Layered collaborative management method for medium and low voltage AC / DC power distribution system

Abstract

The invention discloses a hierarchical collaborative management method for a medium and low voltage AC / DC power distribution system. The hierarchical collaborative management method comprises the steps of constructing a dynamic resource aggregation model, collecting and processing data information, and constructing a coupling equation of hierarchical collaborative management. Based on the coupling equation, constructing a hierarchical collaborative management model of a strategic layer, a tactical layer and an execution layer, and optimizing the coupling equation and the hierarchical collaborative management model; and on the basis of the constructed hierarchical collaborative management model, constructing constraint conditions according to internal conditions of operation of the hierarchical collaborative management model and actual conditions of the distributed energy sources. Real-time monitoring and processing of resource data information are achieved by constructing a dynamic resource aggregation model, and real-time, reliable and accurate basic data support is provided for the follow-up process; a multi-level coupling model of a strategic layer, a tactical layer and an execution layer is constructed, full-chain communication from macroscopic planning to real-time control is achieved, the overall operation efficiency of the hierarchical collaborative management model is effectively improved, and flexible aggregation and calling of large-scale distributed resources are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic control of power systems, and more particularly, to a layered collaborative management method for a medium and low voltage AC / DC power distribution system. BACKGROUND

[0002] With large-scale access of intermittent and volatile renewable energy such as wind power and photovoltaic power to power systems, their power generation is mostly dependent on weather, and the output prediction is not accurate, often not matching the peak electricity consumption, resulting in a significant increase in the complexity and uncertainty of power systems.

[0003] The traditional "top-down" centralized management mode of energy is the basis and dominant mode of the power system in most countries at present, which mainly produces electricity by large centralized power plants, transmits through high-voltage transmission networks, and distributes to users through distribution networks. The users are guided to use electricity at different times and to shave peaks and fill valleys mainly through administrative and economic means. However, it is difficult to adapt to the characteristics of distributed energy such as dispersion and intermittency, so it cannot achieve efficient use of distributed energy and cannot well maintain the balance between power supply and demand and stable operation of the system.

[0004] Based on the above, the existing technology begins to try to build a layered control architecture to realize the management and control of distributed energy, but there are problems such as data information collection delay, rough data processing, and large deviation, which leads to poor subsequent management and control, and problems such as rough modeling and insufficient robustness of optimization algorithms when managing and controlling energy, which makes there is still a lot of room for improvement in the energy regulation accuracy and economy of the power system. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the purpose of the embodiments of the present application is to provide a layered collaborative management method for a medium and low voltage AC / DC power distribution system, which solves the problems of significant increase in complexity and uncertainty of the existing power system due to large-scale access of distributed energy to the power system, and rough modeling of interlayer coupling relationship and insufficient robustness of optimization algorithms in the existing layered management architecture.

[0006] In order to achieve the above-mentioned purpose, the embodiments of the present application provide a layered collaborative management method for a medium and low voltage AC / DC power distribution system, comprising the following steps: A dynamic resource aggregation model is constructed to collect and process data information of the medium and low voltage AC / DC power distribution system, and a coupled equation for layered collaborative management is constructed; Based on the coupled equation, a layered collaborative management model of the strategic layer, the tactical layer and the execution layer is constructed, and an improved simulated annealing algorithm is introduced to optimize the coupled equation and the layered collaborative management model; The layered collaborative management model is deployed, and constraint conditions are constructed according to the internal conditions of its operation and the actual situation of distributed energy.

[0007] In a preferred embodiment, the constructing the dynamic resource aggregation model comprises: constructing a cloud-edge collaborative architecture, collecting and preprocessing resource data information of the medium and low voltage AC / DC power distribution system; extracting data features from the preprocessed resource data information, and performing hierarchical clustering based on the extracted data features to obtain a resource cluster; taking the resource cluster as an analysis object, and constructing a dynamic resource aggregation model based on a long short-term memory neural network technology.

[0008] In a preferred embodiment, the constructing the cloud-edge collaborative architecture, collecting and preprocessing resource data information of the medium and low voltage AC / DC power distribution system comprises: collecting real-time data of the distributed energy medium and low voltage DC power distribution system at the edge and preprocessing the data; deep processing the resource data information preprocessed at the edge in the cloud.

