Method, device and computer equipment for optimizing operation of power market of smart micro-grid

Abstract

The application relates to a power market optimization operation method, device and computer equipment of a smart micro-grid. The method comprises the following steps: building a standardized micro-grid intelligent agent state space through multi-source power data, hierarchically deriving an autonomous supply safety boundary and a tradable power boundary, generating a declaration strategy of the intelligent agent under the boundary constraint, and then combining a dispatching result to carry out source-grid-load-storage collaborative scheduling. Therefore, a complete link from data input, safety constraint demarcation, bidding decision to resource scheduling is formed, safety constraints are nested layer by layer, risks of over-limit transaction and supply-demand imbalance are avoided, and market bidding and local power supply guarantee are cooperated.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method, apparatus and computer equipment for optimizing the operation of a smart microgrid in the power market. Background Technology

[0002] With the advancement of new power system construction, the rapid integration of flexible resources such as distributed photovoltaic power, distributed wind power, energy storage devices, adjustable loads, and electric vehicles on the distribution side has made smart microgrids an important platform for realizing local consumption of new energy, improving power supply reliability, and participating in electricity market transactions. During operation, smart microgrids not only need to ensure continuous power supply to local loads and critical loads, but also need to participate in power purchase, sale, and regulation services based on price signals from the day-ahead market, intraday market, real-time market, and ancillary service market, achieving synergistic optimization of security and economy.

[0003] Existing microgrid optimization methods often aim to minimize operating costs, maximize trading revenue, or maximize renewable energy integration rates. They typically treat the microgrid as a single controlled entity for dispatch, lacking dynamic assessment of its autonomous supply capacity. In situations of increased uncertainty in renewable energy output or sudden changes in load demand, relying solely on price signals for market bidding can easily lead to insufficient energy storage reserves, reduced critical load protection capabilities, exceeding grid connection limits, and increased deviation assessment costs. This makes it difficult to meet the actual needs of smart microgrids for safe and reliable participation in the electricity market. Summary of the Invention

[0004] Based on this, the purpose of this application is to provide a method, apparatus and computer equipment for a smart microgrid to participate in the optimized operation of the electricity market while ensuring local power supply security and continuous power supply to critical loads, so as to solve the technical problems mentioned in the background art.

[0005] Firstly, this application provides a method for optimizing the operation of a smart microgrid's electricity market. Based on a microgrid intelligent agent, it includes: Based on multi-source power data, construct the state space of the microgrid intelligent agent; The autonomous supply security boundary is determined by the state space under different scheduling cycles, and the tradable power boundary is generated based on the autonomous supply security boundary. With the microgrid intelligent agent as the decision-making core, a market application strategy is generated within the tradable power boundary; Based on the market clearing results, implement coordinated dispatch of power generation, grid, load and storage in smart microgrids.

[0006] In one embodiment, determining the autonomous supply security boundary under different scheduling cycles through the state space includes: determining the critical load guarantee rate and energy storage support duration under different scheduling cycles based on the critical load demand associated with the state space and the predicted output of distributed power sources; weighting the normalized energy storage support duration and the determined islanded operation margin, reserve capacity margin, and local supply and demand balance to obtain a comprehensive index; and determining the autonomous supply security boundary based on the comprehensive index.

[0007] In one embodiment, the above method further includes: initially adjusting the tradable power boundary based on any one of critical load guarantee rate, reserve capacity margin, and comprehensive indicators.

[0008] In one embodiment, generating a tradable power boundary based on the autonomous supply security boundary includes: determining multiple power parameters based on the autonomous supply security boundary; the power parameters include at least the maximum discharge power of energy storage, the maximum rechargeable power of energy storage, the adjustable load response capacity, and the maximum allowable exchange power at the grid connection point; determining the upper limit of the tradable power by using the maximum discharge power of energy storage and the adjustable load response capacity; obtaining the upper limit of the purchasable power by using the maximum rechargeable power of energy storage and the maximum allowable exchange power at the grid connection point; and generating the tradable power boundary by combining the upper limit of the tradable power and the upper limit of the purchasable power; negative values ​​in the tradable power boundary indicate the direction of power purchase, and positive values ​​in the interval indicate the direction of power sale.

[0009] In one embodiment, a market application strategy is generated within the tradable power boundary using a microgrid agent as the decision-making core. This includes: processing the state space through the microgrid agent with the tradable power boundary as a constraint to obtain an initial market application action; the market application action includes the applied electricity quantity, applied price, trading direction, and reserve capacity configuration; obtaining the objective function and optimizing the parameters based on the initial market application action and the objective function until the target market application action under the optimal solution of the objective function is obtained; and supplementing the trading period information under the target market application action to obtain the market application strategy.

[0010] In one embodiment, the scheduling constraints satisfied by the coordinated dispatch of source, grid, load, and storage include at least power balance constraints, energy storage state of charge constraints, and grid connection point exchange power constraints. The method further includes: when executing market-traded electricity, determining power balance constraints based on the grid connection point exchange power, load adjustment power, and energy storage charging power of the smart microgrid; determining energy storage state of charge constraints based on the energy storage charging efficiency, energy storage discharging efficiency, dispatch time interval, and rated energy storage capacity of the smart microgrid; and determining grid connection point exchange power constraints based on the maximum allowable exchange power of the grid connection point of the smart microgrid.

