Household power output determination method, apparatus, and electronic device within a community

Abstract

This invention discloses a method, apparatus, and electronic device for determining household power output within a community. The method includes: acquiring household electricity consumption data for multiple households within a predetermined community; performing multi-time-scale prediction operations on the household electricity consumption data to obtain multi-time-scale prediction data; performing a weighted fusion operation on the multi-time-scale prediction data to obtain fused prediction data; determining the predicted energy difference based on the fused prediction data; adjusting the community's shared electricity price based on the community's shared energy storage state of charge (SOC) value, according to dynamic electricity price adjustment rules, when multiple households are not uniformly participating in the sharing system; and determining the shared power value based on the predicted energy difference and the household energy storage SOC value, using the community shared energy storage SOC value, for control purposes. This invention solves the technical problems of low user participation and poor scheduling performance in related technologies for community energy storage.

Description

Technical Field

[0001] This invention relates to the field of data processing, and more specifically, to a method, apparatus, and electronic device for determining household power output within a community. Background Technology

[0002] Among related technologies, distributed photovoltaic (PV) systems and residential energy storage (ES) systems have been widely applied. These new energy systems have become indispensable core equipment for residential energy transition. These systems can not only reduce residential electricity costs but also help the power grid to smooth out peak and off-peak periods. However, existing residential PV and energy storage systems still have many shortcomings. Existing community-level energy storage technologies are prone to low user participation, system dispatch delays, and poor system stability. These shortcomings limit energy utilization and the promotion of new energy systems.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides a method, apparatus, and electronic device for determining household power output within a community, in order to at least address the technical problems of low user participation and poor scheduling performance in community energy storage technologies.

[0005] According to one aspect of the present invention, a method for determining household power output within a community is provided, comprising: acquiring household electricity consumption data corresponding to multiple households within a predetermined community, wherein the household electricity consumption data includes historical power data, environmental data, and time characteristics; performing a multi-time-scale prediction operation on the household electricity consumption data to obtain multi-time-scale prediction data, wherein the multi-time-scale prediction data includes photovoltaic power output prediction values ​​and electricity load prediction values; performing a weighted fusion operation on the multi-time-scale prediction data according to a preset dynamic weighted fusion strategy to obtain fused prediction data for the multiple households at future times; and determining the predicted energy difference corresponding to the multiple households based on the fused prediction data for the multiple households at future times, wherein the prediction... The energy difference is the difference between the predicted output of the integrated photovoltaic system and the predicted electricity load. When multiple households are not evenly participating in the sharing phase, the shared electricity price within the community is adjusted according to the dynamic electricity price adjustment rules and the community shared energy storage state-of-charge value to maintain the target stability condition corresponding to the predetermined community, ensuring that the multiple households are in a state of selective participation in the sharing phase. The target stability condition includes that the additional revenue from sharing is greater than or equal to the household sharing cost. When multiple households are all participating in the sharing phase, the shared power value corresponding to each of the multiple households is determined based on the community shared energy storage state-of-charge value, the predicted energy difference corresponding to each household, and the household energy storage state-of-charge value, for control purposes.

[0006] Optionally, based on the dynamic electricity price adjustment rules, the shared electricity price within the community is adjusted according to the state of charge value of the shared energy storage, in order to maintain the target stability condition corresponding to the predetermined community, so that the multiple households are in a state of choosing to participate in sharing. This includes: determining multiple game payoff matrices under different strategy combinations, wherein the game payoff matrix is ​​a matrix representing the strategy combination corresponding to the sharing participation result of one household and the sharing participation result chosen by another household; determining the payoff value under different strategy combinations based on the multiple game payoff matrices under different strategy combinations; and determining the payoff value corresponding to each of the multiple households based on the payoff value under different strategy combinations. The expected and average expected returns of participating in the sharing strategy are selected; based on the expected and average expected returns of participating in the sharing strategy corresponding to each of the multiple households, a replication dynamic equation is established for each of the multiple households, wherein the replication dynamic equation represents the rate of evolution of the probability of a household choosing to participate in the sharing strategy over time; fixed points of multiple replication dynamic equations are determined, and stability analysis is performed on each fixed point to obtain the stability determination result of each fixed point. When any household's fixed point is determined to be unstable, the community-wide shared electricity price is adjusted according to the dynamic electricity price adjustment rules and the community-wide shared energy storage state of charge value until all stability determination results are stable.

[0007] Optionally, determining multiple game payoff matrices under different strategy combinations includes: obtaining the basic payoffs, shared additional payoffs, and family sharing costs corresponding to the multiple households respectively; determining the payoff functions corresponding to the multiple households respectively based on the basic payoffs, shared additional payoffs, and family sharing costs; setting a set of strategies for participating in and not participating in sharing for each household according to the evolutionary game model, and constructing game payoff matrices under different strategy combinations in the strategy set based on the payoff functions corresponding to the corresponding households.

[0008] Optionally, obtaining the basic income, shared additional income, and household sharing costs corresponding to each of the multiple households includes: obtaining community electricity data, and electricity transaction data and energy loss data corresponding to each of the multiple households within the predetermined community. The community electricity data includes the community shared electricity sales price, the community shared electricity purchase price, and the shared subsidy standard. The electricity transaction data includes the average daily photovoltaic surplus electricity, the average daily electricity purchased from the grid, the grid purchase price, and the grid sales price. The energy loss data includes energy transmission loss costs, energy storage battery unit cycle loss costs, and routing equipment depreciation costs. Based on the electricity transaction data, the basic income of each of the multiple households in independent operation mode is determined. Based on the community electricity data and the predicted energy difference corresponding to each of the multiple households, the household shared electricity amount corresponding to each of the multiple households is determined. Based on the energy loss data and the household shared electricity amount corresponding to each of the multiple households, the household sharing cost corresponding to each of the multiple households is determined. Based on the household shared electricity amount corresponding to each of the multiple households, the shared additional income corresponding to each of the multiple households is determined.

[0009] Optionally, after determining the shared power value corresponding to each of the multiple households based on the community-shared energy storage state of charge value, the predicted energy difference corresponding to each of the multiple households, and the household energy storage state of charge value, the method further includes: obtaining the control target value and actual operating value of the energy storage device corresponding to each of the multiple households, determining the control error corresponding to each of the multiple households, wherein the control target value is the corresponding shared power value; and performing proportional, fractional integral, and fractional derivative operations on the corresponding control error based on the fractional cooperative control law defined by Caputo fractional calculus to generate control instructions corresponding to each of the multiple households, wherein the corresponding control instructions are used to adjust the charging and discharging power of the corresponding energy storage device.

[0010] Optionally, before generating control commands corresponding to the multiple families by performing proportional, fractional integral, and fractional derivative operations on the corresponding control error according to the fractional cooperative control law defined by Caputo fractional calculus, the method further includes: retrieving a multi-objective optimization function, wherein the multi-objective optimization function aims to jointly minimize overshoot, settling time, and steady-state error; and determining the target parameters in the fractional cooperative control law by using the optimal combination of control parameters obtained through global optimization within a preset parameter constraint range using a particle swarm optimization algorithm based on the multi-objective optimization function, wherein the target parameters include proportional coefficient, integral coefficient, derivative coefficient, integral order, and derivative order.

[0011] Optionally, when the multi-timescale prediction operation includes a first timescale prediction operation and a second timescale prediction operation, the first timescale prediction operation is performed on the household electricity consumption data according to the kernel extreme learning machine model to obtain first timescale prediction data, and the second timescale prediction operation is performed on the household electricity consumption data according to the long short-term memory network model with fused attention mechanism to obtain second timescale prediction data, wherein the first timescale is smaller than the second timescale.