[0009] In a preferred embodiment, the extracting data features from the preprocessed resource data information, and performing hierarchical clustering based on the extracted data features to obtain a resource cluster comprises: constructing a forward propagation function to map original high-dimensional resource data to a low-dimensional feature vector; dividing independent clusters, constructing a decoding function, and reconstructing high-dimensional data according to the low-dimensional feature vector; calculating the inter-cluster distance between each independent cluster, and performing same-type aggregation according to the inter-cluster distance to obtain a resource cluster.

[0010] In a preferred embodiment, before the reconstructing high-dimensional data and the calculating the inter-cluster distance between each independent cluster, a loss function is further constructed to adjust and optimize the error between the original high-dimensional resource data and the reconstructed high-dimensional data.

[0011] In a preferred embodiment, the coupling equation is updated in real time based on the improved simulated annealing algorithm under the premise of minimizing cost and minimizing accident handling time.

[0012] In a preferred embodiment, the improved simulated annealing algorithm dynamically adjusts the cooling rate by introducing an acceptance rate cooling coefficient and a variance cooling coefficient; based on the dynamically adjusted cooling rate, the nonlinear logarithmic coupling equation of the update strategy layer to the tactical layer, and the nonlinear logarithmic and quadratic coupling equation of the tactical layer and the execution layer are optimized.

[0013] In a preferred embodiment, the constructing the hierarchical collaborative management model of the strategy layer, the tactical layer, and the execution layer comprises: The strategy layer formulates, generates and outputs a strategy framework of regional energy distribution according to macro data; The tactic layer receives the strategy framework of regional energy distribution output by the strategy layer, combines short-term prediction data, and calculates and outputs power generation proportion and energy storage charging and discharging capacity based on the inter-layer coupling equation from the strategy layer to the tactic layer; The execution layer generates device control instructions and fault handling instructions based on the inter-layer coupling equation between the tactic layer and the execution layer according to the power generation proportion and energy storage charging and discharging capacity output by the tactic layer, combines real-time monitoring data, and performs device control and fault response and feedback to the strategy layer and the tactic layer.

[0014] In a preferred embodiment, the strategy framework of regional energy distribution output by the strategy layer comprises: regions, power generation resources, power generation capacity, energy storage devices and energy storage capacity.

[0015] In a preferred embodiment, the constructed constraint conditions comprise output constraints, energy capacity constraints, power supply and demand balance constraints and stability constraints.

[0016] The present application has the following advantages: (1) The dynamic resource aggregation model is constructed to realize real-time monitoring and processing of resource data information, and to provide real-time and reliable and accurate basic data support for the multi-level coupling model of the strategy layer, the tactic layer and the execution layer. Through the multi-level coupling model of the strategy layer, the tactic layer and the execution layer, the whole chain is connected from macro planning to real-time control, the overall operation efficiency of the hierarchical collaborative management model is effectively improved, and flexible aggregation and calling of large-scale distributed resources are realized. (2) The improved simulated annealing algorithm is introduced to optimize the coupling equation and the hierarchical collaborative management model. The dynamic simulated annealing algorithm is used to quickly converge to obtain the global optimal solution of scheduling, thereby reducing the operation cost of the hierarchical collaborative management of the medium and low voltage AC / DC distribution system; (3) The strategy layer of the hierarchical collaborative management model provides support to the tactic layer, the tactic layer provides guidance to the execution layer, and the execution layer provides real-time feedback and assists the strategy layer and the tactic layer to adjust planning and strategy, thereby forming a closed-loop feedback mechanism, i.e. a closed loop of planning-scheduling-control-feedback, which greatly improves the robustness, reliability and stability of the algorithm in the hierarchical collaborative management model, and realizes efficient management of distributed resources in the power system. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is a flowchart of the internal interaction and work of the hierarchical collaborative management model; Figure 2 It is a flowchart of the execution of the improved simulated annealing algorithm. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0019] As Figure 1 And Figure 2 The hierarchical collaborative management method of the medium and low voltage AC / DC power distribution system comprises the following steps: constructing a dynamic resource aggregation model, collecting and processing data information of the medium and low voltage AC / DC power distribution system, and constructing a coupling equation of hierarchical collaborative management; based on the coupling equation, constructing a hierarchical collaborative management model of the strategic layer, the tactical layer and the execution layer, and introducing an improved simulated annealing algorithm to optimize the coupling equation and the hierarchical collaborative management model; deploying the hierarchical collaborative management model, and constructing constraint conditions according to the internal conditions of its operation and the actual situation of distributed energy.