[0011] In one embodiment, after performing source-grid-load-storage coordinated scheduling of the smart microgrid, the method further includes: collecting the actual operating results after source-grid-load-storage coordinated scheduling, and iteratively updating the microgrid agent based on the actual operating results and market clearing results; obtaining various indicator risks and corresponding target thresholds determined after the microgrid agent is updated; the indicator risks include at least the self-sufficiency supply capacity indicator, critical load guarantee risk, and market deviation assessment risk; and switching the operating mode of the smart microgrid based on the indicator risks and target thresholds.

[0012] In one embodiment, the microgrid agent is iteratively updated based on actual operating results and market clearing results, including: obtaining the microgrid agent's preliminary prediction results; obtaining the prediction deviation based on the actual operating results and the preliminary prediction results; determining the operating deviation between the actual operating conditions and the scheduled operating conditions; obtaining the transaction deviation based on the market declaration strategy and the market clearing results; mapping the penalty costs corresponding to the prediction deviation, operating deviation, and transaction deviation; determining the revenue deviation between the actual revenue and the expected revenue; constructing a learning signal for the microgrid agent through multiple penalty costs and revenue deviations; and performing self-learning updates on the network weights of the microgrid agent using the learning signal.

[0013] Secondly, this application provides a smart microgrid power market optimization operation device. It includes: The space construction and boundary determination module is used to construct the state space of the microgrid intelligent agent based on multi-source power data; determine the autonomous supply security boundary under different scheduling cycles through the state space; and generate the tradable power boundary based on the autonomous supply security boundary.

[0014] The strategy determination and scheduling execution module is used to generate market declaration strategies within the tradable power boundary with the microgrid intelligent agent as the decision-making core, and execute the coordinated scheduling of source, grid, load and storage of the smart microgrid based on the market clearing results.

[0015] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the above-described method for optimizing the operation of the power market in a smart microgrid.

[0016] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-described method for optimizing the operation of the power market in a smart microgrid.

[0017] The aforementioned smart microgrid's power market optimization operation methods, devices, computer equipment, and readable storage media construct a standardized microgrid intelligent agent state space through multi-source power data. It then derives the autonomous supply security boundary and tradable power boundary in layers, using these boundaries to constrain the intelligent agent's declaration strategy. Finally, it combines the clearing results to conduct source-grid-load-storage coordinated scheduling. Therefore, it forms a complete chain from data input, security constraint delineation, bidding decision-making to resource scheduling, with nested security constraints. This not only avoids the risks of over-limit trading and supply-demand imbalance but also achieves coordination between market-based bidding and local power supply security. Attached Figure Description

[0018] Figure 1 This is an application environment diagram of a smart microgrid power market optimization operation method in one embodiment; Figure 2 This is a flowchart illustrating a method for optimizing the operation of a smart microgrid's electricity market in one embodiment. Figure 3 This is a flowchart illustrating a method for optimizing the operation of a smart microgrid's electricity market, as described in another embodiment. Figure 4 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] The smart microgrid power market optimization operation method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. Server 104 acquires multi-source power data sent by terminal 102 and constructs the state space of the microgrid intelligent agent; it determines the autonomous supply security boundary under different scheduling cycles through the state space, and generates the tradable power boundary based on the autonomous supply security boundary. Server 104 is also used to generate market declaration strategies within the tradable power boundary with the microgrid intelligent agent as the decision-making core, and executes the source-grid-load-storage coordinated scheduling of the smart microgrid based on the market clearing results, so that terminal 102 can present and display the scheduling instructions. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, smart vehicle devices, and portable wearable devices. Server 104 can be implemented using a standalone server or a server cluster composed of multiple servers, and can be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Microgrid agents include deep neural networks and can be deployed on servers or independently.

[0021] In one embodiment, such as Figure 2 As shown, a method for optimizing the operation of a smart microgrid's electricity market is provided, which can be applied to... Figure 1 The server in the middle includes the following steps: Step 202: Construct the state space of the microgrid intelligent agent based on multi-source power data.

[0022] The multi-source power data includes smart microgrid operation data, power market data, and external environmental data. Smart microgrid operation data includes distributed photovoltaic power output data, distributed wind power output data, energy storage status of charge data, energy storage charging and discharging power data, base load data, adjustable load data, interruptible load data, critical load data, grid connection point switching power data, microgrid bus voltage data, branch power flow data, and equipment operating status data. Power market data includes day-ahead market price data, intraday market price data, real-time market price data, ancillary service market price data, market declaration rule data, declaration period data, clearing price data, traded volume data, deviation assessment rule data, and deviation assessment fee data. External environmental data includes solar irradiance forecast data, wind speed forecast data, temperature forecast data, weather warning data, grid dispatch instruction data, upper-level grid operation constraint data, and user energy consumption behavior data.

