High-penetration new energy distribution network evaluation method and device based on multi-intelligent algorithm

Abstract

The present invention discloses a high-penetration new energy distribution network evaluation method and device based on a multi-intelligent algorithm. Taking into account the complex and changeable characteristics of the distribution network state, the theory of complementary advantages between intelligent algorithms is used to obtain a distribution network state estimation value based on a multi-intelligent optimization algorithm. The method includes: first designing a distribution network state estimation function for target optimization, using a genetic algorithm to modify the difference weights in the distribution network state estimation function to obtain a first distribution network state estimation value; using a longhorn beard optimization algorithm to modify the difference weights in the distribution network state estimation function to obtain a second distribution network state estimation value; determining the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value and performing weighted calculation to obtain a distribution network state estimation value. By adopting the present invention, a high-precision distribution network state estimation value can be quickly obtained by weighted fusion of the distribution network state estimation values obtained by the two algorithms.

Description

Technical Field

[0001] The present invention relates to the field of power grid operation and maintenance technology, and in particular to a high-penetration new energy distribution network evaluation method and device based on a multi-intelligent algorithm. Background Art

[0002] The earliest method used for distribution network state estimation is node voltage-based state estimation. Using node voltage as the state variable, it continuously iterates and establishes a mapping relationship between real-time measurements and pseudo-measurements to estimate the grid state. However, this method has long iteration times, low execution efficiency, and requires a large amount of storage space, resulting in significant limitations. The most mature method for estimating distribution network state is the branch current method. Using branch current as the state variable, it estimates the state based on real current measurements and pseudo-measurements. However, this method cannot reflect the reactance characteristics of the grid and may even produce pathological convergence results. The branch power method uses power as the state variable to achieve state estimation, but it cannot adapt to node injection-type measurements.

[0003] Distribution network state estimation methods based on intelligent algorithms are a relatively recent phenomenon, primarily including particle swarm optimization, neural networks, genetic algorithms, and ant colony algorithms. However, distribution network state estimation is a typical NP-hard problem. These intelligent algorithms suffer from slow convergence and large estimation errors. Furthermore, a single algorithm cannot meet the complex and ever-changing demands of power grids, making it difficult to obtain highly accurate distribution network state estimates. Summary of the Invention

[0004] The embodiment of the present invention provides a distribution network state estimation method and device, which quickly obtains a high-precision distribution network state estimation value by weighted fusion of distribution network state estimation values obtained by two algorithms.

[0005] To achieve the above objectives, a first aspect of an embodiment of the present application provides a method for estimating a distribution network state, comprising:

[0006] Acquire real-time measurement data and pseudo-measurement data of a high-penetration new energy distribution network system; the real-time measurement data is collected by a measurement device, and the pseudo-measurement data is obtained based on historical user electricity consumption data;

[0007] Establishing a distribution network state estimation function according to a difference between a quantity measurement in each of the pseudo measurement data and a corresponding quantity measurement in each of the real-time measurement data, and a difference weight corresponding to each difference;

[0008] Modifying the difference weights in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value;

[0009] Using a longhorn beard optimization algorithm to modify the difference weight in the distribution network state estimation function to obtain a second distribution network state estimation value;

[0010] The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined and weighted calculation is performed to obtain the distribution network state estimation value.

[0011] In a possible implementation manner of the first aspect, the distribution network state estimation function is specifically:

[0012] Among them, x is the state change of the distribution network system, z i is the real-time measurement value of the quantity i; h i () is the measurement function of measurement i, w i is the difference weight of measurement i.

[0013] In a possible implementation of the first aspect, modifying the difference weight in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value specifically includes:

[0014] Initialize the population of the genetic algorithm and set various basic parameters;

[0015] The fitness value of each individual is determined based on the error between the distribution network state estimation result and the measured data, and individuals with fitness values greater than the preset threshold are found and saved as excellent individuals;

[0016] The excellent individuals are placed into a new population, and crossover and mutation operations are performed on some individuals, and individuals with fitness values greater than a preset threshold are retained to form a new population until the maximum evolutionary generation is reached;

[0017] According to the individuals in the new population, the difference weights in the distribution network state estimation function are modified to obtain a first distribution network state estimation value.