[0020] The data information of the distributed energy of the power system and the market rules are collected, that is, the data information of the medium and low voltage AC / DC power distribution system is collected, including the current data of the power system, such as power grid topology, load curve, existing power generation and / or energy storage facilities, etc.; and also including the land availability and energy distribution conditions such as light and wind energy of the environment in which the distributed energy of the power system is located; the market rules include renewable energy quota, carbon emission target, subsidy, etc.

[0021] By constructing a dynamic resource aggregation model, the data information of the medium and low voltage AC / DC power distribution system is collected and processed in real time, which provides accurate and effective dynamic basic data support for the establishment of the subsequent coupling equation and the construction of the hierarchical collaborative management model; by constructing a multi-level coupling model of the strategic layer, the tactical layer and the execution layer, the whole chain from macro planning to real-time control can be realized, and the overall operation efficiency of the hierarchical collaborative management model is effectively improved; by constructing the coupling equation of hierarchical collaborative management, the state correlation between layers can be accurately described, the precise coupling modeling is achieved, and the adaptability of the hierarchical collaborative management model to the uncertainty of distributed energy is enhanced; by introducing the improved simulated annealing algorithm to optimize the coupling equation between layers and the hierarchical collaborative management model, the operation cost of the hierarchical collaborative management of the medium and low voltage AC / DC power distribution system is reduced. At the same time, the strategic layer of the hierarchical collaborative management model provides support for the tactical layer, the tactical layer provides guidance for the execution layer, the execution layer provides real-time feedback and assists the strategic layer and the tactical layer to adjust the planning and strategy, thereby forming a closed-loop feedback mechanism, that is, a closed loop of planning-scheduling-control-feedback, which greatly improves the robustness, reliability and stability of the algorithm in the hierarchical collaborative management model, and realizes the efficient management of distributed resources in the power system.

[0022] To provide more accurate and reliable basic data support, the dynamic resource aggregation model is constructed based on the embodiment, and the data information of the medium and low voltage AC / DC power distribution system is further limited. That is: S1 constructing a resource aggregation model A cloud-edge collaborative architecture is constructed to collect and preprocess resource data information of the medium and low voltage AC / DC power distribution system. Data features are extracted from the preprocessed resource data information, and hierarchical clustering is performed based on the extracted data features to obtain a resource cluster. The resource cluster is taken as an analysis object, and a dynamic resource aggregation model is constructed based on a long short-term memory neural network technology. That is: When constructing the resource aggregation model, the cloud-edge collaborative architecture is used to collect data information of edge-side distributed energy, energy storage devices, and controllable loads. Self-encoding technology is used to extract the characteristic features of the resource data. Hierarchical clustering method is used to aggregate small-capacity flexible resource data into a resource cluster. A long short-term memory neural network technology (LSTM technology) is used to construct a dynamic resource aggregation model of the virtual power plant. At the same time, the power adjustable range of each device in the resource cluster at each time is collected to the cluster level, providing basic data support for subsequent hierarchical collaborative management, including resource cluster power adjustable range, spatio-temporal distribution characteristics, etc. Specifically: S11 constructing a cloud-edge collaborative architecture The edge end collects real-time data of the medium and low voltage DC power distribution system of distributed energy and performs preprocessing. The cloud end performs deep processing on the resource data information preprocessed by the edge end. That is: At the edge end, edge computing devices are deployed to collect real-time data of distributed energy, energy storage devices, and controllable loads nearby, and to preprocess the collected real-time data. The deployed edge computing devices include edge servers, intelligent gateways, etc. Distributed energy includes distributed photovoltaic, wind power, etc. Real-time data includes power output, operating state, environmental parameters, and geographic location information, etc. Environmental parameters include light, wind speed, etc. Preprocessing of the collected real-time data includes data cleaning, data compression, etc. Data cleaning refers to removing outliers and filling missing values, etc. to reduce the data transmission and processing pressure of the cloud end.

[0023] In the cloud, a cloud server with strong computing and storage capabilities is configured to receive the preprocessed data uploaded by the edge end, and to perform deep and complex data analysis, model training, and collection and management of cluster power adjustable range, etc. At the same time, remote management and control of the edge end devices are realized to ensure the real-time and reliability of data transmission, and to provide efficient architecture support for data processing and model running of the resource aggregation model.