[0023] The server abstracts the smart microgrid into a microgrid intelligent agent with capabilities such as state perception, supply and demand forecasting, autonomous supply assessment, market application decision-making, scheduling execution, and feedback learning. The state space of the microgrid intelligent agent includes source-side state, energy storage-side state, load-side state, grid-side state, market-side state, and risk-side state.

[0024] Specifically, after acquiring the initial multi-source power data, the server performs time-scale unification, missing value correction, outlier removal, and normalization on this data to form multi-source power data that can be identified by the microgrid intelligent agent. Based on the multi-source power data, the server constructs the state space of the microgrid intelligent agent, uniformly mapping the states of the source side, energy storage side, load side, grid side, market side, and risk side into state feature vectors. This provides a data foundation for subsequent autonomous supply capacity assessment and market application strategy generation, and characterizes the supply and demand status, regulation capacity, trading capacity, and operational risks of the smart microgrid during the corresponding scheduling period through the state feature vectors.

[0025] In one embodiment, let the first The state feature vector for each scheduling period is: ;

[0026] in, Predicting output for distributed power sources In the state of energy storage charge, This is the total load forecast value. For critical load demand, For power exchange at the grid connection point, This is a market price forecast. The risk coefficient for deviation assessment This is a comprehensive uncertainty indicator for new energy output and load forecasting.

[0027] Step 204: Determine the autonomous supply security boundary under different scheduling cycles through the state space, and generate the tradable power boundary based on the autonomous supply security boundary.

[0028] The entire scheduling cycle typically covers the time interval of the entire process, including pre-forecasting, safety boundary calculation, generation of declaration strategies, market clearing, internal source-grid-load-storage scheduling execution, and post-cycle review and learning; it includes both market trading periods and the real-time scheduling operation period of microgrid internal equipment after clearing.

[0029] Specifically, based on the state feature vector of the state space, the server initially assesses the autonomous supply capability of the smart microgrid in different scheduling cycles in the future, and obtains a variety of autonomous supply capability indicators, including the predicted available output of new energy sources, the duration of energy storage support, the critical load guarantee rate, the islanding operation margin, the grid-connected switching power margin, the reserve capacity margin, and the local supply and demand balance. Then, the autonomous supply security boundary is determined based on multiple autonomous supply capability indicators.

[0030] The server determines the grid operation constraints associated with the smart microgrid, such as the state of charge of energy storage, energy storage charging and discharging power constraints, critical load guarantee constraints, maximum exchange power at grid connection points, and reserve capacity constraints. Because the deep neural network of the microgrid agent performs load and renewable energy power forecasts before making decisions, the server obtains preliminary forecast results. Therefore, based on the autonomous supply security boundary, various grid operation constraints, and the predicted output of distributed power sources and predicted load demand from the preliminary forecast results, the server determines the tradable power boundary of the smart microgrid in each trading period. This allows the smart microgrid to clearly define the range of purchasable power, the range of tradable power, and the range of reserve capacity before participating in the electricity market.

[0031] In one embodiment, the preliminary forecast results are generated by an independent forecasting module to predict supply and demand, constrain power boundaries, and guide the generation of application actions. After the trading cycle ends, the actual operating results are compared with the preliminary forecast results to calculate the forecast deviation, which is then converted into a penalty cost to construct a learning signal. The agent's strategy, risk, and boundary parameters are updated through gradient iteration to continuously optimize the market application strategy adapted to the forecast conditions.

[0032] Step 206: Using the microgrid smart agent as the decision-making core, generate a market application strategy within the tradable power boundary.

[0033] Specifically, taking the microgrid intelligent agent as the decision-making core and within the tradable power boundary, and combining market price forecasts (including multi-source power data and pre-forecast results), distributed power output forecasts, load forecasts, energy storage operation constraints, and deviation assessment risks, a vector representation of market application actions is generated. Then, the market application actions are analyzed as vectors, and supplemented with supporting information such as trading time period information and trading type identifiers to obtain the market application strategy. The market application strategy includes the applied electricity volume, applied price, trading direction, trading time period, reserve capacity configuration, and deviation risk coefficient.

[0034] Step 208: Based on the market clearing results, perform coordinated scheduling of power generation, grid, load and storage in the smart microgrid.

[0035] The market clearing results include the clearing price, the traded volume, the untraded volume, the trading direction, and the execution period. The coordinated dispatch instructions for power generation, grid, load, and storage include instructions for distributed power generation output adjustment, energy storage charging, energy storage discharging, adjustable load response, interruptible load control, grid connection point switching power control, and critical load protection.

[0036] Specifically, the microgrid smart agent uploads its market declaration strategy to the power trading platform. The power trading platform aggregates the declaration information from all market participants, performs unified matching and calculation, and outputs the market clearing results. Based on the market clearing results, the autonomous supply security boundary, and the grid operation constraints associated with the smart microgrid, the server generates source-grid-load-storage coordinated dispatch instructions to control the operation modes of distributed power sources, energy storage devices, adjustable loads, and grid connection interfaces.

[0037] Furthermore, after executing the source-grid-load-storage coordinated dispatch command, the smart microgrid will collect actual operating results from meters, energy storage, photovoltaics, loads, etc., and determine various deviations between the actual operating results and the market application strategy, market clearing results, and pre-prediction results. These deviations will be fed back to the microgrid intelligent agent as learning signals to update the network parameters of the microgrid intelligent agent and iteratively optimize the market application strategy for the next round.