[0018] In a possible implementation of the first aspect, modifying the difference weight in the distribution network state estimation function using the longhorn beard optimization algorithm to obtain the second distribution network state estimation value specifically includes:

[0019] Initialize the population size, solution space dimension, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the beetle beard algorithm;

[0020] Determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn based on the algorithm distance;

[0021] Calculate and store the fitness function of the longhorn beetle position, and update the step size until the fitness function converges;

[0022] A global optimal solution is determined according to the last position of the longicorn, and the difference weight in the distribution network state estimation function is modified according to the global optimal solution to obtain a second distribution network state estimation value.

[0023] In a possible implementation manner of the first aspect, determining the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value specifically includes:

[0024] The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

[0025] A second aspect of an embodiment of the present application provides a distribution network state estimation device, including:

[0026] A data acquisition module is used to acquire real-time measurement data and pseudo-measurement data of a high-penetration new energy distribution network system; the real-time measurement data is collected by a measurement device, and the pseudo-measurement data is obtained based on the user's historical electricity consumption data;

[0027] A function establishment module, configured to establish a distribution network state estimation function based on a difference between a quantity measurement in each of the pseudo measurement data and a corresponding quantity measurement in each of the real-time measurement data, and a difference weight corresponding to each difference;

[0028] A genetic algorithm module, configured to modify the difference weights in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value;

[0029] A longhorn beard algorithm module, configured to modify the difference weight in the distribution network state estimation function using a longhorn beard optimization algorithm to obtain a second distribution network state estimation value;

[0030] An evaluation module is used to determine a weight of the first distribution network state estimation value and a weight of the second distribution network state estimation value and perform weighted calculation to obtain a distribution network state estimation value.

[0031] In a possible implementation manner of the second aspect, the distribution network state estimation function is specifically:

[0032] Among them, x is the state change of the distribution network system, z i is the real-time measurement value of the quantity i; h i () is the measurement function of measurement i, w i is the difference weight of measurement i.

[0033] In a possible implementation of the second aspect, the genetic algorithm module is specifically configured to:

[0034] Initialize the population of the genetic algorithm and set various basic parameters;

[0035] The fitness value of each individual is determined based on the error between the distribution network state estimation result and the measured data, and individuals with fitness values greater than the preset threshold are found and saved as excellent individuals;

[0036] The excellent individuals are placed into a new population, and crossover and mutation operations are performed on some individuals, and individuals with fitness values greater than a preset threshold are retained to form a new population until the maximum evolutionary generation is reached;

[0037] According to the individuals in the new population, the difference weights in the distribution network state estimation function are modified to obtain a first distribution network state estimation value.

[0038] In a possible implementation of the second aspect, the longicorn beetle whisker algorithm module is specifically configured to:

[0039] Initialize the population size, solution space dimension, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the beetle beard algorithm;

[0040] Determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn based on the algorithm distance;

[0041] Calculate and store the fitness function of the longhorn beetle position, and update the step size until the fitness function converges;

[0042] A global optimal solution is determined according to the last position of the longicorn, and the difference weight in the distribution network state estimation function is modified according to the global optimal solution to obtain a second distribution network state estimation value.

[0043] In a possible implementation manner of the second aspect, determining the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value specifically includes:

[0044] The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

[0045] Compared to the prior art, the present invention provides a distribution network state estimation method and device. This method initializes a population of genetic algorithms and sets parameters such as crossover and mutation. The fitness value of each individual is determined based on the error between the distribution network state estimation result and the actual measurement. The optimal individual is found and saved. Some of the better individuals are directly selected into a new population, and crossover and mutation operations are performed on some of the individuals, retaining the better individuals to form a new population. If the maximum evolution number is reached, the optimal individual is output to obtain a first distribution network state estimate. Furthermore, the method initializes the population size, spatial dimension of the solution, maximum number of iterations, initial step size, step size attenuation factor, hunting space, and location information of the longhorn beetle whisker algorithm. The positions of the longhorn beetle's left and right whiskers are determined, and the position of the longhorn beetle is updated based on the algorithm distance. The fitness function of the longhorn beetle's position is calculated and stored, and the step size is updated. It is determined whether the iteration termination condition is met. If so, the global optimal solution is output to obtain a second distribution network state estimate. The first distribution network state estimate and the second distribution network state estimate are weightedly fused to obtain a final distribution network state estimate. Weighted fusion represents an effective fusion of the results of the genetic algorithm and the beetle beard optimization algorithm, accurately depicting the characteristics of distribution network state changes and improving the accuracy of distribution network state estimation. The estimation results can provide valuable information for distribution network companies and managers.