[0024] S12 encoding, decoding, and aggregation of resource data The forward propagation function is constructed to map the original high-dimensional resource data to a low-dimensional feature vector; according to the low-dimensional feature vector, independent clusters are divided, a decoding function is constructed, and high-dimensional data is reconstructed; the inter-cluster distance between each independent cluster is calculated, and the same type is aggregated according to the inter-cluster distance to obtain a resource cluster.

[0025] Before the reconstruction of the high-dimensional data and the calculation of the inter-cluster distance between each independent cluster, a loss function is constructed to adjust and optimize the error between the original high-dimensional resource data and the reconstructed high-dimensional data. That is: The self-encoding technology is used to extract low-dimensional and essential data features from high-dimensional and complex resource data, and hierarchical clustering processing is performed based on the extracted data features. When the self-encoding technology is used to extract data features, the self-encoder can be used to extract data features. Specifically: S121 resource data encoding The forward propagation process is constructed, that is, the original high-dimensional resource data collected is encoded as follows:

[0026] Where x is the input collected original high-dimensional resource data, h is the low-dimensional feature vector to be extracted, and , n represents the number of features of the original high-dimensional resource data before dimension reduction, and m represents the number of features of the low-dimensional feature vector after dimension reduction; is the weight matrix of the encoder; is the bias vector of the encoder; f is the activation function.

[0027] S122 resource data decoding

[0028] Where, is the data after dimension reduction and reconstruction; is the weight matrix of the decoder; is the bias vector of the decoder; g is the activation function.

[0029] S123 constructing a loss function using mean square error

[0030] The constructed loss function is used to iteratively adjust the data set involved in resource data encoding and decoding, that is, , the error between the original high-dimensional resource data and the reconstructed high-dimensional data can be minimized, and it is ensured that the extracted low-dimensional feature vector can accurately represent the essential characteristics of the original high-dimensional resource data, and the reliability of data acquisition and processing is improved.

[0031] S124 hierarchical clustering of resource data S1241 encodes the low-dimensional feature vectors extracted based on the step S121 resource data, initializes each extracted low-dimensional feature vector, and takes each extracted low-dimensional feature vector as a cluster, namely: 、 ... .

[0032] S1242 calculates the distance between clusters as:

[0033] wherein, represents the number of features of the low-dimensional feature vector after dimension reduction; represents the i-th cluster; represents the j-th cluster.

[0034] The distance between each cluster is combined to form a distance matrix as:

[0035] S1243 merges clusters The two clusters with the shortest distance in the distance matrix are merged to form a new cluster .

[0036] S1244 updates the iteration to obtain the aggregation result The distance between the synthesized new clusters and the distance matrix are updated until the shortest distance between the adjacent two new clusters is greater than a pre-set threshold, and the number of final clusters is obtained, that is, the aggregation result is obtained.

[0037] S1245 constructs a dynamic resource aggregation model Based on the aggregation result, a long short-term memory neural network technology is used to construct and train a dynamic resource aggregation model; after deploying and running the dynamic resource aggregation model, the power adjustable range of each resource cluster at t time is obtained as: The time required from receiving the instruction to reaching the target power value, that is, the response time is: The probability of the schedulable capacity is: .

[0038] The long short-term memory neural network technology (LSTM, Long Short-Term Memory) is a special recurrent neural network RNN specially designed to solve the "gradient disappearance" or "gradient explosion" problem of traditional RNN when processing long sequence data, which can effectively capture the long-term dependence relationship in time series data and is one of the core technologies for processing time series tasks. LSTM is good at processing time series data and adapting to dynamic changes in power, etc. to obtain the power adjustable range of each resource cluster at t time, the response time and the probability of schedulable capacity.

[0039] In summary, the resource aggregation model realizes real-time monitoring and preprocessing of edge resource data information through a cloud-edge collaborative architecture, the cloud performs deep processing based on the accurate data information collected and preprocessed by the edge, and performs hierarchical clustering on the resource data after deep processing, that is, the power adjustable of each device in each resource cluster is collected to the cluster level, providing real-time and reliable, accurate basic data support for subsequent hierarchical collaborative management, laying a foundation for better control of power energy, and helping to realize flexible aggregation and calling of large-scale distributed resources, and improving the feasibility of participating in grid dispatching and power market.

