[0152] ;

[0153] ;

[0154] ;

[0155] In the formula, For the first The first micro-network Class parameters at time The smoothed value; For the previous moment The smoothed value; For smoothing coefficients; For the first The first micro-network Class parameters at time Trend items; Historical trend value; Trend coefficient; This is the compensation value; This is a continuous packet loss counter.

[0156] Furthermore, this invention models the dynamic evolution process of a microgrid as an equivalent linear discrete state-space model, and utilizes the process noise term. To absorb the potential nonlinear residuals and unmodeled dynamics of the physical system, the state-space model is as follows:

[0157] ;

[0158] In the formula, This is the state transition matrix; The input matrix; The observation matrix; Input for the system; This is process noise; To observe noise.

[0159] The improved Kalman filter method is as follows:

[0160] The improved Kalman filter introduces a state correction term. Sum of squares correction term These are used for state prediction and covariance update, respectively, to suppress error accumulation caused by long-term packet loss. The core iterative formula is:

[0161] ;

[0162] ;

[0163] In the formula, It is the first Individual micro-networks Predicted parameter values ​​at time; It is the first At the last moment, the micro-network State estimates; This is a state correction item; As a correction factor; It is a moment The prediction error covariance matrix; It was the previous moment The posterior error covariance matrix; It is the state transition matrix The transpose of the matrix; It is the process noise covariance matrix.

[0164] When communication is restored, the system uses the initial observation data to perform measurement updates on the filter. This aims to use Kalman gain to smooth out the step deviation between predicted and actual values ​​and simultaneously calibrate the error covariance, thus reserving accurate prior states for potential subsequent disconnections. During prolonged packet loss, with no observations available, only the state and covariance prediction steps are performed. Correction terms maintain the validity of the predictions, ensuring that the compensation values ​​accurately reflect the current dynamics of the system.

[0165] Step S3: Establish a power allocation optimization model with the objective of minimizing the total power supply cost of multiple microgrids and introducing a fairness constraint that the power undertaken by each microgrid is proportional to its comprehensive contribution; use mathematical expectation to process random variables, transforming the power allocation problem under random packet loss constraints into a deterministic optimization problem; design an iterative solution algorithm based on the Karush-Kuhn-Tucker conditions, iteratively updating the power allocation values ​​and Lagrange multipliers of each microgrid until the convergence condition is met to obtain the optimal power allocation result;

[0166] The power allocation optimization model is as follows:

[0167] The objective function of the power allocation optimization model is to minimize the total power supply cost, as shown in the formula:

[0168] ;

[0169] The fairness constraint is that the power borne by each microgrid is proportional to its overall contribution:

[0170] ;

[0171] The physical constraints are:

[0172] ;

[0173] ;

[0174] in, For micro-network The power supply cost function, Using a linear function not only fits the physical characteristic that the marginal cost of distributed power generation in microgrids is approximately constant, but also allows the parameter to remain slowly changing when there is packet loss in communication, thereby significantly reducing the difficulty of prediction and improving the compensation accuracy of the algorithm. For micro-network The marginal cost is the first derivative of the total power supply cost with respect to the output power. For fixed costs; and The first The micro-network and the first The active power of each microgrid; and The first The micro-network and the first The overall contribution of each microgrid; For the active power deficit of the distribution network; For micro-network Maximum adjustable power.

[0175] The Lagrangian function considering random packet loss is constructed as follows:

[0176] ;

[0177] In the formula, For micro-network The power supply cost function; This is the marginal cost compensation value in the event of packet loss; This is the maximum adjustable power compensation value during packet loss; and For micro-network Hewei.com The overall contribution compensation value; , and All are Lagrange multipliers; The total allocated power must meet the active power deficit to ensure a balance between power supply and demand. Guarantee The power allocated to each microgrid shall not exceed its maximum adjustable power. Guarantee The power distribution of each microgrid is non-negative; This is a fairness penalty coefficient used to adjust the strictness of matching power allocation with overall contribution. and For micro-network Hewei.com The overall contribution compensation value;

[0178] By taking the mathematical expectation of the Lagrange function The power allocation problem under random packet loss constraints is transformed into a deterministic optimization problem, as shown in the formula:

[0179] ;

[0180] ;

[0181] ;

[0182] ;

[0183] In the equation, equation (20) represents the stationarity condition (first derivative condition) in the desired sense. This equation shows that the Lagrangian function has a certain relationship with the active power output of each microgrid. The expected value of the first-order partial derivative is zero; Equation (21) is the feasibility condition in the expected sense, that is, the total output of the multi-microgrid system must meet the given power demand in the expected sense, and the optimal power allocation of each microgrid must be strictly limited to the upper and lower limits of its physical capacity; Equation (22) is the non-negativity condition of the multiplier in the expected sense, which is a mathematical constraint to ensure that the optimization direction does not deviate in the opposite direction; Equation (23) is the complementary relaxation condition in the expected sense, which reveals the strong coupling mechanism between the Lagrange multiplier and the corresponding physical constraints at the optimal solution of the system.