[0038] Furthermore, after the microgrid smart agent is updated and iterated, new autonomous supply capacity indicators, critical load guarantee risks, and market deviation assessment risks are determined. The smart microgrid can switch operating modes based on the new autonomous supply capacity indicators, critical load guarantee risks, market deviation assessment risks, and grid connection point operating status.

[0039] In the aforementioned method for optimizing the operation of the smart microgrid's power market, a state space for the microgrid's intelligent agent is constructed using multi-source power data. Power supply capacity is assessed at each level, and dual boundaries for safety and trading power are defined. The microgrid's intelligent agent generates market bidding strategies within these constraints, and the clearing results are used to implement coordinated dispatching of power generation, grid, load, and storage. Next, learning signals are constructed using the actual operational results after dispatching, and the microgrid's intelligent agent is continuously iterated and optimized. Simultaneously, multiple operating modes are dynamically switched based on system evaluation. Therefore, integrated linkage is achieved in data evaluation, constraint generation, market bidding, resource dispatching, feedback optimization, and operating condition switching. Layered safety constraints reduce the risk of exceeding limits and assessment risks, balancing market trading revenue with power supply reliability in both grid-connected and islanded scenarios, and possessing adaptive and self-optimizing capabilities.

[0040] In one embodiment, determining the autonomous supply security boundary under different scheduling cycles through the state space includes: determining the critical load guarantee rate and energy storage support duration under different scheduling cycles based on the critical load demand associated with the state space and the predicted output of distributed power sources; weighting the normalized energy storage support duration and the determined islanded operation margin, reserve capacity margin, and local supply and demand balance to obtain a comprehensive index; and determining the autonomous supply security boundary based on the comprehensive index.

[0041] Specifically, the critical load guarantee rate can be expressed in the following ways: ;

[0042] in, For the first Critical load guarantee rate for each scheduling period For energy storage, dischargeable power, To allow for the purchase of electricity, Available power released for adjustable loads. For critical load demand.

[0043] The duration that energy storage can support can be expressed in the following ways: ;

[0044] in, For the duration of energy storage, This represents the current energy storage capacity. To store the minimum amount of electricity, Based on basic load demand, Predicting output for distributed power sources.

[0045] Constructing comprehensive indicators It can be expressed in the following ways: ;

[0046] in, The normalized duration of energy storage support. For the operational margin of an isolated island, As a reserve capacity margin, To maintain local supply and demand balance, These are the corresponding weighting coefficients, and the sum of these weighting coefficients is 1. Therefore, after the server constructs a comprehensive index based on multiple autonomous supply capability indicators, the autonomous supply security boundary can be determined based on this comprehensive index.

[0047] In this embodiment, by integrating energy storage support indicators and multiple margins to obtain comprehensive indicators and delineate safety boundaries, the power supply security level of the microgrid can be fully quantified, avoiding the bias of single indicator evaluation, identifying supply and demand risks in advance, and providing a quantitative basis for subsequent dynamic adjustment of tradable power.

[0048] In one embodiment, the above method further includes: initially adjusting the tradable power boundary based on any one of critical load guarantee rate, reserve capacity margin, and comprehensive indicators.

[0049] Specifically, when the comprehensive indicator is lower than the preset safety threshold, or the critical load guarantee rate is lower than the preset guarantee threshold, or the reserve capacity margin is lower than the preset reserve threshold, the range of market tradable power boundary is narrowed; when the comprehensive indicator meets the preset safety threshold, and the critical load guarantee rate and reserve capacity margin meet the corresponding threshold requirements, the tradable power boundary capacity is released.

[0050] In one embodiment, generating a tradable power boundary based on the autonomous supply security boundary includes: determining multiple power parameters based on the autonomous supply security boundary; determining the upper limit of tradable power by means of the maximum discharge power of energy storage and the adjustable load response capacity; obtaining the upper limit of purchasable power by means of the maximum chargeable power of energy storage and the maximum allowable exchange power at the grid connection point; and generating a tradable power boundary by combining the upper limit of tradable power and the upper limit of purchasable power.

[0051] The power parameters include at least the maximum discharge capacity of energy storage, the maximum recharge capacity of energy storage, the adjustable load response capacity, and the maximum allowable exchange capacity at the grid connection point. Negative values ​​within the tradable power boundary indicate the direction of electricity purchase, while positive values ​​indicate the direction of electricity sale.

[0052] Specifically, the upper limit of salable electrical power can be expressed in the following ways: ;

[0053] in, For the first The upper limit of available electricity for sale during each dispatch period To store the maximum discharge power, For adjustable load response capacity, This is the total load forecast value. Reserve power for backup capacity.

[0054] The maximum purchasable power capacity can be expressed in the following ways: ;

[0055] in, For the first The upper limit of the available power capacity for each dispatch period The maximum allowable switching power at the grid connection point. This represents the maximum rechargeable power of the energy storage. Reserve power for backup capacity. Predicting output for distributed power sources.