[0040] In this embodiment, on the premise of obtaining real-time and effective basic data by running the dynamic resource aggregation model, the following limitations are made for better hierarchical collaborative management of medium and low voltage AC / DC distribution systems: S2 construct coupling equation The constructed coupling equation is updated in real time based on the improved simulated annealing algorithm under the premise of minimizing cost and minimizing accident handling time. At the same time, based on the data information collected and processed by the dynamic resource aggregation model, the constructed coupling equation contains the characteristics of cluster power adjustable range, space-time distribution, etc., providing accurate and effective basic data support for hierarchical collaborative management.

[0041] The cost includes power generation cost, power purchase cost, energy storage cost, and fault handling time, etc.

[0042] The improved simulated annealing algorithm dynamically adjusts the cooling rate by introducing acceptance rate cooling coefficient and variance cooling coefficient; based on the dynamically adjusted cooling rate, the nonlinear logarithmic coupling equation from the strategic layer to the tactical layer, and the nonlinear logarithmic and quadratic coupling equation between the tactical layer and the execution layer are optimized and updated.

[0043] Specifically: S21 construct nonlinear pointer logarithmic coupling equation from strategic layer to tactical layer: Let the distributed energy regional distribution of the strategic layer be S; The distributed energy generation prediction of the tactical layer is , the user power consumption mode prediction is , the proportion of each distributed energy generation output by the tactical layer is , and the proportion of energy storage charge and discharge capacity is , wherein , , The nonlinear pointer logarithmic coupling equation from the strategic layer to the tactical layer is:

[0044] where a, b, and c are corresponding adjustment parameters, and t is the time.

[0045] In establishing the coupling equation of the strategic layer to the tactical layer, the distributed energy regional distribution of the strategic layer is associated with the distributed energy power generation prediction, the user power consumption mode prediction, the distributed energy power generation proportion and the energy storage charging and discharging capacity proportion of the tactical layer.

[0046] S22 constructs a nonlinear logarithmic and quadratic coupling equation of the tactical layer and the execution layer: The distributed energy power generation proportion output by the tactical layer is , and the energy storage charging and discharging strategy is ; the distributed energy real-time monitoring information of the execution layer is , the fault alarm information is , the equipment control instruction is , and the fault handling instruction is ; wherein , , The nonlinear logarithmic and quadratic coupling equation of the tactical layer and the execution layer is:

[0047] Wherein a, b and c are corresponding adjustment parameters, and t is the time.

[0048] In establishing the coupling equation of the tactical layer and the execution layer, the distributed energy power generation proportion of the tactical layer, the energy storage charging and discharging strategy, the distributed energy real-time monitoring information of the execution layer, the fault alarm information, the equipment control instruction and the fault handling instruction are associated.

[0049] Therefore, in constructing the coupling equation, the distributed energy power generation plan and the energy storage system charging and discharging strategy output by the tactical layer are the basis for the execution layer to generate the equipment control instruction and the fault handling instruction.

[0050] S3 constructs a hierarchical collaborative management model S31 constructs strategic layer management The strategic layer formulates, generates and outputs the strategy framework of regional energy distribution according to macro data.

[0051] Specifically, the macro data includes power system status data and geographic information data, such as power grid topology, load curve, existing power generation and / or energy storage facilities, etc.; geographic information data such as illumination and / or wind energy resource distribution, land availability.

[0052] The strategic layer management is constructed according to macro data guidance, based on power system present situation data and geographic information data, to formulate and output a long-term distribution strategy of regional energy, so as to solve the problems of where to build, how much to build and how to layout. The strategy framework of the regional energy distribution output by the strategic layer includes: region, power generation resource, power generation capacity, energy storage device and energy storage capacity.

[0053] The strategic layer outputs a five-dimensional matrix including region, power generation resource, power generation capacity, energy storage device and energy storage capacity, and the five-dimensional matrix The output example is as follows:

[0054] S32 constructing the tactical layer management The tactical layer receives the strategy framework of the regional energy distribution output by the strategic layer, combines short-term prediction data, and calculates and outputs power generation proportion and energy storage charging and discharging capacity based on the inter-layer coupling equation from the strategic layer to the tactical layer.

[0055] Specifically: The strategy framework of the regional energy distribution output by the strategic layer includes regional distribution information of distributed energy; and the short-term prediction data includes power generation prediction, energy storage state and user electricity consumption mode prediction of the distributed energy.