[0184] The iterative solution algorithm based on KKT conditions is as follows:

[0185] S31. Initialize the power allocation to First, initial power is allocated according to the proportion of overall contribution, taking fairness into account; second, the Lagrange multipliers are initialized, and the power balance multipliers are adjusted. To minimize marginal cost and avoid over-incentivizing, use the upper limit multiplier. lower limit multiplier There is no initial penalty; finally, the compensation parameters are initialized. , , By default, there is no packet loss, and the actual values ​​are used to ensure that the initial state is unbiased.

[0186] S32. Power and multiplier updates are performed based on KKT conditions, and the fair mean is updated as follows: The formula for updating the ideal power without projection is:

[0187] ;

[0188] In the formula, The updated value is the ideal power value before projection. For micro-network The overall contribution compensation value; For the first At the current moment, the micro-network The actual output value of active power; The difference between the marginal value of power balance and the marginal cost of the microgrid is used to incentivize the microgrid to output power. The power boundary is adjusted by penalizing the difference between the upper and lower limit constraints. For fairness penalty coefficient, The larger the value, the stronger the effect of fairness on power adjustment;

[0189] S33. Project the ideal power onto the feasible region to avoid power exceeding the upper limit or becoming negative, ensuring that physical constraints are met:

[0190] ;

[0191] if Less than 0, below the lower limit, truncated to 0; if Greater than If it exceeds the upper limit, it will be truncated to... ;

[0192] S34. Update the Lagrange multipliers:

[0193] ;

[0194] Equation (26) is for power balance multiplier updates, where the penalty increases when the total power is insufficient; upper limit multiplier updates, where the penalty increases when the upper limit is exceeded; and lower limit multiplier updates, where the penalty increases when the power is negative, with a step size of [missing information]. Attenuation method is used;

[0195] S35. To ensure the algorithm can reliably obtain the optimal solution and to ensure that the iteration process does not oscillate, a convergence check is performed:

[0196] ;

[0197] To determine the convergence threshold, it is not necessary to check the lower limit multiplier, power, and projection to values ​​greater than or equal to 0.

[0198] The iteration terminates when equation (27) is satisfied, and the optimal solution is output.

[0199] Step S4: Send the calculated optimal power allocation results to the local controllers of each microgrid for execution, thereby compensating for the active power deficit in the distribution network.

[0200] The following describes in detail the specific execution process of the method proposed in the distribution network multi-microgrid system with reference to specific embodiments.

[0201] Combination Figure 3 The schematic diagram shown illustrates the overall scenario of the coordination and control strategy of the present invention. This embodiment also proposes a multi-micronet coordination and control system based on contribution and dynamic compensation under packet loss conditions, used to execute the above-mentioned control method. This system includes:

[0202] Calculation of overall contribution:

[0203] Based on the collected microgrid data, the overall contribution of each microgrid is calculated. Taking five microgrids as an example, their parameters are shown in Table 1. Table 1 shows the specific parameters of the five microgrids.

[0204] Table 1

[0205] The overall contribution is calculated using the entropy weight method combined with subjective weighting. The specific steps are as follows:

[0206] 1) Standardize the indicators (positive indicators: capacity, adjustable power; negative indicators: cost).

[0207] 2) Calculate the information entropy and objective weight of each indicator.

[0208] 3) Combine subjective weights (1 / 3 for daily use, and the cost weight can be increased if economic efficiency is emphasized) and obtain the comprehensive weights through least squares optimization.

[0209] 4) Finally, the overall contribution of each micronetwork is obtained.

[0210] Data collection and packet loss compensation:

[0211] The central controller collects real-time operating parameters of each microgrid via the communication link during each control cycle, including: power supply capacity. (kW), adjustable power (kW) and electricity supply cost (Yuan / kWh). To simulate a real communication environment, it is assumed that packet loss follows a Bernoulli distribution.