[0046] The combination of the initialization genetic algorithm and the beetle whisker algorithm overcomes the shortcomings of the single genetic algorithm, such as slow convergence speed and large estimation error. In practical applications, the corresponding weight ratio can be adjusted according to different situations to adapt to the complex and changeable characteristics of the power grid state. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of a method for estimating a distribution network state provided by one embodiment of the present invention;

[0048] Figure 2 is a schematic diagram of a genetic algorithm employed in one embodiment of the present invention;

[0049] Figure 3 Schematic diagram of an embodiment of the present invention using a longicorn beard optimization algorithm. DETAILED DESCRIPTION

[0050] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0051] See Figure 1An embodiment of the present invention provides a method for estimating a distribution network state, comprising:

[0052] S10. Acquire real-time measurement data and pseudo-measurement data of a high-penetration new energy distribution network system; the real-time measurement data is collected by a measurement device, and the pseudo-measurement data is obtained based on historical electricity consumption data of users.

[0053] S11 . Establish a distribution network state estimation function according to a difference between a quantity measurement in each pseudo measurement data and a corresponding quantity measurement in each real-time measurement data, and a difference weight corresponding to each difference.

[0054] S12. Modify the difference weights in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value.

[0055] S13. Using the beetle beard optimization algorithm, modify the difference weight in the distribution network state estimation function to obtain a second distribution network state estimation value.

[0056] S14: Determine a weight of the first distribution network state estimation value and a weight of the second distribution network state estimation value and perform weighted calculation to obtain a distribution network state estimation value.

[0057] Before implementing this method, it is necessary to collect measured data from the distribution network and preprocess it. The 3σ principle and case elimination method are used for data cleaning to remove abnormal or useless data and reduce data noise. The distribution network status data is then normalized. Max-Min is set as the continuous data normalization method to normalize continuous data, and One-Hot encoding is selected to complete discrete data encoding. It should be noted that this embodiment is also applicable to distribution networks with high penetration of new energy access.

[0058] Currently, there are numerous types of distributed power sources connected to distribution networks. These systems estimate their state based on continuous measured data and supplemented power load data. To analyze the performance of a multi-intelligent optimization algorithm for state estimation, we used voltage and current estimates as evaluation criteria and collected data points from the output values of the distribution network system.

[0059] Real-time measurement data is collected by measuring devices. Current measurement data includes load power, branch power, etc. Pseudo-measurement data mainly refers to the power load data used by users, which plays a vital role in estimating the state of the distribution network. For a regional distribution network system, the power load at the input and output points can be measured. The total load of the region is:

[0060] Where: s d and Ω are the loads at the entry and exit points respectively; sj is the load flowing through switch j.

[0061] The first step of this method embodiment is to establish the objective function of state estimation, that is, to establish the distribution network state estimation function. For example, the distribution network state estimation function is specifically:

[0062] Among them, x is the state change of the distribution network system, and the state change is the characteristic quantity of the distribution network state, such as power, voltage, and current; z i is the real-time measurement value of the quantity i, obtained from the real-time measurement data; h i () is the measurement function of the measurement i, which is the fitting function of the historical data and is obtained from the pseudo-measurement data; w i is the difference weight of measurement i.

[0063] The state estimation of the distribution network can be transformed into a constrained optimization function. To accurately estimate the state of the distribution network, the error between the measured value and the estimated value must be minimized as much as possible.

[0064] Exemplarily, S12 specifically includes:

[0065] S120: Initialize the population of the genetic algorithm and set various basic parameters.

[0066] S121. Determine the fitness value of each individual based on the error between the distribution network state estimation result and the measured data, and find individuals whose fitness values are greater than a preset threshold and save them as excellent individuals.

[0067] S122. Place the excellent individuals into a new population, perform crossover and mutation operations on some individuals, and retain individuals with fitness values greater than a preset threshold to form a new population until the maximum evolutionary generation is reached.

[0068] S123: Modify the difference weights in the distribution network state estimation function according to the individuals in the new population to obtain a first distribution network state estimation value.