[0056] Under the strategy framework of the regional energy distribution output by the strategic layer, the tactical layer formulates a day-ahead power generation and / or energy storage scheduling plan in combination with short-term prediction data, so as to solve the problems of how much to generate, how much to store and how to allocate, and calculates and outputs power generation proportion and energy storage charging and discharging capacity based on the inter-layer coupling equation from the strategic layer to the tactical layer. The tactical layer specifically outputs a two-dimensional matrix including power generation proportion and energy storage charging and discharging capacity, and the two-dimensional matrix The output example is as follows:

[0057] S33 constructing the execution layer management The execution layer generates device control instructions and fault handling instructions based on the inter-layer coupling equation from the tactical layer to the execution layer in combination with real-time monitoring data according to the power generation proportion and energy storage charging and discharging capacity output by the tactical layer, and feeds back device control and fault response to the strategic layer and the tactical layer.

[0058] Specifically, the real-time monitoring data includes real-time monitoring information data and fault alarm information data of the distributed energy.

[0059] The execution layer executes the device control and fault response of the distributed energy based on the scheduling plan of the tactical layer to solve the problems of "how to adjust in real time" and "how to repair quickly", controls the devices in real time and handles the faults, realizes real-time adjustment and quick repair, and guarantees stable operation of the system. The specific output of the execution layer is a four-dimensional matrix containing device control instructions and fault handling instructions, and the four-dimensional matrix is The output example is as follows:

[0060] S4 introduces the simulated annealing algorithm S41 improves the simulated annealing algorithm The simulated annealing algorithm is introduced, which dynamically adjusts the cooling rate by introducing an acceptance rate cooling coefficient and a variance cooling coefficient. Specifically, it includes: S411 initializes the conditions , , ; S412 calculates whether to accept according to the Metropolis criterion The Metropolis criterion refers to generating a new state j from the current state i, and if the internal energy of the new state j is less than that of the current state i, accepting the new state j as the new current state; otherwise, accepting the new state j with a probability P.

[0061] According to the Metropolis criterion, the probability P of the particle tending to equilibrium at temperature T is where E is the internal energy at temperature T, and ΔE is the change in its internal energy.

[0062] When , P = 1; when , P = 0. .

[0063] When , it is not accepted; when , it is accepted.

[0064] where is the reference value; k is the Boltzmann constant.

[0065] S413 calculates the iteration acceptance rate

[0066] where represents the number of iterations at this time, represents the number of times accepted according to the Metropolis criterion in K iterations.

[0067] S414 constructs the acceptance rate cooling coefficient

[0068]

[0069] wherein, is an acceptance rate reference value, is a linear parameter factor.

[0070] S415 calculating variance, variance of N steps

[0071] S416 constructing variance cooling coefficient

[0072]

[0073] wherein, is a variance reference value, is a linear parameter factor.

[0074] S317 updating cooling coefficient

[0075] based on the constructed acceptance rate cooling coefficient and the variance cooling coefficient , the cooling coefficient is updated by using a weighting method .

[0076]

[0077] wherein, is a weighting parameter, satisfying .

[0078] by introducing the acceptance rate cooling coefficient and the variance cooling coefficient , the cooling coefficient is updated in real time , and thus the simulated annealing algorithm is improved, and the annealing temperature control is updated in real time according to the improved simulated annealing algorithm.

[0079] S42 optimizing the interlayer coupling equation from the strategic layer to the tactical layer According to the hierarchical collaborative management model, the interlayer coupling equation from the strategic layer to the tactical layer is updated by the improved simulated annealing algorithm, which specifically includes the following: S421 constructing a target function:

[0080] wherein, represents the power generation cost of the i th device; represents the power purchase cost of the i th device; represents the energy storage cost of the i th device.

[0081]

[0082] wherein, represents the i-th device t time distributed resource power generation cost coefficient; represents the i-th device t time distributed resource power generation power.

[0083]

[0084] wherein, represents the i-th device t time purchase power cost; represents the i-th device t time purchase power.

[0085]

[0086] wherein, represents the i-th device t time energy storage device cost coefficient; represents the i-th device t time energy storage power.

[0087] S422 initialization of particle to be optimized parameters Initialization of particle to be optimized parameters , according to the nonlinear index coupling equation from the strategic layer to the tactical layer, the output of each distributed energy generation proportion of the tactical layer is , the energy storage charging and discharging strategy is , according to the random particle position, the initial objective function value is obtained .