[0212] If communication is successful (i.e., the central controller receives the actual data from the microgrid), the data is used directly; if packet loss occurs, an adaptive compensation mechanism is triggered. The specific compensation steps are as follows:

[0213] 1. Calculate the fluctuation intensity of the parameter: For the first... The first micro-network Calculate fluctuation intensity using parameters such as marginal cost, overall contribution, and maximum adjustable power.

[0214] (9);

[0215] 2. Scene segmentation: Based on preset thresholds Divide into stable scenarios ( ) and fluctuation scenarios ( ), and scenarios with extremely low packet loss rates (packet loss rate is ), ).

[0216] 3. Adaptive compensation process:

[0217] Stable scenario: An improved exponential smoothing method is adopted, and a trend term is introduced. To compensate for parameter changes during continuous packet loss, the compensation value is extrapolated from the following formula:

[0218] (10);

[0219] (11);

[0220] (12);

[0221] Fluctuating scenarios: An improved Kalman filter is employed, introducing a state correction term. To suppress the accumulation of errors caused by long-term packet loss, the prediction update formula is as follows:

[0222] (13);

[0223] (14);

[0224] For scenarios with extremely low packet loss rates: A simple strategy of directly transmitting data and compensating with fixed old data can meet the requirements for stable system operation.

[0225] Power allocation optimization model construction:

[0226] A power allocation optimization model is proposed, aiming to minimize the total power supply cost of multiple microgrids and incorporating a fairness constraint that the power undertaken by each microgrid is proportional to its overall contribution:

[0227] (15);

[0228] (16);

[0229] Physical constraints:

[0230] (17);

[0231] (18);

[0232] Considering the parameter uncertainty caused by packet loss in communication, we take the expected value of the Lagrange function, transforming the original problem into a deterministic optimization problem. The Lagrange function is: (19);

[0233] KKT-based iterative algorithm design:

[0234] (1) Initialize the power allocation to Secondly, initialize the Lagrange multipliers and balance the power multipliers. upper limit multiplier lower limit multiplier Finally, initialize the compensation parameters. , , .

[0235] (2) The formula for updating the ideal power without projection is:

[0236] (twenty four);

[0237] (3) Project the ideal power onto the feasible region to avoid power exceeding the upper limit or being negative, and ensure that physical constraints are met:

[0238] (25);

[0239] (4) Update the Lagrange multipliers:

[0240] (26);

[0241] (5) Perform convergence judgment:

[0242] (27);

[0243] The iteration terminates when the above equation is satisfied, and the optimal solution is output.

[0244] This embodiment also proposes a multi-micronet coordinated control system based on contribution and dynamic compensation under communication packet loss, including:

[0245] Comprehensive contribution calculation module: Based on the collected microgrid data, a multi-dimensional comprehensive contribution index is constructed, and the comprehensive contribution of each microgrid is calculated through a subjective and objective weighted evaluation model;

[0246] Data acquisition and compensation module: Real-time acquisition of operating parameters of each micronet, monitoring of communication status, and dynamic reconstruction of missing information by quantifying the intensity of parameter fluctuations and adaptively switching compensation strategies based on preset thresholds when packet loss occurs.

[0247] The optimization model construction and solution module aims to minimize the total power supply cost and introduces a fairness constraint that the allocated power is proportional to the overall contribution to establish a power allocation optimization model. By taking the mathematical expectation of the Lagrangian function containing random variables, the stochastic optimization problem is transformed into a deterministic optimization problem, and iterative solutions are performed based on the KKT conditions.

[0248] Command delivery module: Used to send the calculated optimal power allocation results to the local controllers of each microgrid for execution.

[0249] The present invention will now be described in detail with reference to specific embodiments.

[0250] like Figure 4 The diagram shown is of the IEEE 33-node topology model used in the simulation verification of this invention. A setup was built in the MATLAB environment as follows... Figure 4 The IEEE 33-node distribution network model shown has a system base voltage of 12.66 kV, a base power of 100 MVA, and an active power deficit of 200 kW. The network exhibits a typical radial topology, with node 1 designated as the slack node and its voltage amplitude set at a constant reference of 1.0 pu. The base loads and line impedances of the remaining nodes follow the standard parameters of the IEEE 33-node model. To verify the voltage support capability of the proposed power allocation algorithm under multi-microgrid grid-connected conditions, five microgrids were connected to nodes 6, 10, 18, 22, and 30 of the system. During power flow calculations, each microgrid was set to operate at a constant power factor of 0.95. The packet loss rate varied randomly between 5% and 50%, with an average packet loss rate of 26.3%. The power supply capacity, adjustable power, and power supply cost of each microgrid were set according to Table 1 of the implementation example, and periodic fluctuations and random disturbances were introduced to simulate dynamic operation.