[0069] See Figure 2Inspired by the natural law of "survival of the fittest," Holland et al. proposed an intelligent optimization algorithm, the genetic algorithm. Genetic algorithms require little prior knowledge of the problem being solved and have few constraints. They optimize and solve problems by simulating population evolution. They possess excellent global search capabilities and implicit parallelism, making them highly efficient for solving complex problems. They have been introduced by scholars into distribution network state estimation. During this process, the genetic algorithm determines an initial population—the set of possible optimal solutions for the distribution network state. It initializes the state variables by setting an evolution counter and a maximum evolution generation number, and randomly generates individuals as the initial population. Individuals in the initial population are evaluated, primarily using the fitness function value. In this embodiment, the distribution network state estimation error is selected as the fitness function value. Then, by simulating the "survival of the fittest" process, the population is continuously evolved toward a more optimal direction, searching for the direction with the minimum distribution network state estimation error. Finally, through selection, crossover, and mutation, the distribution network state estimate with the minimum error is found.

[0070] Exemplarily, S13 specifically includes:

[0071] S130, initializing the population size, spatial dimension of the solution, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the longhorn beetle whisker algorithm.

[0072] S131, determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn according to the algorithm distance.

[0073] S132, calculating and storing the fitness function of the longicorn position, and updating the step size until the fitness function converges.

[0074] S133. Determine a global optimal solution according to the last location of the longicorn, and modify the difference weight in the distribution network state estimation function according to the global optimal solution to obtain a second distribution network state estimation value.

[0075] See Figure 3The Beetle Antennae search algorithm (BAS) is an intelligent optimization algorithm proposed by Jiang et al. in 2017. This algorithm simulates the search method of longhorn beetles when looking for food. It is a single-body search algorithm with a simple principle, few parameters, and low computational complexity. It has advantages when dealing with low-dimensional optimization targets. During the foraging process, longhorn beetles are attracted by the smell of food. Longhorn beetles perceive the smell of food in the air through their two antennae. Since the distances of food from the two antennae are different, the concentrations of the odors perceived by the antennae are also different. When the food is on the left side of the longhorn beetle, the odor concentration perceived by the left antennae is stronger than the odor concentration perceived by the right antennae. Therefore, the longhorn beetle can randomly move towards the side with stronger concentration based on the concentration difference perceived by the two antennae. After multiple iterations, the location of the food is finally found.

[0076] Exemplarily, S14 specifically includes:

[0077] The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

[0078] When allocating weights to the two evaluation values (the weight of the first distribution network state estimation value and the second distribution network state estimation value), the inherent statistical laws and authoritative values between the indicator data should be considered. A reasonable decision indicator weighting method is proposed, which is a combined weighting method that combines subjective weighting method (hierarchical analysis method) and objective weighting method (entropy method) to make up for the shortcomings of single weighting. The weighting method that combines the two weighting methods is called the combined weighting method. The subjective and objective combined weights are:

[0079] The comprehensive weight of the indicator Wj:

[0080] Among them, α j is the weight calculated by the hierarchical analysis method; β j is the weight calculated by the entropy method Q.

[0081] Compared to the prior art, the present invention provides a distribution network state estimation method and device. This method initializes a population of genetic algorithms and sets parameters such as crossover and mutation. The fitness value of each individual is determined based on the error between the distribution network state estimation result and the actual measurement. The optimal individual is found and saved. Some of the better individuals are directly selected into a new population, and crossover and mutation operations are performed on some of the individuals, retaining the better individuals to form a new population. If the maximum evolution number is reached, the optimal individual is output to obtain a first distribution network state estimate. Furthermore, the method initializes the population size, spatial dimension of the solution, maximum number of iterations, initial step size, step size attenuation factor, hunting space, and location information of the longhorn beetle whisker algorithm. The positions of the longhorn beetle's left and right whiskers are determined, and the position of the longhorn beetle is updated based on the algorithm distance. The fitness function of the longhorn beetle's position is calculated and stored, and the step size is updated. It is determined whether the iteration termination condition is met. If so, the global optimal solution is output to obtain a second distribution network state estimate. The first distribution network state estimate and the second distribution network state estimate are weightedly fused to obtain a final distribution network state estimate.