[0088] S423 update particle position Update particle position, if , update; if not, according to the Metropolis criterion of step S312 to judge whether to update.

[0089] Each iteration is cooled according to the cooling equation Cooling treatment. Iterate until Iteration is stopped, and the maximum value is taken to obtain the optimized target parameter result .

[0090] S43 optimization of interlayer coupling equation of tactical layer and execution layer According to the hierarchical collaborative management model, the interlayer coupling equation of the tactical layer and the execution layer is updated by the improved simulated annealing algorithm, which specifically includes the following: S431 construct objective function:

[0091] wherein, represents the fault handling time, represents the control time.

[0092] S432 initialize the particle to be optimized parameter initialize the particle to be optimized parameter , the device control instruction of the execution layer output is obtained according to the nonlinear logarithmic and quadratic coupling equation of the tactical layer and the execution layer , the fault handling instruction is , the initial target function value is obtained according to the random particle position .

[0093] S423 update the particle position update the particle position, if is met, update; if not met, judge whether to update according to the Metropolis criterion of step S312.

[0094] temperature reduction processing is carried out according to the temperature reduction equation each time. Iterate until iteration is stopped, take the maximum value, and get the optimized target parameter result .

[0095] S5 construct model constraints The constructed constraint conditions include: output constraint, energy capacity constraint, power supply and demand balance constraint and stability constraint.

[0096] ①Output constraint: The power output power P of the power supply cannot exceed the upper and lower limits: ; ②Resource capacity C constraint: ; ③Power supply and demand balance:

[0097] Among them, represents the output of all power supplies; represents the energy storage output; represents the predicted user demand; ④Stability constraint: Voltage V: ; Frequency f: .

[0098] By constructing the model constraint condition of the hierarchical collaborative management model, it can be ensured that the hierarchical collaborative management model is stable in the safe and feasible domain, so as to achieve the collaborative relationship of the strategic layer, the tactical layer and the execution layer in the hierarchical collaborative management process of the power system, and realize the mutual correlation between each layer.

[0099] Through the introduction of acceptance rate temperature reduction coefficient and variance temperature reduction coefficient , the weight method is used to update the temperature reduction coefficient in real time The improved simulated annealing algorithm is used to update the annealing temperature control in real time, and the improved simulated annealing algorithm is used to update the interlayer coupling equations from the strategic layer to the tactical layer and the interlayer coupling equations from the tactical layer to the execution layer in real time, so that the interlayer coupling relationship of the hierarchical collaborative management model is greatly improved, the interlayer coupling relationship modeling is clear, the robustness problem is improved based on the improved simulated annealing algorithm, and then the energy regulation accuracy and economy of the power system are greatly improved.

[0100] In summary, the hierarchical collaborative management of the medium and low voltage AC-DC power distribution system adopts a hierarchical collaborative optimization mode, constructs a three-layer architecture of a strategic layer, a tactical layer and an execution layer, constructs a nonlinear logarithmic coupling equation from the strategic layer to the tactical layer and a nonlinear logarithmic and quadratic coupling equation from the tactical layer to the execution layer, accurately depicts the interlayer state correlation, enhances the adaptability of the hierarchical collaborative management model to distributed energy uncertainty, realizes the whole chain from macro planning to real-time control, and then improves the overall operation efficiency of the hierarchical collaborative management model; in the hierarchical collaborative management method of the medium and low voltage AC-DC power distribution system, the simulated annealing algorithm is introduced by accepting the cooling rate of the cooling coefficient and the variance cooling coefficient to dynamically adjust the cooling rate, the simulated annealing algorithm is improved, the nonlinear logarithmic coupling equation from the strategic layer to the tactical layer and the nonlinear logarithmic and quadratic coupling equation from the tactical layer to the execution layer are optimized by the improved simulated annealing algorithm, the simulated annealing algorithm with dynamically adjusted cooling rate can quickly converge to the global optimal solution, and the operation cost of the hierarchical collaborative management of the medium and low voltage AC-DC power distribution system is reduced; at the same time, the strategic layer of the hierarchical collaborative management model supports the tactical layer, the tactical layer guides the execution layer, the execution layer provides real-time feedback and assists the strategic layer and the tactical layer to adjust the planning and strategy, and then forms a closed loop feedback mechanism, forms a closed loop of planning-scheduling-control-feedback, greatly improves the robustness and reliability of the algorithm in the hierarchical collaborative management model, and then improves the regulation accuracy and economy of the distributed energy of the power system.