[0012] According to one aspect of the present invention, a device for determining household power output within a community is provided, comprising: an acquisition module for acquiring household electricity consumption data corresponding to multiple households within a predetermined community, wherein the household electricity consumption data includes historical power data, environmental data, and time characteristics; a prediction module for performing multi-time-scale prediction operations on the household electricity consumption data to obtain multi-time-scale prediction data, wherein the multi-time-scale prediction data includes photovoltaic power output prediction values ​​and electricity load prediction values; a fusion module for performing a weighted fusion operation on the multi-time-scale prediction data according to a preset dynamic weighted fusion strategy to obtain fused prediction data for multiple households at future times; and a first determination module for determining the predicted energy corresponding to each of the multiple households based on the fused prediction data for the multiple households at future times. The difference, wherein the predicted energy difference is the difference between the predicted value of integrated photovoltaic output and the predicted value of integrated electricity load; the adjustment module, used to adjust the shared electricity price within the community according to the dynamic electricity price adjustment rules and the community shared energy storage state of charge value when multiple households are not evenly participating in the sharing state, so as to maintain the target stability condition corresponding to the predetermined community and keep multiple households in the state of choosing to participate in the sharing, wherein the target stability condition includes that the additional sharing income is greater than or equal to the household sharing cost; the second determination module, used to determine the shared power value corresponding to each of the multiple households when multiple households are all participating in the sharing state, based on the community shared energy storage state of charge value, and the predicted energy difference and household energy storage state of charge value corresponding to each household, so as to control according to the corresponding shared power value.

[0013] According to one aspect of the present invention, an electronic device is provided, comprising: a processor; and a memory for storing processor-executable instructions; wherein the processor is configured to execute the instructions to implement the method for determining household power output within a community as described in any of the preceding claims.

[0014] According to one aspect of the present invention, a computer-readable storage medium is provided, which, when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, enables the electronic device to perform the household power output determination method within a community as described in any of the preceding claims.

[0015] In this embodiment of the invention, household electricity consumption data corresponding to multiple households within a predetermined community are acquired. This household electricity consumption data includes historical power data, environmental data, and time characteristics. A multi-timescale prediction operation is performed on the household electricity consumption data to obtain multi-timescale prediction data, which includes predicted photovoltaic output and predicted electricity load. Based on a preset dynamic weighted fusion strategy, a weighted fusion operation is performed on the multi-timescale prediction data to obtain fused prediction data for each household at future times. Based on the fused prediction data for each household at future times, the predicted energy difference for each household is determined, where the predicted energy difference is the fused photovoltaic output prediction. The difference between the value and the predicted value of the integrated electricity load; when multiple households are not evenly participating in the sharing state, the community-wide sharing price is adjusted according to the dynamic electricity price adjustment rules based on the community-wide shared energy storage state of charge value, in order to maintain the target stability condition corresponding to the predetermined community, so that multiple households are in a state of choosing to participate in sharing. The target stability condition includes that the additional sharing revenue is greater than or equal to the household sharing cost; when multiple households are all participating in the sharing state, the shared power value corresponding to each household is determined based on the community-wide shared energy storage state of charge value, the predicted energy difference corresponding to each household, and the household energy storage state of charge value, so as to control according to the corresponding shared power value. By adopting a multi-timescale forecasting and dynamic electricity price collaborative decision-making approach, and adjusting the community-wide sharing price based on the community-wide shared energy storage state of charge value to maintain the target stability condition, the aim is to guide multiple households to actively participate in sharing and determine their respective shared power values. This achieves the technical effect of collaborative optimization of energy sharing among households within the community, and thus solves the technical problems of low user participation and poor dispatch performance in related technologies for community energy storage. Attached Figure Description

[0016] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0017] Figure 1 This is a flowchart of a method for determining household power output within a community according to an embodiment of the present invention;

[0018] Figure 2 This is an overall operation flowchart of the photovoltaic-storage sharing micro-circulation system provided by an optional method of the present invention;

[0019] Figure 3 This is a flowchart of the multi-scale prediction model provided by an optional method of the present invention;

[0020] Figure 4 This is an evolutionary game shared decision-making flowchart provided by an optional method of the present invention;

[0021] Figure 5 This is a flowchart of the fractional-order cooperative control algorithm provided by an optional method of the present invention;

[0022] Figure 6 This is a structural block diagram of a household power output determination device in a community according to an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "including" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] Example 1

[0026] According to an embodiment of the present invention, an embodiment of a method for determining household power output within a community is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] Figure 1 This is a flowchart of a method for determining household power output within a community according to an embodiment of the present invention, such as... Figure 1 As shown, the method includes the following steps:

[0028] Step S102: Obtain household electricity consumption data for multiple households within the predetermined community. The household electricity consumption data includes historical power data, environmental data, and time characteristics.

[0029] Household electricity data refers to the raw dataset used to describe household energy generation and consumption, including historical power data, environmental data and time characteristics, which can provide multi-dimensional input information for prediction models, such as the photovoltaic output and load power sequence recorded every 15 minutes in the past 24 hours for a household.

[0030] Historical power data refers to the measured records of household photovoltaic power generation and electricity load power in various past periods, which can reflect the historical patterns of household energy behavior, such as the active power value of each hour of the previous day.

[0031] Among them, environmental data refers to natural environmental parameters that are strongly correlated with photovoltaic power generation, which can improve the accuracy of photovoltaic output prediction, such as solar irradiance, ambient temperature, and wind speed.

[0032] Among them, time characteristics refer to the time stamp attributes or time codes of the predicted time, which can capture the time periodicity of household energy consumption behavior, such as hour index, number of days of the week, whether it is a weekday, etc.

[0033] In this step, household electricity consumption data, including historical power data, environmental data, and temporal characteristics, is acquired for multiple households within the designated community. Through this step, the system collects multi-source basic data necessary to support multi-timescale predictions, laying a data foundation for subsequent accurate perception of energy supply and demand fluctuations in each household.

[0034] Step S104: Perform multi-time-scale prediction operation on household electricity consumption data to obtain multi-time-scale prediction data, which includes photovoltaic power output prediction value and electricity load prediction value.

[0035] Among them, multi-timescale prediction operation refers to the process of using at least two prediction models with different time resolutions to process the input data simultaneously. This can take into account both the high-frequency random fluctuations of photovoltaic output and the overall trend of electricity load changes, such as the parallel execution of ultra-short-term 15-minute prediction and short-term 1-hour prediction.

[0036] Among them, multi-timescale prediction data refers to a data set that contains prediction results at different time scales, obtained through multi-timescale prediction operations. It can provide a richer level of prediction information, such as obtaining the photovoltaic output prediction results for the next 15 minutes and the next hour at the same time.

[0037] Among them, the photovoltaic output forecast value refers to the estimated value of the output power of a household photovoltaic power generation system at a specific time in the future. It can identify the period of energy surplus in advance, such as predicting that a household's photovoltaic output will be 3.5kW at 10:15.

[0038] Among them, the electricity load forecast refers to the estimated value of the total power demand of various electrical devices in a household at a specific time in the future. It can predict the period of energy shortage in advance. For example, it can predict that the electricity load of a household at 20:00 will be 4.2kW.

[0039] In this step, multi-timescale forecasting is performed on the acquired household electricity consumption data to obtain multi-timescale forecast data that includes predicted photovoltaic power output and predicted electricity load. This step enables forward-looking perception of future household energy supply and demand at multiple time granularities, providing differentiated information for subsequent fusion to improve forecast accuracy.

[0040] Step S106: Based on the preset dynamic weighted fusion strategy, perform a weighted fusion operation on the multi-time-scale prediction data to obtain the fused prediction data of multiple families at future times.

[0041] Among them, the preset dynamic weighted fusion strategy refers to the fusion rules that are set in advance and automatically adjust the weight of predictions at different time scales based on the time interval between the prediction time and the current time. For example, the fusion rules for the weight of short-term predictions and ultra-short-term predictions can make short-term predictions more reliant on ultra-short-term models that capture fluctuations, and long-term predictions more reliant on short-term models that grasp trends. For example, the closer the prediction point is, the greater the weight of ultra-short-term predictions.

[0042] Among them, the fused prediction data refers to the unified prediction result obtained by applying a dynamic weighted fusion strategy to weight and synthesize the photovoltaic power output prediction value and the electricity load prediction value at different time scales. It can combine the advantages of the two models and obtain a prediction value with higher accuracy than the single-scale model. For example, it can obtain a photovoltaic power output fused value with final determinism.