[0056] therefore, For the first The tradable power boundary for each scheduling period can be represented as follows: ;

[0057] In this embodiment, the power purchase and sale limit is calculated by layering the power supply security boundary, and the upper and lower limits are determined by combining the energy storage charging and discharging capacity, adjustable load and tie line constraints. This can take into account multiple constraints of equipment, power grid and power supply security, and accurately delineate the tradable range.

[0058] In one embodiment, a market application strategy is generated within the tradable power boundary using a microgrid agent as the decision-making core. This includes: processing the state space through the microgrid agent with the tradable power boundary as a constraint to obtain an initial market application action; obtaining an objective function and optimizing parameters based on the initial market application action and the objective function until the target market application action under the optimal solution of the objective function is obtained; and supplementing the trading period information under the target market application action to obtain the market application strategy.

[0059] The market reporting process includes reporting electricity volume, reporting price, transaction direction, and reserve capacity configuration.

[0060] Specifically, the initial market application action of a microgrid smart agent can be represented in the following ways: ;

[0061] in, For the first The initial market reporting action for each scheduling period, In order to declare the electricity volume, For the declared price, For the direction of the transaction, Configured for standby capacity. The requested power consumption meets the requirements. .

[0062] The market application strategy aims to maximize transaction revenue, minimize electricity purchase costs, minimize deviation assessment risks, prioritize critical load protection, and limit energy storage lifespan losses. The objective function is expressed as: ;

[0063] in, For the overall benefit within the scheduling period, To clear electricity prices in the market, For the volume of electricity traded, For operating costs, For deviation assessment costs, Costs related to energy storage lifespan depletion. To mitigate the risks and costs associated with autonomous supply, when the state feature vector of the state space is input into the deep neural network of the microgrid agent, an initial market application action is output, constrained by the tradable power boundary. The server substitutes the initial market application action into the objective function to calculate the comprehensive revenue within the scheduling cycle and optimizes to obtain the target market application action that maximizes the revenue. Finally, business information is supplemented to generate a market application strategy.

[0064] In this embodiment, the application range is rigidly constrained by the tradable power boundary. The microgrid agent processes the state space to generate initial application actions, and the optimal application action is obtained by optimizing the objective function. Finally, business information is supplemented to form a complete application strategy. This approach not only eliminates the power supply and performance risks associated with out-of-boundary transactions but also balances multiple objectives such as revenue, losses, and supply security, achieving quantitative optimal bidding decisions.

[0065] In one embodiment, after market clearing, the penalty weights corresponding to the objective function are adjusted based on the deviations generated by the actual operating results, and the network output is iteratively optimized to perform the next round of new market declaration actions.

[0066] In one embodiment, when the market price is higher than the electricity sales revenue threshold and the autonomous supply security boundary meets the trading conditions, the microgrid agent generates an electricity sales declaration strategy; when the market price is lower than the electricity purchase cost threshold or the local supply and demand gap is greater than the preset gap threshold, the microgrid agent generates an electricity purchase declaration strategy.

[0067] In one embodiment, the policy parameters in the microgrid agent Specifically, it consists of the weight matrices of each hidden layer in the deep policy network (ActorNetwork). and bias vector The composition, whose parameter set is represented as: ;

[0068] in, This represents the total number of layers in the deep policy network. Microgrid state feature vector. As network input, it is mapped to market reporting actions through layer-by-layer forward propagation. , of which The nonlinear forward propagation process of a layer is represented as: ; in, and The first The input and output feature vectors of the hidden layer, when Input time is , This is a non-linear activation function. In the network parameter self-learning update step, the gradient vector... In essence, it is about the composition Weight matrices of each layer and bias vector Calculate the partial derivatives and perform numerical iterations.

[0069] In one embodiment, the method further includes: when executing market-traded electricity, determining power balance constraints based on the grid connection point exchange power, load adjustment power, and energy storage charging power of the smart microgrid; determining energy storage state of charge constraints based on the energy storage charging efficiency, energy storage discharging efficiency, dispatching time interval, and rated energy storage capacity of the smart microgrid; and determining grid connection point exchange power constraints based on the maximum allowable exchange power at the grid connection point of the smart microgrid.

[0070] Among them, the scheduling constraints satisfied by the coordinated scheduling of source, grid, load and storage include at least power balance constraints, energy storage state of charge constraints and grid connection point exchange power constraints.

[0071] Specifically, when executing market-traded electricity volumes, the power balance constraints that smart microgrids must meet can be expressed in the following way: ;

[0072] in, For power exchange at the grid connection point, Adjust power for load, Power for charging energy storage. Energy storage operation meets the following requirements: ;

[0073] in, For energy storage charging efficiency, For energy storage and discharge efficiency, For scheduling time intervals, This represents the rated capacity of the energy storage. Therefore, the energy storage state-of-charge constraints satisfy: ;

[0074] in, This is the limit of the state of charge of energy storage. This represents the upper limit of the energy storage state of charge. The maximum allowable switching power at the grid connection point is given by the following constraints: ;

[0075] In this embodiment, by simultaneously establishing three refined constraints—power balance, energy storage state of charge, and grid connection point power exchange—real-time power supply and demand balance can be guaranteed, ensuring that dispatching instructions strictly conform to the physical limits of equipment and the power grid, and taking into account transaction fulfillment, power supply security, and energy storage operation and maintenance benefits.