[0251] 1. Comparison of compensation accuracy;

[0252] Taking the power supply capacity of microgrid 1 as an example, the average absolute error (MAE) of 150 time steps was compared between the improved data compensation method and the fixed old data method: the improved method was 1.78 kW, the fixed data method was 1.82 kW, the relative error decreased from 2.01% to 1.89%, and the accuracy improved by about 2.0%.

[0253] 2. Comparison of distributive fairness and economic efficiency;

[0254] The method of this invention was compared with the traditional capacity-based allocation method. The fairness index (variance of power / contribution ratio) decreased from 49.57 to 24.34, a reduction of 50.9%; the average total power supply cost decreased from 96.60 yuan to 90.73 yuan, a saving of 6.08%.

[0255] 3. Performance with extremely low packet loss rate;

[0256] When the packet loss rate is less than 5%, the simple strategy of directly transmitting data and compensating with fixed old data can meet the requirements. The average relative error of marginal cost, comprehensive contribution, and maximum adjustable power is less than 0.1%, indicating that accuracy can be maintained without complex compensation algorithms.

[0257] 4. Voltage stability verification;

[0258] Based on the power allocation results above, power flow calculations were performed using both the method of this invention and the traditional capacity proportional allocation method, and the voltage deviations at key nodes were compared. Under the method of this invention, the maximum voltage deviation at each node was 0.0227%, and the average deviation was 0.0147%, which is far below the safety threshold of ±7%. The total active power loss of the system was 0.1384 MW, which is basically equivalent to 0.1379 MW under the traditional method, proving that the proposed method can still maintain voltage stability under communication packet loss.

[0259] It should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any modifications or equivalent changes made based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A multi-micronetwork coordination and control method based on contribution and dynamic compensation under packet loss conditions, characterized in that, Includes the following steps: Step S1: Construct a multi-dimensional comprehensive contribution index that considers power supply capacity, adjustable power and power supply cost. Determine the weight of each index of the multi-dimensional comprehensive contribution index through a subjective and objective weighting method, and calculate the comprehensive contribution of each microgrid. Step S2: The central controller collects the operating parameters of each microgrid in real time, including marginal cost, comprehensive contribution and maximum adjustable power. If normal communication is detected, the original measurement data is used directly. If packet loss occurs, the adaptive compensation mechanism is triggered. In the adaptive compensation mechanism, the fluctuation intensity of the key parameters of the microgrid is calculated, and the current operating condition is judged based on the preset threshold. If the packet loss rate is less than the preset threshold, a simple strategy of directly transmitting data and compensating with fixed old data is adopted. If the packet loss rate is equal to the preset threshold, the improved exponential smoothing method is used for compensation; if the packet loss rate is greater than the preset threshold, the improved Kalman filtering method is used for compensation. The core formula of the improved exponential smoothing method is: ； ； ； In the formula, For the first The first micro-network Class parameters at time The smoothed value; For the previous moment The smoothed value; For smoothing coefficients; For the first The first micro-network Class parameters at time Trend items; Historical trend value; Trend coefficient; This is the compensation value; For continuous packet loss counter; The improved Kalman filtering method is as follows: The improved Kalman filter introduces a state correction term. Sum of squares correction term These are used for state prediction and covariance update, respectively, to suppress error accumulation caused by long-term packet loss. The core iterative formula is: ； ； In the formula, It is the first Individual micro-networks Predicted parameter values ​​at time; It is the first At the last moment, the micro-network State estimates; This is a state correction item; As a correction factor; It is a moment The prediction error covariance matrix; This is the state transition matrix; The input matrix; Input for the system; It was the previous moment The posterior error covariance matrix; It is the state transition matrix The transpose of the matrix; It is the process noise covariance matrix; Step S3: Establish a power allocation optimization model with the objective of minimizing the total power supply cost of multiple microgrids and introducing a fairness constraint that the power undertaken by each microgrid is proportional to its comprehensive contribution; use mathematical expectation to process random variables, transforming the power allocation problem under random packet loss constraints into a deterministic optimization problem; design an iterative solution algorithm based on the Karush-Kuhn-Tucker conditions, iteratively updating the power allocation values ​​and Lagrange multipliers of each microgrid until the convergence condition is met to obtain the optimal power allocation result; Step S4: Send the calculated optimal power allocation results to the local controllers of each microgrid for execution, thereby compensating for the active power deficit in the distribution network.