[0043] In this step, a weighted fusion operation is performed on the multi-timescale forecast data according to a preset dynamic weighted fusion strategy, resulting in fused forecast data for multiple households at various future times. This step eliminates the limitations of single-scale forecasting, improves the overall accuracy of household energy supply and demand forecasting, and provides more reliable data for accurately calculating energy deficits.

[0044] Step S108: Based on the combined prediction data of multiple households at future times, determine the predicted energy difference for each household, where the predicted energy difference is the difference between the combined photovoltaic output prediction value and the combined electricity load prediction value.

[0045] Among them, the predicted energy difference refers to the result obtained by subtracting the predicted value of a household's integrated photovoltaic output from the predicted value of its integrated electricity load. It can directly quantify the degree of energy surplus or shortage of the household at a specific time in the future. For example, a positive difference indicates that there is a predicted surplus of electricity available for sharing, while a negative difference indicates that there is an electricity shortage that needs to be obtained from the outside.

[0046] In this step, based on the converged forecast data for each household at future times, the difference between the predicted converged photovoltaic output and the predicted converged electricity load is calculated to obtain the predicted energy difference for each household. Through this step, the forecast information is transformed into a quantitative indicator of energy surplus / shortage that can be directly used for sharing decisions, providing a basic criterion for determining the sharing power allocation at the community level.

[0047] Step S110: When multiple households are not evenly participating in the sharing state, the sharing price within the community is adjusted according to the dynamic electricity price adjustment rules and the community shared energy storage state value, so as to maintain the target stability condition corresponding to the predetermined community and keep multiple households in the state of choosing to participate in the sharing. The target stability condition includes that the additional sharing income is greater than or equal to the household sharing cost.

[0048] Among them, the participation sharing state refers to the strategy state of a household when it chooses to participate in the community energy sharing strategy in the evolutionary game decision. It can represent whether the household is currently willing to engage in energy sharing within the community. For example, when the household's participation sharing state is yes, its excess electricity can be sent to the community sharing pool.

[0049] Among them, the dynamic electricity price adjustment rule refers to the rule that adjusts the shared electricity sales price and purchase price within the community in reverse according to the real-time changes in the state of charge value of the shared energy storage. It can increase the electricity sales price to encourage households to supply electricity and reduce the electricity purchase price to attract households to take electricity when the energy storage capacity is low, thereby automatically maintaining the economic attractiveness of participating in sharing.

[0050] Among them, the community shared energy storage state of charge value refers to the ratio of the current remaining capacity of the shared energy storage unit set up in the community to its rated capacity. It can reflect the sufficiency of community-level energy storage resources in real time. For example, a value of 0.3 indicates that the energy storage pool needs to be replenished.

[0051] Among them, the shared electricity price within the community refers to the electricity price used to settle energy exchanges between households within the community sharing system. This includes the electricity sales price to shared energy storage and the electricity purchase price from shared energy storage, which can directly affect the shared benefits of each household. For example, the shared electricity sales price can be set higher than the grid purchase price.

[0052] Among them, the target stability condition refers to the economic condition that must be met to enable all households to voluntarily and stably maintain the state of participation in sharing, as derived from evolutionary game analysis. Its content is that the additional benefits of sharing are greater than or equal to the household sharing costs, which can serve as the ultimate basis for electricity price adjustments.

[0053] Among them, the additional benefits of sharing refer to the additional net benefits that households obtain from participating in community energy sharing, which exceed the benefits they would obtain from their individual household independent operation mode. This can economically incentivize households to participate in sharing, such as the price difference benefits generated by the shared electricity price difference and the sharing subsidies provided by the community.

[0054] Among them, the cost of household sharing refers to the total cost that households need to pay in order to participate in community energy sharing. It can be used to measure the cost of participation, such as the cost of energy transmission loss caused by sharing, the depreciation of energy storage batteries due to cycle decay, and the amortization of two-way routing equipment.

[0055] In this step, when not all households are participating in the sharing mechanism, the shared electricity price within the community is adjusted using dynamic electricity price adjustment rules based on the community's shared energy storage state of charge value. This ensures that the target stability condition (extra revenue from sharing ≥ household sharing costs) is continuously met, thereby guiding all households to actively choose to participate in the sharing mechanism. Through this step, the closed-loop adjustment using real-time electricity price feedback dynamically guarantees the economic rationality of participation in sharing, fundamentally solving the problem of low user participation willingness and providing a stable user base for community-level collaboration.

[0056] Step S112: When multiple households are participating in the sharing state, based on the community shared energy storage state of charge value, the predicted energy difference corresponding to each household and the household energy storage state of charge value, determine the shared power value corresponding to each household, and control is carried out based on the corresponding shared power value.

[0057] Among them, the state of charge (SCC) of household energy storage refers to the ratio of the current remaining available electricity of a household's private energy storage device to its rated capacity. It can measure each household's energy storage capacity to cope with energy fluctuations. For example, a household energy storage SCC of 0.7 means that it still has 70% of its energy storage capacity remaining.

[0058] The shared power value refers to the target power calculated for each household to exchange with the community's shared energy storage unit, based on the household's predicted energy difference and the energy storage status of the community and the household. It can serve as a control command to guide the actual charging and discharging operation. For example, a value of +2kW means that the household should supply 2kW of power to the community.

[0059] In this step, when all households are participating in the energy sharing process, the shared energy storage state of charge (SOC) value of the community, the predicted energy difference of each household, and their respective household energy storage SOC values ​​are integrated to calculate a customized shared power value for each household, and energy control is performed accordingly. Through this step, while ensuring that all members have the same economic willingness, the community's energy allocation and collaborative optimization are achieved, thereby improving the overall photovoltaic absorption rate and energy storage utilization rate, and enhancing the system's operational economy and stability.

[0060] Through steps S102-S112 above, household electricity consumption data for multiple households within the predetermined community are obtained. This household electricity consumption data includes historical power data, environmental data, and time characteristics. Multi-timescale prediction operations are performed on the household electricity consumption data to obtain multi-timescale prediction data, which includes predicted photovoltaic output and predicted electricity load. Based on a preset dynamic weighted fusion strategy, a weighted fusion operation is performed on the multi-timescale prediction data to obtain fused prediction data for each household at future times. Based on the fused prediction data for each household at future times, the predicted energy difference for each household is determined, where the predicted energy difference is the fused photovoltaic output... The difference between the predicted power load and the predicted combined electricity load; when multiple households are unevenly participating in energy sharing, the shared electricity price within the community is adjusted according to the dynamic electricity price adjustment rules based on the shared energy storage state-of-charge value of the community, in order to maintain the target stability condition corresponding to the predetermined community, so that multiple households are in a state of choosing to participate in energy sharing. The target stability condition includes that the additional revenue from sharing is greater than or equal to the household's sharing cost. When multiple households are all participating in energy sharing, the shared power value corresponding to each household is determined based on the shared energy storage state-of-charge value of the community, the predicted energy difference corresponding to each household, and the household's energy storage state-of-charge value, so as to control the energy sharing based on the corresponding shared power value. By adopting a multi-timescale forecasting and dynamic electricity price collaborative decision-making approach, and adjusting the shared electricity price within the community based on the shared energy storage state-of-charge value of the community to maintain the target stability condition, the aim is to guide multiple households to actively participate in energy sharing and determine their respective shared power values. This achieves the technical effect of collaborative optimization of energy sharing among households within the community, and thus solves the technical problems of low user participation and poor dispatch performance in related technologies for community energy storage.

[0061] As an optional embodiment, based on dynamic electricity price adjustment rules, the shared electricity price within the community is adjusted according to the state of charge value of the community-shared energy storage to maintain the target stability condition corresponding to the predetermined community, so that the multiple households are in a state of choosing to participate in sharing. This includes: determining multiple game payoff matrices under different strategy combinations, wherein the game payoff matrix is ​​a matrix representing the strategy combination corresponding to the sharing participation result of one household and the sharing participation result chosen by another household; determining the payoff value under different strategy combinations based on the multiple game payoff matrices under different strategy combinations; and determining the sharing price with the multiple households based on the payoff value under different strategy combinations. The expected and average expected returns of each household choosing to participate in the sharing strategy are determined. Based on these expected and average expected returns, a replication dynamic equation is established for each household, where the replication dynamic equation represents the rate of change of the probability of a household choosing to participate in the sharing strategy over time. Fixed points of the replication dynamic equations are determined, and stability analysis is performed on each fixed point to obtain the stability determination result. When any household's fixed point is determined to be unstable, the shared electricity price within the community is adjusted according to the dynamic electricity price adjustment rules and the state of charge value of the community shared energy storage, until all stability determination results are stable.