[0076] In one embodiment, after performing source-grid-load-storage coordinated scheduling of the smart microgrid, the method further includes: collecting the actual operating results after source-grid-load-storage coordinated scheduling, and iteratively updating the microgrid agent based on the actual operating results and market clearing results; obtaining the various indicator risks and corresponding target thresholds determined after the microgrid agent is updated; and switching the operating mode of the smart microgrid based on the indicator risks and target thresholds.

[0077] The actual operating results include actual distributed power output, actual load demand, actual energy storage state of charge, actual grid connection point switching power, critical load power supply results, and actual operating costs. Risk indicators include at least self-sufficiency capacity indicators, critical load guarantee risks, and market deviation assessment risks. Operating modes include grid-connected trading mode, self-sufficiency priority mode, isolated supply guarantee mode, and restored grid-connected trading mode. Target thresholds include the safety threshold corresponding to self-sufficiency capacity, the risk threshold corresponding to critical load guarantee risks, and the assessment threshold corresponding to market deviation assessment risks.

[0078] Specifically, after the market trading cycle ends, the server compares the actual operating results collected with the market declaration strategy, market clearing results, and pre-prediction results to obtain prediction deviation, trading deviation, operating deviation, and revenue deviation. The above-mentioned deviations and deviation assessment fees are used as learning signals and fed back to the microgrid intelligent agent to update the state weights, risk weights, declaration strategy parameters, energy storage scheduling parameters, and autonomous supply security boundary parameters of the network model, so as to realize the self-learning iteration of the smart microgrid's participation in the power market optimization operation strategy.

[0079] In simple terms, the microgrid intelligent agent update includes safety boundary weights and weights for various sub-indicators. After weight iteration, the calculated autonomous supply capacity index value will change based on the same source-load-storage measured data, and the corresponding safety threshold will be updated simultaneously. Meanwhile, the risk weight coefficient in risk assessment is also a parameter to be updated for the microgrid intelligent agent. After parameter optimization, the standard for assessing the risk of critical load guarantee will change under the same load deficit and energy storage support duration. Furthermore, deviation costs and assessment penalty weights are also core update parameters for the microgrid intelligent agent. After multiple rounds of deviation feedback learning, the assessment of market deviation assessment risk will be re-evaluated. Therefore, the server needs to obtain various indicator risks and corresponding target thresholds after the microgrid intelligent agent update, as well as the grid connection point operating status, to switch the smart microgrid's operating mode.

[0080] Furthermore, when the self-sufficiency supply capacity indicator meets the corresponding safety threshold and the grid connection point is operating normally, the smart microgrid enters the grid-connected trading mode and participates in electricity market transactions according to the tradable power boundary. When the self-sufficiency supply capacity indicator is lower than the corresponding safety threshold, the critical load guarantee risk is higher than the corresponding risk threshold, or the market deviation assessment risk is higher than the corresponding assessment threshold, the smart microgrid enters the self-sufficiency power supply priority mode and reduces the market-declared electricity volume or increases the reserve capacity retention ratio. When the grid connection point is abnormal or the upper-level grid issues an off-grid operation instruction, the smart microgrid enters the islanded supply guarantee mode, prioritizing the power supply to critical loads and suspending or reducing market trading activities that do not meet the safety boundary. When the grid connection point returns to normal and the self-sufficiency supply capacity indicator meets the trading conditions, the smart microgrid enters the restored grid-connected trading mode and regenerates the market declaration strategy.

[0081] In the above embodiments, the intelligent agent is iteratively optimized based on the actual operation results and clearing results after scheduling, updating power supply and assessment risk indicators and judgment thresholds, and automatically switching between grid-connected, islanded, and other operating modes according to the thresholds. This achieves closed-loop adaptive control of scheduling, feedback, and operating condition switching, which can avoid power supply and assessment risks in advance and dynamically balance market revenue and local supply security.

[0082] In one embodiment, the server solves and generates new source-grid-load-storage coordinated scheduling instructions based on the objectives, constraints, and power boundaries corresponding to the switched operating mode.

[0083] In one embodiment, the microgrid agent is iteratively updated based on actual operating results and market clearing results, including: obtaining the microgrid agent's preliminary prediction results; obtaining the prediction deviation based on the actual operating results and the preliminary prediction results; determining the operating deviation between the actual operating conditions and the scheduled operating conditions; obtaining the transaction deviation based on the market declaration strategy and the market clearing results; mapping the penalty costs corresponding to the prediction deviation, operating deviation, and transaction deviation; determining the revenue deviation between the actual revenue and the expected revenue; constructing a learning signal for the microgrid agent through multiple penalty costs and revenue deviations; and performing self-learning updates on the network weights of the microgrid agent using the learning signal.

[0084] Specifically, the prediction deviation obtained from the actual operating results and the preliminary prediction results can be expressed as: ;

[0085] in, For the first Prediction deviation for each scheduling period To provide actual power output for distributed power sources Predicting output for distributed power sources For actual load demand, To predict load demand.