2. The multi-micronet coordination and control method based on contribution and dynamic compensation under communication packet loss as described in claim 1, characterized in that, In step S1, the calculation of the comprehensive contribution of each microgrid is as follows: Step S11: Select power supply capacity, adjustable power and power supply cost as evaluation indicators, where power supply capacity and adjustable power are positive indicators and power supply cost is a negative indicator. Step S12: Standardize the three indicators separately: For positive indicators: ； For negative indicators: ； In the formula, For micro-network The original index values ​​are for microgrids. Real-time data measured, calculated, or reported by the local control system; This is the minimum value of this indicator for all microgrids; This represents the maximum value of this indicator for all microgrids; The standardized value of this indicator; Step S13: Calculate the objective weights of each indicator using the entropy weight method. And combined with preset subjective weights The comprehensive weight of each indicator is calculated by using the optimal combination coefficient method that minimizes the difference between the comprehensive weight and the subjective and objective weights. The details are as follows: Substitute the standardized value of the indicator calculated in step S12 into the formula to calculate the indicator percentage: ； In the formula, Indicates the first Category 1 Indicators The percentage of indicators for individual micro-networks; Indicates the first Category 1 Indicators Indicators for individual microgrids; Represents a local minimum, preventing hour, This renders subsequent calculations meaningless; Calculate the information entropy and objective weight of each indicator. : ； ； In the formula, Represents the number of microgrids; It is its maximum entropy value; Information entropy The smaller its value, the higher the dispersion of the indicator and the greater its information value; It is the coefficient of difference. The larger the value, the higher the weight. The sum of the objective weights of all indicators is ultimately 1. The overall weight is obtained by combining subjective and objective weights: ； In the formula, subjective weight It was set by humans. These are objective weighting coefficients; These are subjective weighting coefficients; and The difference between the comprehensive weight and the subjective and objective weights is minimized using the least squares criterion: To obtain the optimal combination coefficients , The final comprehensive weight is obtained. ; Step S14: Obtain the overall contribution of each microgrid. : ； In the formula, Indicates the first Category 1 Indicators The standardized value of a microgrid; It is a positive number and its value range is (0,1); For the first The higher the overall contribution value of a microgrid, the better its overall performance in terms of fairness and economy, and the higher its priority should be in the allocation of active power deficit.

3. The multi-micronet coordination and control method based on contribution and dynamic compensation under packet loss according to claim 1, characterized in that, In step S2, the central controller collects the operating parameters of each microgrid in real time, as follows: Define the set of key parameters for the microgrid that need compensation: ； In the formula, k = 1, 2, 3 respectively correspond to the first, second, third, and third... Micro-network at any time marginal cost Overall contribution With maximum adjustable power The actual value is collected directly by the central controller when communication is successful; when packet loss occurs, an estimated value is generated by a compensation algorithm to replace the actual value in the power allocation calculation.

4. The multi-micronet coordination and control method based on contribution and dynamic compensation under communication packet loss as described in claim 1, characterized in that, In step S2, the fluctuation intensity of the key parameters of the microgrid is calculated. The calculation formula is: ； In the formula, For the first Individual micro-networks The true values ​​of key parameters at any given moment; The actual value at the previous moment; Based on parameter fluctuation intensity With preset threshold By comparing the scenarios, packet loss can be categorized as follows: In scenarios with extremely low packet loss rates, the packet loss rate is... ; Stable scenarios, ; Fluctuation scenarios, ; For the different scenarios mentioned above, a simple strategy of directly transmitting data combined with fixed old data compensation, an improved exponential smoothing method, and an improved Kalman filtering method are used for compensation.