[0062] Among them, the game payoff matrix refers to a data structure that records the payoffs of two families under different strategy combinations in matrix form. It can clearly show the quantitative correspondence between strategy choice and payoff. For example, when family A chooses to participate in sharing and family B chooses not to participate, the corresponding position in the matrix gives the payoff value of A and B respectively.

[0063] The payoff value under different strategy combinations refers to the specific payoff value extracted from the game payoff matrix corresponding to a specific strategy combination. It can provide the basic input for calculating the expected payoff. For example, the payoff of family A under the participation-non-participation combination can be read from the matrix as 3.2 yuan.

[0064] The expected return of choosing to participate in the sharing strategy refers to the weighted average return that a family can obtain by choosing to participate in sharing, considering that another family may choose different strategies with a certain probability. It can reflect the overall economic attractiveness of participating in sharing. For example, if the other party has a 60% probability of participating, the expected return of our participation is 0.6 × the return of both parties participating plus 0.4 × the return of our participation if the other party does not participate.

[0065] Among them, the average expected return refers to the overall weighted expected return of a family under its own mixed strategy, taking into account the probability of choosing to participate or not participate. It can serve as a benchmark for comparing strategy evolution.

[0066] Among them, the replication dynamic equation refers to the differential equation describing the rate of change of the probability of a family choosing to participate in a sharing strategy over time. It can characterize the dynamic evolution trajectory of group strategies. Its core idea is that the higher the benefit of a strategy, the faster its proportion grows in the group.

[0067] Among them, the fixed point refers to the point in the replication dynamic equation where the rate of change of the policy probability is zero, that is, the policy probability value when the system is in equilibrium. It can identify the candidate states where the policy may stabilize, such as a probability of 0, 1 or some intermediate value.

[0068] Among them, stability analysis refers to the analytical process of determining whether a system can automatically recover to a fixed point after being subjected to a small disturbance. It can distinguish between a truly stable strategy that can be maintained in the long term and a temporary equilibrium that cannot withstand fluctuations.

[0069] Among them, the stability determination result refers to the conclusion of stability or instability given for each fixed point after stability analysis, which can directly indicate which fixed points correspond to the actual maintainable group strategy state.

[0070] In this embodiment, a complete evolutionary game theory analysis process is introduced to address the specific implementation of dynamic electricity price adjustment rules. First, a payoff matrix is ​​constructed for each pair of households, and the payoff values ​​under each strategy combination are extracted. Then, the expected and average expected payoffs for each household choosing to participate in the shared pricing are calculated. Based on this, their respective replication dynamic equations are established. By solving for fixed points and assessing stability one by one, the shared electricity price within the community is adjusted when an unstable household is identified, until the stability assessment results for all households are stable. This approach bases the decision of when and how to adjust electricity prices on a rigorous mathematical analysis of the group's evolutionary dynamics, rather than relying on preset fixed rules or human experience. This gives electricity price adjustments a clear target, eliminating unstable equilibria. Every electricity price adjustment is based on evidence, avoiding imbalances in payoff distribution caused by blind price adjustments or inappropriate price increases. Furthermore, since evolutionary game theory does not rely on the assumption of perfect rationality, each household can start from any initial intention and, through continuous imitation and learning of efficient strategies, eventually spontaneously and stably converge to a state where all members participate in the shared pricing. This self-organizing mechanism significantly reduces the cost of manual intervention in community operation and management, and theoretically guarantees the ability to resist disturbances after strategy convergence. Even if individual households temporarily deviate from the participation state due to temporary factors, the system can recover on its own under stable conditions, fundamentally solving the technical problem of large fluctuations in user participation willingness and the inability to form lasting cooperation in traditional community energy storage solutions.

[0071] As an optional embodiment, determining multiple game payoff matrices under different strategy combinations includes: obtaining the basic income, shared additional income, and family sharing costs corresponding to the multiple households respectively; determining the payoff function corresponding to the multiple households respectively based on the basic income, shared additional income, and family sharing costs corresponding to the multiple households respectively; setting a strategy set for each household to participate in sharing and not participate in sharing based on the evolutionary game model, and constructing game payoff matrices under different strategy combinations in the strategy set based on the payoff function corresponding to the corresponding household.

[0072] Among them, the basic income refers to the net income of a household operating independently without participating in any community sharing model. It can serve as a benchmark for measuring whether participating in sharing is worthwhile. It consists of the surplus photovoltaic revenue fed into the grid minus the cost of purchasing electricity from the grid.

[0073] Among them, the additional benefits of sharing refer to the net additional benefits that households receive beyond the basic benefits from participating in community energy sharing. This can directly incentivize households to join the sharing effort, and includes benefits from the electricity price difference and the sharing subsidy.

[0074] Among them, household sharing costs refer to the additional costs that households need to bear to participate in sharing. They can be used to measure the economic threshold for participating in sharing, including energy transmission loss costs, energy storage battery cycle loss costs, and routing equipment depreciation costs.

[0075] The payoff function is a mathematical expression that maps a household’s strategy choice to its overall payoff. It consists of a combination of basic payoff, shared additional payoff, and shared costs, and can quantify the final economic gain under any strategy.

[0076] The strategy set refers to the set of optional action plans set for each family in the evolutionary game model, including two strategies: participating in sharing and not participating in sharing, which can define the alternative space for family decision-making.

[0077] Among them, evolutionary game theory models refer to mathematical analysis frameworks that study how boundedly rational individuals continuously adjust their strategies through imitation and learning in repeated interactions, eventually achieving group equilibrium. They can describe the dynamic evolutionary process of strategy choices in community households and do not rely on the assumption of perfect rationality.

[0078] In this embodiment, the specific path for constructing the game payoff matrix is ​​clarified: first, the basic revenue of each household operating independently is calculated; then, the additional revenue and cost of sharing after participation are calculated by combining the predicted energy difference and the community electricity price; these three factors are combined to form a unique revenue function for each household; and finally, the revenue values ​​of each strategy combination in the game matrix are filled in based on this function. In this way, the energy surplus / shortage state at the household physical level is completely transformed into a game payoff matrix at the economic level, avoiding the one-size-fits-all bias caused by using uniform preset revenue parameters. Since each household's photovoltaic installed capacity, electricity consumption habits, and energy storage configuration are different, their surplus periods and shortage degrees are also quite different. The payoff matrix constructed using personalized revenue functions can truly reflect the differentiated revenue structure and cost-bearing of each household in the process of participating in sharing, significantly improving the applicability and accuracy of subsequent evolutionary analysis for heterogeneous households. At the same time, the revenue function consists of the sum of basic revenue, additional sharing revenue, and sharing costs, with a clear structure that can be traced item by item. This facilitates modular updates when the system expands or electricity price policies change, without needing to redefine the overall game rules, enhancing the scalability and engineering practicality of the solution.

[0079] As an optional embodiment, obtaining basic revenue corresponding to multiple households, sharing additional revenue, and household sharing costs includes: obtaining community electricity data, as well as electricity transaction data and energy loss data corresponding to multiple households within the predetermined community. The community electricity data includes the community shared electricity sales price, the community shared electricity purchase price, and the shared subsidy standard; the electricity transaction data includes the average daily photovoltaic surplus electricity, the average daily electricity purchased from the grid, the grid purchase price, and the grid sales price; and the energy loss data includes energy transmission loss costs, the unit cycle loss cost of energy storage batteries, and the depreciation cost of routing equipment. Based on the electricity transaction data, the basic revenue of each household in independent operation mode is determined; based on the community electricity data and the predicted energy difference corresponding to each household, the household shared electricity for each household is determined; based on the energy loss data and household shared electricity for each household, the household sharing costs for each household are determined; and based on the household shared electricity for each household, the shared additional revenue for each household is determined.