[0086] The trading deviation obtained based on the market reporting strategy and market clearing results is expressed as follows: ;

[0087] in, For trading deviations, For the volume of electricity traded, For the declared electricity volume. Revenue deviation is expressed as: .in, For actual benefits, For expected returns.

[0088] The learning signals for constructing a microgrid agent can be represented in the following ways: ;

[0089] in, For the first The learning signal for each scheduling period, i.e., the environmental reward in reinforcement learning; penalty cost. The cost for deviation assessment is derived from the mapping of transaction deviations; penalty costs. The risk cost of autonomous supply is derived from the mapping of forecast deviation; penalty cost. The energy storage lifetime loss cost is obtained by mapping from the operating deviation. These are the corresponding weighting coefficients. Therefore, by calculating the multi-dimensional deviations of prediction, operation, transaction, and revenue in a hierarchical manner, various deviations are quantified and mapped to corresponding penalty costs. The revenue deviations are then integrated to construct a standardized learning signal for iterative updates of the agent.

[0090] The time difference error of the microgrid agent in the current state action pair is calculated based on the learned signal. :

[0091] in, As a discount factor, For the agent in state The value function evaluation value is obtained. The calculated various deviations and TD errors are transformed into mathematical gradients, and the policy parameters of the microgrid agent are updated through self-learning. ; in, For the first The strategy parameters for the next iteration. For the updated policy parameters, For learning rate, In the state The following actions will be taken in the current reporting and scheduling process. The strategy probability distribution.

[0092] In this embodiment, the aforementioned feedback learning mechanism can comprehensively quantify various losses caused by prediction inaccuracies, scheduling mismatches, and bidding discrepancies, uniformly transforming them into model optimization guidelines to accurately update the state weights, risk weights, strategy parameters, energy storage scheduling parameters, and autonomous supply safety boundary parameters of the microgrid intelligent agent. Therefore, it achieves closed-loop error feedback throughout the entire process of prediction, scheduling, and market transactions, enabling continuous self-learning iteration of market application strategies and source-grid-load-storage coordinated scheduling.

[0093] In one embodiment, such as Figure 3 As shown, a method for smart microgrid agents to participate in the optimized operation of the power market is provided, constrained by the autonomous supply security boundary and using market clearing feedback as a learning signal. The method includes the following steps: S1: Collect multi-source power data and construct the state space of the microgrid agent. S2: Assess the autonomous supply capability under future scheduling cycles. S3: Determine the autonomous supply security boundary based on autonomous supply capability indicators. S4: Generate a tradable power boundary based on the autonomous supply security boundary and grid operation constraints. S5: Generate a market application strategy within the tradable power boundary, using the microgrid agent as the decision-making core. S6: Execute the source-grid-load-storage coordinated scheduling command of the smart microgrid based on the market clearing result. S7: Collect the actual operating results after executing the source-grid-load-storage coordinated scheduling command, and iteratively update the microgrid agent based on the actual operating results and the feedback from the market clearing result as learning signals. S8: Based on the state assessment of the smart microgrid, switch between grid-connected trading, autonomous power supply, islanded supply guarantee, and grid restoration operation modes.

[0094] In the above embodiments, by introducing an autonomous supply security boundary, the local energy security of the smart microgrid is prevented from being reduced due to excessive pursuit of market gains; by constructing a tradable power boundary, the market bidding strategy is constrained by the actual regulation capacity of the power generation, grid, load, and storage, thereby improving the executability of the bidding results; by using a unified modeling system for the microgrid's intelligent agents to model the states of the source side, energy storage side, load side, grid side, market side, and risk side, the decision-making adaptability in complex scenarios is improved; and through a market clearing feedback learning mechanism, the smart microgrid can continuously optimize its operation strategy based on actual transaction results, operational deviations, and revenue deviations, thereby improving the benefits of participating in the electricity market and operational stability.

[0095] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0096] Based on the same inventive concept, this application also provides a smart microgrid power market optimization operation device for implementing the above-mentioned smart microgrid power market optimization operation method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more smart microgrid power market optimization operation device embodiments provided below can be found in the limitations of the smart microgrid power market optimization operation method described above, and will not be repeated here.

[0097] In one embodiment, a power market optimization operation device for a smart microgrid is provided, comprising: a spatial construction and boundary determination module and a strategy determination and scheduling execution module, wherein: The space construction and boundary determination module is used to construct the state space of the microgrid intelligent agent based on multi-source power data; determine the autonomous supply security boundary under different scheduling cycles through the state space; and generate the tradable power boundary based on the autonomous supply security boundary.

[0098] The strategy determination and scheduling execution module is used to generate market declaration strategies within the tradable power boundary with the microgrid intelligent agent as the decision-making core, and execute the coordinated scheduling of source, grid, load and storage of the smart microgrid based on the market clearing results.

[0099] The various modules in the optimized operation of the electricity market in the aforementioned smart microgrid can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0100] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores multi-source power data. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a smart microgrid power market optimization operation method.

[0101] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0102] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0103] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0104] In one embodiment, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the steps in the above method embodiments.