5. The multi-micronet coordination and control method based on contribution and dynamic compensation under communication packet loss as described in claim 1, characterized in that, The power allocation optimization model in step S3 is as follows: The objective function of the power allocation optimization model is to minimize the total power supply cost, as shown in the formula: ； The fairness constraint is that the power borne by each microgrid is proportional to its overall contribution: ； The physical constraints are: ； ； in, For micro-network The power supply cost function; For micro-network The marginal cost is the first derivative of the total power supply cost with respect to the output power. For fixed costs; and The first The micro-network and the first The active power of each microgrid; and The first The micro-network and the first The overall contribution of each microgrid; For the active power deficit of the distribution network; For micro-network Maximum adjustable power.

6. The multi-micronet coordination and control method based on contribution and dynamic compensation under communication packet loss as described in claim 1, characterized in that, In step S3, the Lagrangian function considering random packet loss is constructed as follows: ； In the formula, For micro-network The power supply cost function; , and All are Lagrange multipliers; This is the marginal cost compensation value in the event of packet loss; This is the maximum adjustable power compensation value during packet loss; This is a fairness penalty coefficient; and For micro-network Hewei.com The overall contribution compensation value; By taking the mathematical expectation of the Lagrange function The power allocation problem under random packet loss constraints is transformed into a deterministic optimization problem, as shown in the formula: ； ； ； ； in: Equation (20) is the stationarity condition in the sense of expectation; Equation (21) represents the feasibility conditions in the expected sense; Equation (22) is the nonnegativity condition of the multipliers in the sense of expectation; Equation (23) represents the complementary relaxation condition in the desired sense.

7. The multi-micronet coordination and control method based on contribution and dynamic compensation under communication packet loss as described in claim 6, characterized in that, In step S3, the iterative solution algorithm based on the Karush-Kuhn-Tucker conditions is as follows: S31. Initialize the power allocation to First, initial power is allocated according to the proportion of overall contribution, taking fairness into account; second, the Lagrange multipliers are initialized, and the power balance multipliers are adjusted. To minimize marginal cost and avoid over-incentivizing, use the upper limit multiplier. lower limit multiplier There is no initial penalty; finally, the compensation parameters are initialized. , , By default, there is no packet loss, and the actual values ​​are used to ensure that the initial state is unbiased. S32. Power and multiplier updates are performed based on KKT conditions, and the fair mean is updated as follows: The formula for updating the ideal power without projection is: ； In the formula, The updated value is the ideal power value before projection. For micro-network The overall contribution compensation value; For the first At the current moment, the micro-network The actual output value of active power; The difference between the marginal value of power balance and the marginal cost of the microgrid is used to incentivize the microgrid to output power. The power boundary is adjusted by penalizing the difference between the upper and lower limit constraints. For fairness penalty coefficient, The larger the value, the stronger the effect of fairness on power adjustment; S33. Project the ideal power onto the feasible region to avoid power exceeding the upper limit or becoming negative, ensuring that physical constraints are met: ； if Less than 0, below the lower limit, truncated to 0; if Greater than If it exceeds the upper limit, it will be truncated to... ; S34. Update the Lagrange multipliers: ； Equation (26) is for power balance multiplier updates, where the penalty increases when the total power is insufficient; upper limit multiplier updates, where the penalty increases when the upper limit is exceeded; and lower limit multiplier updates, where the penalty increases when the power is negative, with a step size of [missing information]. Attenuation method is used; S35. Perform a convergence test: ； Since the allocated power has been projected into the non-negative interval, there is no need to determine the lower bound multiplier as the convergence threshold. The iteration terminates when equation (27) is satisfied, and the optimal solution is output.

8. The control system of the multi-micronet coordinated control method based on contribution and dynamic compensation under communication packet loss according to any one of claims 1-7, characterized in that, include: Comprehensive contribution calculation module: Based on the collected microgrid data, a multi-dimensional comprehensive contribution index is constructed. The weight of each index is determined by a subjective and objective weighting method, and the comprehensive contribution of each microgrid is calculated. Data acquisition and compensation module: Real-time acquisition of operating parameters of each micronet, monitoring of communication status, and dynamic reconstruction of missing information by quantifying the intensity of parameter fluctuations and adaptively switching compensation strategies based on preset thresholds when packet loss occurs. The optimization model construction and solution module aims to minimize the total power supply cost and introduces a fairness constraint that the allocated power is proportional to the overall contribution to establish a power allocation optimization model. By taking the mathematical expectation of the Lagrangian function containing random variables, the stochastic optimization problem is transformed into a deterministic optimization problem, and iterative solutions are performed based on the KKT conditions. Command distribution module: Distributes the optimal power allocation results to each microgrid for execution.

Images