[0080] Among them, community electricity data refers to the set of electricity price and subsidy parameters on which the community sharing system operates, including the community shared electricity sales price, shared electricity purchase price and shared subsidy standards, which can determine the economic settlement rules for energy sharing.

[0081] Among them, electricity trading data refers to the statistical data of energy exchange between households and the power grid in the independent operation mode, including the average daily surplus of photovoltaic power, the average daily electricity purchased from the power grid, the power grid purchase price and the power grid sales price, which can reflect the basic energy income and expenditure of households.

[0082] Among them, energy loss data refers to the set of various additional loss parameters generated during the process of households participating in sharing, including energy transmission loss cost, energy storage battery unit cycle loss cost and routing equipment depreciation cost, which can quantify the equipment and energy costs brought about by sharing behavior.

[0083] The independent operation mode refers to a single-household operation mode in which the household does not exchange energy with other households in the community and only interacts with the power grid. This mode can be used as a reference scenario for calculating basic benefits.

[0084] Among them, household shared electricity refers to the planned exchange of electricity between households and community shared energy storage units, which can serve as the core physical quantity for settling shared benefits and costs.

[0085] In this embodiment, after acquiring three types of data—community electricity data, electricity transaction data, and energy loss data—the revenue composition is calculated along three paths: basic revenue is calculated based on electricity transaction data; household shared electricity is calculated based on community electricity data and the predicted energy difference; sharing costs are calculated based on energy loss data and shared electricity; and additional sharing revenue is calculated based on shared electricity. This approach clearly breaks down the data sources and calculation paths required for revenue calculation layer by layer. Each type of data corresponds to a component of the revenue function, making the entire revenue calculation process transparent and traceable. The clear division of data sources ensures a complete closed loop in the calculation chain: the predicted energy difference determines the shared electricity; the shared electricity, combined with electricity price data, determines the additional sharing revenue; the loss data determines the sharing cost; and the basic revenue is determined by independent operating parameters. There is no redundancy or omission among the four components. This clear data-driven architecture not only improves the automation of game payoff matrix construction, but also provides a standardized template for system deployment at different community scales: simply collect local weather, electricity prices and equipment parameters according to the corresponding dimensions, and a payoff function adapted to the local conditions can be automatically generated, which greatly reduces the workload of customized development and parameter verification when the solution is implemented, and improves the universality and portability of the technical solution.

[0086] As an optional embodiment, after determining the shared power value corresponding to each of the multiple households based on the shared energy storage state of charge value of the community, the predicted energy difference corresponding to each of the multiple households, and the energy storage state of charge value of each household, the method further includes: obtaining the control target value and actual operating value of the energy storage device corresponding to each of the multiple households, determining the control error corresponding to each of the multiple households, wherein the control target value is the corresponding shared power value; and performing proportional, fractional integral, and fractional derivative operations on the corresponding control error based on the fractional cooperative control law defined by Caputo fractional calculus to generate control instructions corresponding to each of the multiple households, wherein the corresponding control instructions are used to adjust the charging and discharging power of the corresponding energy storage device.

[0087] The control target value refers to the reference value of the operating state that the energy storage device is expected to achieve. Here, it is equal to the shared power value of the corresponding household and can be used as the setting input for the control closed loop.

[0088] Among them, the actual operating value refers to the measured value of the current true operating status of the energy storage device, which can reflect the current actual working condition of the controlled object. The difference between the actual operating value and the control target value is the control error.

[0089] Among them, control error refers to the deviation between the control target value and the actual operating value. It can be used as the driving signal for the controller to calculate the adjustment amount. The larger the error, the greater the adjustment force required.

[0090] Among them, Caputo fractional calculus refers to fractional integral and differential operations given in the form of Caputo's definition. The order can be non-integer and can accurately characterize physical processes with memory effects and hereditary characteristics. When the order is an integer, it degenerates into classical calculus.

[0091] Among them, the fractional-order cooperative control law refers to a control algorithm based on Caputo fractional-order calculus that includes proportional, fractional-order integral, and fractional-order derivative terms. It can adapt to the nonlinear and time-delay characteristics of the photovoltaic storage system by adjusting the integral and derivative orders, and has higher degrees of freedom and control accuracy compared to traditional integer-order proportional-integral-derivative control.

[0092] Among them, the control command refers to the specific command value calculated by the fractional-order cooperative control law based on the control error, which is used to adjust the charging and discharging power of the energy storage device and can directly act on the energy storage converter to achieve power regulation.

[0093] In this embodiment, after determining the shared power value for each household, a closed-loop control mechanism is further introduced: the shared power value is set as the control target, compared with the actual operating value of the energy storage device to obtain the control error, and then processed by a fractional-order cooperative control law to generate precise charging and discharging control commands. This method establishes a closed-loop feedback bridge between the ideal power allocation scheme given by the game theory decision and the actual physical system. This bridge is crucial: the shared power reference value given by the game theory decision layer is a static ideal value. If it is executed directly in an open-loop manner without closed-loop control, random fluctuations in photovoltaic output, intermittent fluctuations in electricity load, and the dynamic response delay of the energy storage device itself will all cause the actual power to continuously deviate from the reference value, resulting in error accumulation. This can range from affecting the accuracy of income settlement for each household to causing short-term fluctuations in system voltage and frequency. Fractional-order cooperative control achieves rapid response to instantaneous deviations through proportional terms, eliminates historical accumulated errors through fractional-order integral terms, and suppresses overshoot trends through fractional-order derivative terms. Furthermore, its non-integer order adapts to the memory effect and time delay characteristics of energy storage batteries, resulting in smooth and precise control outputs and avoiding the overshoot oscillations and slow convergence problems common in integer-order control. Consequently, the system can stably track shared power commands under various disturbances, ensuring that the ideal energy allocation scheme optimized at the game-theoretic layer is faithfully executed at the actual physical level, achieving seamless integration of decision-making and execution.

[0094] As an optional embodiment, before generating control commands corresponding to multiple households by performing proportional, fractional integral, and fractional derivative operations on the corresponding control error according to the fractional cooperative control law defined based on Caputo's fractional calculus, the method further includes: retrieving a multi-objective optimization function, wherein the multi-objective optimization function aims to jointly minimize overshoot, settling time, and steady-state error; and determining the target parameters in the fractional cooperative control law by using the optimal combination of control parameters obtained through global optimization within a preset parameter constraint range using a particle swarm optimization algorithm based on the multi-objective optimization function, wherein the target parameters include proportional coefficient, integral coefficient, derivative coefficient, integral order, and derivative order.

[0095] Among them, the multi-objective optimization function refers to a comprehensive optimization function that aims to minimize three control quality indicators at the same time: overshoot, settling time, and steady-state error. It can avoid sacrificing other performance aspects in pursuit of optimization of a single indicator. For example, optimizing overshoot alone may lead to an excessively long settling time.

[0096] Overshoot refers to the percentage of the maximum magnitude by which the system output exceeds the target value during the response process. It can measure the degree of transient impact of the control system. The smaller the overshoot, the more stable the system.

[0097] The settling time refers to the time required from the occurrence of a disturbance to the system output re-entering and maintaining within the allowable error band. It measures the system response speed; the shorter the settling time, the faster the system recovers.

[0098] Steady-state error refers to the continuous deviation between the output value and the target value after the system stabilizes. It can measure the final control accuracy of the system. The smaller the steady-state error, the higher the accuracy.

[0099] Among them, the particle swarm optimization algorithm refers to a swarm intelligence optimization algorithm that simulates the foraging behavior of bird flocks. It gradually approaches the global optimum through the cooperation and information sharing of individual particles in the search space. It can handle multi-parameter optimization problems without gradient information and achieve efficient parameter tuning.

[0100] Among them, the parameter constraint range refers to the upper and lower bounds of the allowed values ​​of each control parameter, which can limit the optimization search to a physically realizable and safe and reasonable range, and avoid the algorithm giving extreme parameter values ​​that have no practical meaning.