[0105] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0106] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0107] The above embodiments merely illustrate several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of this application's patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for optimizing the operation of a smart microgrid's electricity market, characterized in that, The method is implemented based on microgrid smart agents and includes: Based on multi-source power data, construct the state space of the microgrid intelligent agent; The autonomous supply security boundary under different scheduling cycles is determined through the state space, and the tradable power boundary is generated based on the autonomous supply security boundary. With the microgrid intelligent agent as the decision-making core, a market declaration strategy is generated within the tradable power boundary; Based on the market clearing results, implement coordinated dispatch of power generation, grid, load and storage in smart microgrids.

2. The method according to claim 1, characterized in that, Determining the autonomous supply security boundary under different scheduling periods through the state space includes: Based on the critical load demand associated with the state space and the predicted output of distributed power sources, the critical load guarantee rate and the energy storage support duration under different scheduling cycles are determined. The normalized energy storage support duration, the determined island operation margin, the reserve capacity margin, and the local supply and demand balance are weighted to obtain a comprehensive index. Determine the self-sufficiency security boundary based on the aforementioned comprehensive indicators; The above method also includes: initially adjusting the tradable power boundary based on any one of the critical load guarantee rate, the reserve capacity margin, and the comprehensive index.

3. The method according to claim 1, characterized in that, Based on the aforementioned autonomous supply security boundary, a tradable power boundary is generated, including: Multiple power parameters are determined based on the autonomous supply safety boundary; the power parameters include at least the maximum discharge power of energy storage, the maximum recharge power of energy storage, the adjustable load response capacity, and the maximum allowable switching power at the grid connection point; The upper limit of the saleable power is determined by the maximum dischargeable power of the energy storage and the adjustable load response capacity; The upper limit of the purchasable power is obtained based on the maximum rechargeable power of the energy storage and the maximum allowable exchange power of the grid connection point; By combining the upper limit of the saleable power and the upper limit of the purchaseable power, a tradable power boundary is generated; negative values ​​in the interval of the tradable power boundary indicate the direction of power purchase, and positive values ​​in the interval indicate the direction of power sale.

4. The method according to claim 1, characterized in that, Using a microgrid agent as the decision-making core, a market application strategy is generated within the tradable power boundary, including: Using the tradable power boundary as a constraint, the state space is processed by a microgrid agent to obtain an initial market application action; the market application action includes the applied electricity amount, applied price, trading direction, and reserve capacity configuration; Obtain the objective function, and optimize the parameters based on the initial market declaration action and the objective function until the target market declaration action under the optimal solution of the objective function is obtained; Supplementing the trading session information under the target market reporting action yields the market reporting strategy.

5. The method according to any one of claims 1 to 4, characterized in that, The scheduling constraints satisfied by the source-grid-load-storage coordinated scheduling include at least power balance constraints, energy storage state of charge constraints, and grid connection point switching power constraints; the method further includes: When executing market-traded electricity volumes, power balance constraints are determined based on the grid connection point exchange power, load adjustment power, and energy storage charging power of the smart microgrid. Based on the energy storage charging efficiency, energy storage discharging efficiency, dispatching time interval, and rated energy storage capacity of the smart microgrid, determine the energy storage state of charge constraints; The grid connection point exchange power constraint is determined based on the maximum allowable exchange power at the grid connection point of the smart microgrid.

6. The method according to claim 1, characterized in that, After performing source-grid-load-storage coordinated scheduling of the smart microgrid, the method further includes: The actual operation results after the coordinated scheduling of source, grid, load and storage are collected, and the microgrid agent is iteratively updated based on the actual operation results and the market clearing results; The risk of various indicators and corresponding target thresholds determined after the microgrid intelligent agent is updated are obtained; the risk of indicators includes at least the self-sufficiency capacity indicator, the critical load guarantee risk and the market deviation assessment risk. Based on the risk indicators and the target threshold, the operation mode of the smart microgrid is switched.

7. The method according to claim 6, characterized in that, Based on the actual operating results and the market clearing results, the microgrid intelligent agent is iteratively updated, including: Obtain the preliminary prediction results of the microgrid intelligent agent, and based on the actual operating results and the preliminary prediction results, obtain the prediction deviation and determine the operating deviation between the actual operating conditions and the scheduling plan operating conditions; Based on the market reporting strategy and the market clearing results, the trading deviation is obtained; Map the penalty costs corresponding to the prediction deviation, the operational deviation, and the transaction deviation; Determine the profit deviation between actual and expected profits; A learning signal for the microgrid agent is constructed using multiple penalty costs and revenue deviations, and the network weights of the microgrid agent are updated through self-learning using the learning signal.

8. A smart microgrid power market optimization operation device, characterized in that, The device includes: The space construction and boundary determination module is used to construct the state space of the microgrid intelligent agent based on multi-source power data; determine the autonomous supply security boundary under different scheduling cycles through the state space; and generate the tradable power boundary based on the autonomous supply security boundary. The strategy determination and scheduling execution module is used to generate market declaration strategies within the tradable power boundary with the microgrid intelligent agent as the decision-making core, and to execute the source-grid-load-storage coordinated scheduling of the smart microgrid based on the market clearing results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images