[0082] The combination of the initialization genetic algorithm and the beetle whisker algorithm overcomes the drawbacks of a single genetic algorithm, such as slow convergence and large estimation errors. In practical applications, the algorithm can also adjust the corresponding weights according to different situations to adapt to the complex and changing state of the power grid. Furthermore, the collected real-time data accurately reflects the state changes of the distribution network under high penetration of new energy access, achieving higher distribution network state estimation accuracy than the competing algorithms, providing a new research method for distribution network researchers.

[0083] An embodiment of the present application provides a distribution network state estimation device, including: a data acquisition module, a function establishment module, a genetic algorithm module, a longhorn beard algorithm module and an evaluation module.

[0084] The data acquisition module is used to obtain real-time measurement data and pseudo-measurement data of the high-penetration new energy distribution network system; the real-time measurement data is collected by the measurement device, and the pseudo-measurement data is obtained based on the user's historical electricity consumption data.

[0085] A function establishment module, configured to establish a distribution network state estimation function based on a difference between a quantity measurement in each of the pseudo measurement data and a corresponding quantity measurement in each of the real-time measurement data, and a difference weight corresponding to each difference;

[0086] The genetic algorithm module is used to modify the difference weight in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value.

[0087] The beetle whisker algorithm module is used to modify the difference weight in the distribution network state estimation function using the beetle whisker optimization algorithm to obtain a second distribution network state estimation value.

[0088] An evaluation module is used to determine a weight of the first distribution network state estimation value and a weight of the second distribution network state estimation value and perform weighted calculation to obtain a distribution network state estimation value.

[0089] Exemplarily, the distribution network state estimation function is specifically:

[0090] Among them, x is the state change of the distribution network system, z i is the real-time measurement value of the quantity i; h i () is the measurement function of measurement i, w i is the difference weight of measurement i.

[0091] Exemplarily, the genetic algorithm module is specifically used to:

[0092] Initialize the population of the genetic algorithm and set various basic parameters;

[0093] The fitness value of each individual is determined based on the error between the distribution network state estimation result and the measured data, and individuals with fitness values greater than the preset threshold are found and saved as excellent individuals;

[0094] The excellent individuals are placed into a new population, and crossover and mutation operations are performed on some individuals, and individuals with fitness values greater than a preset threshold are retained to form a new population until the maximum evolutionary generation is reached;

[0095] According to the individuals in the new population, the difference weights in the distribution network state estimation function are modified to obtain a first distribution network state estimation value.

[0096] Exemplarily, the longicorn beard algorithm module is specifically used to:

[0097] Initialize the population size, solution space dimension, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the beetle beard algorithm;

[0098] Determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn based on the algorithm distance;

[0099] Calculate and store the fitness function of the longhorn beetle position, and update the step size until the fitness function converges;

[0100] A global optimal solution is determined according to the last position of the longicorn, and the difference weight in the distribution network state estimation function is modified according to the global optimal solution to obtain a second distribution network state estimation value.

[0101] Exemplarily, determining the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value specifically includes:

[0102] The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

[0103] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment and will not be described again here.

[0104] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A high-penetration new energy distribution network evaluation method based on a multi-intelligent algorithm, characterized in that: include: Obtain real-time measurement data and pseudo-measurement data from high-penetration renewable energy distribution network systems; The real-time measurement data is collected by a measurement device, and the pseudo-measurement data is obtained based on the user's historical electricity consumption data; Establishing a distribution network state estimation function according to a difference between a quantity measurement in each of the pseudo measurement data and a corresponding quantity measurement in each of the real-time measurement data, and a difference weight corresponding to each difference; Modifying the difference weights in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value; Using a longhorn beard optimization algorithm to modify the difference weight in the distribution network state estimation function to obtain a second distribution network state estimation value; Determining a weight of the first distribution network state estimation value and a weight of the second distribution network state estimation value and performing weighted calculation to obtain a distribution network state estimation value; The distribution network state estimation function is specifically: Among them, x is the state change of the distribution network system, z i is the real-time measurement value of the quantity i; h i () is the measurement function of measurement i, w i is the difference weight of measurement i.