[0101] The optimal control parameter combination refers to a set of parameter values ​​that achieve the best comprehensive control quality index after global optimization under a multi-objective optimization function using the particle swarm optimization algorithm. These values ​​include the proportional coefficient, integral coefficient, derivative coefficient, integral order, and derivative order.

[0102] In this embodiment, before generating control commands using a fractional-order cooperative control law, a particle swarm optimization algorithm is first employed to perform offline global optimization of five control parameters. The goal is to jointly minimize three quality indicators: overshoot, settling time, and steady-state error. Within preset constraints, the optimal parameter combination is searched and embedded into the controller. This method transforms parameter tuning from a traditional approach relying on manual experience and trial-and-error to an automated, systematic multi-objective optimization process. Fractional-order cooperative control laws have up to five parameter dimensions, with complex coupling relationships between them. Manual tuning is not only extremely time-consuming but also prone to falling into locally optimal rather than globally optimal configurations, leading to significant deterioration in control quality under certain operating conditions. The particle swarm optimization algorithm, through group collaboration and information sharing mechanisms, can efficiently explore in a high-dimensional parameter space. Simultaneously, the multi-objective optimization function ensures that the optimal parameter combination achieves a balance across all control quality dimensions, avoiding excessively reducing overshoot to prolong settling time or amplifying steady-state error by solely pursuing rapid response. After offline optimization, the controller is already in the best preset state before being put into online operation. When the actual photovoltaic and energy storage system faces complex disturbances such as photovoltaic sudden drop and load sudden increase, it can respond quickly with preset optimal parameters, achieving a comprehensive control effect of fast response, small overshoot and high precision, which greatly improves the robustness of the system and the consistency of control quality under complex working conditions.

[0103] As an optional embodiment, when the multi-timescale prediction operation includes a first timescale prediction operation and a second timescale prediction operation, the first timescale prediction operation is performed on the household electricity consumption data according to the kernel extreme learning machine model to obtain the first timescale prediction data, and the second timescale prediction operation is performed on the household electricity consumption data according to the long short-term memory network model with fused attention mechanism to obtain the second timescale prediction data, wherein the first timescale is smaller than the second timescale.

[0104] Among them, the first timescale prediction operation refers to the ultra-short-term prediction operation performed at a finer time granularity using a kernel extreme learning machine model, which can quickly capture high-frequency random fluctuations in photovoltaic power output and electricity load.

[0105] Among them, the kernel extreme learning machine model refers to a single hidden layer feedforward neural network that introduces a kernel function to replace random mapping. It can solve the problem of blindness in the traditional extreme learning machine that requires a preset number of hidden layer nodes, while retaining the advantages of fast training speed and strong generalization ability. It is suitable for handling high-frequency time series prediction tasks that require fast response.

[0106] Among them, the first time scale prediction data refers to the prediction data obtained through the first time scale prediction operation with a relatively close sampling interval. It can reflect the details of power fluctuations in a short period of time and provide high-frequency correction information for dynamic weighted fusion.

[0107] The second timescale prediction operation refers to the short-term prediction operation performed at a coarser time granularity using a long short-term memory network model that incorporates attention mechanisms. This operation can capture the overall trend of power changes and the daily peak-valley patterns.

[0108] Among them, the long short-term memory network model that integrates attention mechanism refers to a time series prediction model that introduces attention mechanism on the basis of long short-term memory network. It can adaptively focus on key periods in the historical sequence that have a greater impact on the current prediction through attention weight. Combined with the long-term dependency capture ability of long short-term memory network, it can achieve accurate grasp of trend patterns.

[0109] Among them, the second time scale prediction data refers to the prediction data obtained through the second time scale prediction operation with a relatively sparse sampling interval. It can reflect the overall trend and time-specific characteristics of power over a longer period of time, and provide a trend benchmark for dynamic weighted fusion.

[0110] This embodiment clarifies the specific implementation of multi-timescale prediction operations: a kernel-based Extreme Learning Machine (KLM) is used to perform short-term first-timescale predictions, while a Long Short-Term Memory (LSTM) network incorporating an attention mechanism is used to perform long-term second-timescale predictions. The two complement each other in terms of temporal granularity and model capabilities. In this approach, high-frequency random fluctuations are tracked in real-time by the KLM model. This model trains extremely quickly and can adapt to drastic jumps in photovoltaic power output caused by factors such as cloud drift within minutes, providing timely and accurate fluctuation correction information for ultra-short-term scheduling. Long-term trends are accurately captured by the LSTM model, which possesses memory and attention-focusing capabilities. The attention mechanism adaptively identifies the key periods with the greatest impact on the current prediction from historical sequences up to 24 hours long. Combined with the long-term dependency modeling capabilities of the LSTM network, it accurately grasps the daily peak-valley patterns and time-specific characteristics, providing a reliable trend benchmark for day-ahead scheduling planning. The advantages of the two models are comprehensively leveraged in subsequent dynamic weighted fusion: short-term predictions rely more on the high-frequency tracking capabilities of the ultra-short-term model, while long-term predictions rely more on the trend-grasping capabilities of the short-term model, avoiding the predicament of a single model failing to address all aspects across timescales. As a result, the overall prediction accuracy is significantly improved compared to the single model scheme, and the estimation error of energy difference is effectively reduced, making the supply and demand information on which subsequent game decisions rely more accurate, thereby reducing improper allocation of shared power and waste of energy resources caused by prediction deviations from the source.

[0111] Based on the above embodiments and optional embodiments, an optional implementation method is provided, which is described in detail below.

[0112] In related technologies, existing residential photovoltaic and energy storage systems still have many shortcomings, which limit energy utilization efficiency and the promotion of new energy systems. At least the following drawbacks exist:

[0113] 1) The independent operation mode of a single household has inherent defects:

[0114] Typically, when a standalone photovoltaic (PV) and energy storage system operates independently, the average daily fluctuation coefficient is 30%-50% due to the randomness and intermittency of PV power generation. At night and during cloudy or rainy weather, PV output can drop to zero. Furthermore, due to the high cost of the system itself, most households can only afford small-capacity energy storage. Therefore, a single-household system struggles to simultaneously meet the three important requirements of peak shaving and valley filling, backup power, and PV power integration. Mathematical quantification of single-household system operation: Let the daily PV power generation be... Daily load consumption is The maximum charge / discharge capacity of the energy storage is Then the photovoltaic absorption rate With light discard The core formula is as follows:

[0115]

[0116]

[0117] in, This indicates the charging efficiency of the energy storage battery.

[0118] Substitute the actual parameters ( , , , The results show an absorption rate of 82.5%, a curtailment volume of 3.5 kWh, and a curtailment rate of 17.5%, which verifies the irrationality of the single-household model.

[0119] 2) Lack of community-level energy coordination and imbalance in resource allocation:

[0120] Most existing home energy storage systems operate independently, and surpluses and shortages between households cannot be compensated for. Typically, the overall energy efficiency of a community is below 60%. Assuming the community has... The family, the first The daily energy difference for each household is:

[0121] (A positive charge indicates a surplus of electricity, and a negative charge indicates a shortage of electricity), and its quantification formula is as follows:

[0122]

[0123]

[0124] It is the first Energy difference between individual families;

[0125] It represents the total surplus electricity of all households in the community;

[0126] This represents the total power shortage for all households in the community.

[0127] If surpluses and shortages within a community cannot be shared and complemented, it will result in a total waste of energy.

[0128] When the energy stored between households in a community cannot be mutually complementary, the waste is the greatest, and households lacking electricity need to purchase electricity from the grid, increasing their electricity costs.

[0129] 3) The centralized scheduling model has shortcomings, resulting in low user participation:

[0130] Currently, most commonly used community-level shared energy storage systems adopt a centralized architecture, using a single control center to manage signal processing and system administration across the entire community. The drawback of this approach is obvious: as the number of users increases, the burden on the communication center also increases. Let's assume the number of households in the community is... The communication delay formula can be:

[0131]

[0132] in, It is the basic latency between communication base stations in the energy storage system control center. This represents the communication delay coefficient for each household. It's evident that there's a linear relationship between the number of households and communication delay; the more households, the greater the communication burden. Too many households can lead to lag in power regulation, causing short-term fluctuations in system voltage and frequency. Furthermore, centralized control systems are costly to expand later, have long construction periods, and are highly susceptible to single-point failures; a failure of a core node can easily paralyze the entire system.