[0101] The above formulas are dimensionless numerical calculations, the formulas are obtained by collecting a large amount of data to simulate the latest real situation, and the preset parameters in the formulas are set by a person skilled in the art according to the actual situation.

[0102] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0103] Those skilled in the art can understand that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0104] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0105] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0106] Finally: the above is only the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems, characterized in that, Includes the following steps: A dynamic resource aggregation model is constructed to collect and process data information from medium and low voltage AC / DC power distribution systems, and a coupling equation for hierarchical collaborative management is constructed. Based on the coupling equation, a hierarchical collaborative management model of strategic, tactical, and execution layers is constructed, and an improved simulated annealing algorithm is introduced to optimize the coupling equation and the hierarchical collaborative management model. The hierarchical collaborative management model is deployed, and constraints are constructed based on its internal operating conditions and the actual situation of distributed energy.

2. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 1, characterized in that, The construction of the dynamic resource aggregation model includes: Construct a cloud-edge collaborative architecture to collect and preprocess resource data information from medium and low voltage AC / DC power distribution systems; Data features are extracted from the preprocessed resource data, and hierarchical clustering is performed based on the extracted data features to obtain resource clusters; Taking the resource cluster as the analysis object, a dynamic resource aggregation model is constructed based on long short-term memory neural network technology.

3. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 2, characterized in that, The construction of the cloud-edge collaborative architecture includes collecting and preprocessing resource data information from medium- and low-voltage AC / DC power distribution systems; including: The edge device collects real-time data from the low-voltage DC power distribution system of distributed energy and performs preprocessing. The cloud performs in-depth processing on the pre-processed resource data information at the edge.

4. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 2 or 3, characterized in that, The step of extracting data features from the preprocessed resource data and performing hierarchical clustering based on the extracted data features to obtain resource clusters includes: Construct a forward propagation function to map the original high-dimensional resource data into low-dimensional feature vectors; Based on the low-dimensional feature vectors, independent clusters are formed, decoding functions are constructed, and high-dimensional data is reconstructed; Calculate the inter-cluster distance between each independent cluster, and aggregate similar clusters based on the inter-cluster distance to obtain resource clusters.

5. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 4, characterized in that, After reconstructing the high-dimensional data and before calculating the inter-cluster distance between each independent cluster, the method further includes: constructing a loss function to adjust and optimize the error between the original high-dimensional resource data and the reconstructed high-dimensional data.

6. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 1, characterized in that, The constructed coupling equations are updated in real time based on the improved simulated annealing algorithm, with the objectives of minimizing cost and shortest accident handling time.

7. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 6, characterized in that, The improved simulated annealing algorithm dynamically adjusts the cooling rate by introducing an acceptance rate cooling coefficient and a variance cooling coefficient. The dynamically adjusted cooling rate is used to optimize and update the nonlinear exponential-logarithmic coupling equation from the strategic layer to the tactical layer, and the nonlinear logarithmic and quadratic coupling equation between the tactical layer and the execution layer.

8. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 1, characterized in that, The hierarchical collaborative management model, comprising strategic, tactical, and execution layers, includes, in sequence: Based on macro data, the strategic level formulates, generates, and outputs a regional-level energy distribution strategy framework; The tactical layer receives the regional energy distribution strategy framework output by the strategic layer, combines it with short-term forecast data, and calculates and outputs the power generation ratio and energy storage charging and discharging capacity based on the inter-layer coupling equation from the strategic layer to the tactical layer. Based on the power generation ratio and energy storage charging and discharging capacity output by the tactical layer, combined with real-time monitoring data, the execution layer generates equipment control commands and fault handling commands based on the inter-layer coupling equation between the tactical and execution layers, executes equipment control and fault response, and feeds back to the strategic and tactical layers.

9. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 8, characterized in that, The strategic framework for regional energy distribution output at the strategic level includes: Region, power generation resources, power generation capacity, energy storage equipment and energy storage capacity.

10. The hierarchical collaborative management method for medium and low voltage AC / DC power distribution systems according to claim 1, characterized in that, The constraints constructed include: output constraints, energy capacity constraints, power supply and demand balance constraints, and stability constraints.

Images