2. The high-penetration new energy distribution network evaluation method based on a multi-intelligent algorithm according to claim 1 is characterized in that: The method of modifying the difference weight in the distribution network state estimation function by using a genetic algorithm to obtain a first distribution network state estimation value specifically includes: Initialize the population of the genetic algorithm and set various basic parameters; The fitness value of each individual is determined based on the error between the distribution network state estimation result and the measured data, and individuals with fitness values greater than the preset threshold are found and saved as excellent individuals; The excellent individuals are placed into a new population, and crossover and mutation operations are performed on some individuals, and individuals with fitness values greater than a preset threshold are retained to form a new population until the maximum evolutionary generation is reached; According to the individuals in the new population, the difference weights in the distribution network state estimation function are modified to obtain a first distribution network state estimation value.

3. The high-penetration new energy distribution network evaluation method based on multi-intelligent algorithm according to claim 1 is characterized in that: The method of modifying the difference weight in the distribution network state estimation function by using the beetle beard optimization algorithm to obtain a second distribution network state estimation value specifically includes: Initialize the population size, solution space dimension, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the beetle beard algorithm; Determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn based on the algorithm distance; Calculate and store the fitness function of the longhorn beetle position, and update the step size until the fitness function converges; A global optimal solution is determined according to the last position of the longicorn, and the difference weight in the distribution network state estimation function is modified according to the global optimal solution to obtain a second distribution network state estimation value.

4. The high-penetration new energy distribution network evaluation method based on a multi-intelligent algorithm according to claim 1 is characterized in that: The determining of the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value specifically includes: The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

5. A high-penetration new energy distribution network evaluation device based on a multi-intelligent algorithm, characterized in that: include: A data acquisition module is used to obtain real-time measurement data and pseudo-measurement data of a high-penetration new energy distribution network system; The real-time measurement data is collected by a measurement device, and the pseudo-measurement data is obtained based on the user's historical electricity consumption data; A function establishment module, configured to establish a distribution network state estimation function based on a difference between a quantity measurement in each of the pseudo measurement data and a corresponding quantity measurement in each of the real-time measurement data, and a difference weight corresponding to each difference; A genetic algorithm module, configured to modify the difference weights in the distribution network state estimation function using a genetic algorithm to obtain a first distribution network state estimation value; A longhorn beard algorithm module, configured to modify the difference weight in the distribution network state estimation function using a longhorn beard optimization algorithm to obtain a second distribution network state estimation value; an evaluation module, configured to determine a weight of the first distribution network state estimation value and a weight of the second distribution network state estimation value and perform weighted calculation to obtain a distribution network state estimation value; The distribution network state estimation function is specifically: Among them, x is the state change of the distribution network system, z i is the real-time measurement value of the quantity i; h i () is the measurement function of measurement i, w i is the difference weight of measurement i.

6. The high-penetration new energy distribution network evaluation device based on multi-intelligent algorithm according to claim 5, characterized in that: The genetic algorithm module is specifically used to: Initialize the population of the genetic algorithm and set various basic parameters; The fitness value of each individual is determined based on the error between the distribution network state estimation result and the measured data, and individuals with fitness values greater than the preset threshold are found and saved as excellent individuals; The excellent individuals are placed into a new population, and crossover and mutation operations are performed on some individuals, and individuals with fitness values greater than a preset threshold are retained to form a new population until the maximum evolutionary generation is reached; According to the individuals in the new population, the difference weights in the distribution network state estimation function are modified to obtain a first distribution network state estimation value.

7. The high-penetration new energy distribution network evaluation device based on multi-intelligent algorithm according to claim 5, characterized in that: The longicorn beard algorithm module is specifically used for: Initialize the population size, solution space dimension, maximum number of iterations, initial step size, step size attenuation factor, hunting space and location information of the beetle beard algorithm; Determine the positions of the left and right whiskers of the longicorn, and update the position of the longicorn based on the algorithm distance; Calculate and store the fitness function of the longhorn beetle position, and update the step size until the fitness function converges; A global optimal solution is determined according to the last position of the longicorn, and the difference weight in the distribution network state estimation function is modified according to the global optimal solution to obtain a second distribution network state estimation value.

8. The high-penetration new energy distribution network evaluation device based on multi-intelligent algorithm according to claim 5, characterized in that: The determining of the weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value specifically includes: The weight of the first distribution network state estimation value and the weight of the second distribution network state estimation value are determined respectively by using the hierarchical analysis method and the entropy method.

Images