[0133] Furthermore, current shared energy storage reward mechanisms are simplistic, often failing to consider peak and off-peak electricity demand and applying only a uniform subsidy standard, resulting in low returns for users. Many users also worry that frequent charging and discharging will lead to equipment damage, further reducing their willingness to share. These differences in willingness can be quantified mathematically.

[0134] Set up the community The willingness coefficient of each user is , Average community participation:

[0135]

[0136] when At that time, the daily utilization rate of shared energy storage is less than 30%, which means that a large amount of photovoltaic power generation is abandoned. This not only fails to effectively reduce users' electricity costs, but also fails to play its role in valley filling and peak shaving, and the system's payback period will be greatly extended.

[0137] 4) Traditional control algorithms have weak adaptability:

[0138] Photovoltaic-storage sharing systems are typical strongly nonlinear, time-delay coupled systems. Their photovoltaic output is constantly affected by many factors, including sunlight intensity, temperature, and wind speed. User behavior also varies intermittently. Furthermore, the charging and discharging of energy storage devices is strongly correlated with factors such as battery state of charge (SOC) and temperature. These factors are dynamic and complex. The most commonly used PID control algorithm in photovoltaic-storage systems is as follows:

[0139]

[0140] in, It is the error between the actual power and the reference power of the photovoltaic-storage system. These are the proportional, integral, and differential coefficients, respectively.

[0141] However, this algorithm can only describe the "linear, time-invariant" characteristics of the system and cannot accurately show the nonlinearity and time delay of the system. More importantly, traditional PID control parameters are tuned offline. When the number of households in the community and the parameters of the energy storage system change, these PID parameters cannot be dynamically adjusted accordingly, which will lead to a significant decrease in control accuracy, thereby affecting the balance of the entire system and increasing the pressure on interaction with the power distribution network.

[0142] In view of this, an optional embodiment of the present invention provides a method for determining household power output within a community, which can also be called a photovoltaic-storage sharing micro-circulation system for households and communities and its method based on multi-scale prediction, evolutionary game theory, and fractional-order collaborative control algorithms. This method can balance household autonomy and community collaboration, enabling accurate prediction and fair decision-making in the photovoltaic-storage system and management. The present invention constructs a "household-community two-layer architecture," integrating multi-scale prediction, evolutionary game theory, and fractional-order control algorithms, while simultaneously improving photovoltaic absorption rate, energy storage utilization rate, and system stability. The core solution consists of four main parts. Starting from the basic model and mathematical framework, a multi-scale prediction model is used to predict photovoltaic output and electricity load. Then, an evolutionary game theory model is used to make decisions on the community's shared energy power. Finally, a fractional-order collaborative control algorithm is used to achieve system stability. These four modules form a complete technical closed loop, ensuring efficient system operation and maximizing energy utilization. Figure 2 This is an overall operation flowchart of the photovoltaic-storage sharing micro-circulation system provided by an optional method of the present invention. Figure 3 This is a flowchart of the multi-scale prediction model provided by an optional method of the present invention. Figure 4 This is an evolutionary game-sharing decision-making flowchart provided by an optional method of the present invention. Figure 5 This is a flowchart of the fractional-order cooperative control algorithm provided by an optional method of the present invention, which is described below:

[0143] (I) Constructing the architecture and mathematical model of a shared photovoltaic-storage micro-circulation system:

[0144] The aim is to overcome the limitations of existing algorithms and establish a two-tiered photovoltaic energy storage sharing micro-circulation system framework that balances household autonomy and community collaboration. The model used in this section will clarify the rules within the framework and the relationships between them, while also considering the impact of factors such as losses, environment, and equipment that may be encountered in reality. Mathematical models at both the household and community levels will be established to accurately describe the energy flow patterns and equipment operating states within the system, providing theoretical support for subsequent design.

[0145] (1) System overall architecture definition:

[0146] This invention employs a two-layer distributed collaborative framework, consisting of a community layer and a household layer. These layers are interconnected via a bidirectional energy routing module and a high-speed communication network. Furthermore, the system strictly adheres to the operating principle of "local priority, community complementarity, and grid backup," achieving a three-tiered energy cycle within households, between households and the community, and between the community and the grid. This ensures that each household can control its own energy, achieves optimized energy resource allocation at the community level, maximizes energy utilization efficiency, and minimizes the impact on the power grid. The overall energy constraint equations satisfied by the system are as follows:

[0147] +

[0148] in, Indicates the first Households in time The photovoltaic power generation capacity is expressed in kW. It is the first Households in time The local power load is expressed in kW. Indicates time The power exchange between the system and the distribution network, expressed in kW. >0 indicates that the system purchases electricity from the grid. <0 indicates that the system sells electricity to the grid. This indicates that the community-centralized shared energy storage unit is in time The charging and discharging power is expressed in kW. >0 indicates that the shared energy storage is charging. <0 indicates that discharge is in progress. This represents the overall energy routing efficiency of the system, taking into account energy routing conversion losses at the household and community levels, as well as line transmission losses. Its value range is... ∈[0.95,0.97].

[0149] (2) Family-level core mathematical model:

[0150] This level of model accurately describes the operating state of a single residential photovoltaic energy storage system, directly determining the execution accuracy of subsequent algorithms. This model focuses on maintaining the household power balance equations and the SOC dynamic model. Furthermore, it fully considers the actual impact of losses and the environment on performance, and modifications have been made to achieve more accurate predictions.

[0151] 1) Power balance equation:

[0152] exist Time of the first The power balance equation for a household needs to comprehensively consider photovoltaic output, energy storage charging and discharging, local load, community shared power, and various energy losses. The core formula is:

[0153]

[0154] The definitions and ranges of the parameters in the formula are as follows:

[0155] ) is the first Home energy storage devices The charging and discharging power at any given time, measured in kW. )>0 indicates charging. ) < 0 indicates discharge;

[0156] It refers to the overall charging and discharging efficiency of energy storage devices. = ,in For charging efficiency, To determine the discharge efficiency, the geometric mean is used to reflect the total loss during the entire charge and discharge process, with a range of values. ;

[0157] It is the first family Shared power with the community at all times, in kW. >0 indicates that households are supplying surplus electricity to the community. <0 indicates that the household is obtaining a shortage of electricity from the community;

[0158] It is the first The equivalent resistance of a household's internal wiring, measured in Ω, is determined by the material, diameter, and length of the wiring, and its value ranges from [value missing]. [0.3, 1.0];

[0159] This is the rated voltage for a single-phase household power supply. =220V, which is the national standard value;

[0160] This is the power factor of household electrical load, which is affected by the proportion of inductive load, and its value range is... [0.85, 0.95];

[0161] This refers to the home-level bidirectional energy routing conversion efficiency, with a value range of... [0.96, 0.98].

[0162] in the formula The term represents the active power loss of the lines caused by the interaction between households and communities. This is a key indicator reflecting the actual energy transmission loss and makes up for the deficiency of traditional models that ignore line losses.

[0163] 2) Dynamic model of home energy storage SOC:

[0164] State of Charge (SOC) is a core indicator describing the remaining capacity of energy storage devices, and its prediction accuracy directly affects the accuracy of energy dispatch and control. Traditional models do not consider the impact of ambient temperature on the charging and discharging efficiency of energy storage devices. This invention introduces a temperature correction term and establishes a high-precision dynamic model for the SOC of home energy storage devices. The core formula is:

[0165] +

[0166] The model simultaneously satisfies the SOC safe operation constraints: ≤ ≤

[0167] The definitions and ranges of the parameters in the formula are as follows:

[0168] Home energy storage unit The value of SOC at time t. It is a home energy storage unit The value of SOC at time t is dimensionless and ranges from [0,1].

[0169] , Considering ambient temperature The corrected energy storage charging and discharging efficiency is a nonlinear function of temperature. , The value ∈ [0.85, 0.95] indicates that the further the temperature deviates from 25℃ (the optimal operating temperature of the battery), the lower the efficiency.

[0170] , , No. Home energy storage devices The charging and discharging power at any given time, in kW, are all non-negative values ​​and satisfy the following conditions: - ;

[0171] The model sampling time interval, combined with the real-time control requirements of the photovoltaic storage system, is taken as follows: =0.25h (15min);

[0172] : No. The rated capacity of each household energy storage device is in kWh, which conforms to the current mainstream configuration of household energy storage.

[0173] Range of values ∈[2,5]

[0174] , : Upper and lower limits for safe operation of energy storage SOC, take =0.2、 =0.8, to avoid overcharging and discharging, which would shorten the battery cycle life.

[0175] (3) Community-level core mathematical model:

[0176] This invention focuses on establishing a dynamic model of community-based centralized shared energy storage SOC and community energy balance constraints. It fully considers the charging and discharging power constraints of shared energy storage, the coupling characteristics of energy interaction among multiple households, and the grid connection constraints of the distribution network, so as to accurately reflect the energy supply and demand balance law at the community level.

[0177] 1) Shared Energy Storage SOC Dynamic Model:

[0178] The SOC variation of a community-centralized shared energy storage unit is affected by the combined power of all households' shared power, and is also constrained by its own charging and discharging power and charging and discharging efficiency. The core dynamic model is as follows:

[0179] +

[0180]

[0181]

[0182] The model simultaneously satisfies the security constraints of shared energy storage SOC:

[0183] ∈[0.15,0.85]

[0184] The definitions and ranges of the parameters in the formula are as follows:

[0185] , Community shared energy storage unit , The SOC value at time t is dimensionless and ranges from [0,1].

[0186] Rated capacity of community shared energy storage units, in kWh, according to ∈[0.3 0.5 Configuration;

[0187] : No. family The charging power of the shared energy storage unit to the community at all times, in kW, is a non-negative value;

[0188] Community shared energy storage unit At all times towards the first The discharge power of a household, in kW, is a non-negative value.

[0189] , The charging and discharging efficiency of community-shared energy storage units, achieved using high-power energy storage converters, is higher than that of residential energy storage. The value range is... ∈[0.92,0.97], ∈[0.88,0.93];

[0190] , The maximum charging and discharging power of the community shared energy storage unit, in kW. These are the factory-rated parameters of the equipment.

[0191] The sampling time interval is consistent with that of the home-level model. =0.25h, to ensure data synchronization.

[0192] 2) Community energy balance constraints:

[0193] Energy flow at the community level must follow the law of conservation of energy. This can be divided into internal self-balancing within the community and interactive balancing with the power grid.

[0194] When the total surplus photovoltaic power within the community can completely cover the total power shortage, and the shared energy storage capacity is sufficient to buffer spatiotemporal energy disparities, the system does not need to interact with the power grid, thus achieving a micro-circulation of energy within the community. Its constraints are:

[0195]

[0196] In the formula: This means that the total surplus electricity sent to the community by all households is equal to the total shortage electricity received by the community, thus replenishing the energy within the community.

[0197] When the total excess power generated by the photovoltaic system within the community cannot compensate for the total power shortage, or when the excess power exceeds the maximum charging capacity of the shared energy storage, the system needs to interact with the distribution network to compensate for the energy difference or consume the excess power. The constraints are:

[0198]

[0199] Meanwhile, the power exchange capacity of the distribution network must meet the grid connection power constraints:

[0200]

[0201] In the formula: , The maximum power that can be purchased and sold in the distribution network is determined by the power grid company and is expressed in kW. This is to prevent large-scale power interaction between the system and the distribution network from affecting the stability of the grid voltage and frequency.

[0202] (II) Design and convergence proof of multi-scale prediction model:

[0203] Photovoltaic power output is often random and intermittent due to natural factors; household electricity load is affected by user habits. If changes in photovoltaic output and electricity load cannot be accurately predicted, it will lead to unreasonable energy distribution, exacerbate energy loss, and increase electricity costs. This invention designs an "ultra-short-term-short-term" dual-scale prediction model, combining ultra-short-term prediction (15min-1h) and short-term prediction (1-24h).

[0204] (1) Prediction scale division and prediction accuracy index:

[0205] Given the real-time control and scheduling requirements of the optical storage shared micro-loop system, the prediction scale, prediction period, sampling interval, model selection, prediction target, and accuracy requirements are clearly defined to ensure that prediction models of different scales can accurately adapt to the different operational needs of the system. Table 1 shows an example of the division provided by the optional embodiments of the present invention, and the specific division is shown in Table 1 below.

[0206] Table 1

[0207]

[0208] The goal of ultra-short-term forecasting is to capture the high-frequency random fluctuations in photovoltaic output and power load, and to track real-time power dynamics. Short-term forecasting aims to capture the overall changing trends of photovoltaic output and power load, daily peak-valley characteristics, and time-specific patterns. Forecast accuracy is evaluated using the core indicator, "Mean Absolute Percentage Error (MAPE)." This indicator directly reflects the relative deviation between the predicted and actual values. The smaller the value, the higher the forecast accuracy. The core calculation formula is:

[0209]

[0210] In the formula: To predict the sample size, in ultra-short-term forecasting Take 4-16 (corresponding to a 15min-4h prediction period) for short-term prediction. Take 24 (corresponding to a 24-hour prediction period); For the first The actual power value (photovoltaic output or electrical load) of each sample is in kW; For the first The model-predicted power value for each sample, in kW.

[0211] (2) Ultra-short prediction model (based on KELM):

[0212] Ultra-short-term forecasting requires the ability to rapidly process time-series data with high-frequency fluctuations and output high-precision prediction results. Kernel Extreme Learning Machine (KELM) can effectively solve the problem of blindly selecting the number of hidden layer nodes, while retaining the advantages of ELM such as fast training speed and strong generalization ability, making it one of the optimal models for ultra-short-term forecasting of photovoltaic-storage systems.

[0213] 1) Core output formula of the model:

[0214] For the ultra-short-term electricity load forecast of the i-th household, the core output formula of the KELM model is:

[0215]

[0216] In the formula: For the first The first family The predicted electricity load at each ultra-short-term forecast time, in kW; The number of training samples for the model; The model outputs a dimensionless weight vector. The radial basis function (RBF) is the kernel function type chosen in this invention. It has the advantages of strong nonlinear mapping capability and low computational cost. The core formula is:

[0217] ,in The width of the kernel function; For the first The input feature vector at each prediction time includes historical power data, ambient temperature, time features, and other features strongly correlated with the prediction target, with dimensions of [dimensionality missing]. ( (for feature dimensions) This is the model bias term, measured in kW, used to correct model prediction bias.

[0218] The ultra-short-term photovoltaic power output prediction model has the same structure as the electricity load prediction model, only the input feature vector is changed. Replace it with a vector that includes features such as historical photovoltaic output, light intensity, solar irradiance, and ambient temperature to ensure that the model accurately captures the high-frequency fluctuation features of photovoltaic output.

[0219] 2) Model training and convergence verification:

[0220] The training process of the KELM model involves solving the output weights. With bias To avoid overfitting, a regularization term is introduced, and the objective function is constructed as follows:

[0221]

[0222] In the formula: This is a regularization parameter used to balance the model's fitting accuracy and generalization ability.

[0223] For the first The prediction error of each training sample.

[0224] By solving the above objective function using the Lagrange multiplier method, the analytical solution for the output weight β can be obtained:

[0225]

[0226] In the formula: It is an identity matrix with dimension 1. ; The kernel matrix has the following elements: ; The actual value vector of the training samples, with dimension . 1.

[0227] Convergence proof: When the radial basis kernel function Satisfying the Mercer condition, and the regularization parameter At time 0, the prediction error of the KELM model monotonically decreases as the number of training samples L increases, that is:

[0228]

[0229] At this point, the model's prediction accuracy meets the accuracy requirements for ultra-short-term prediction (MAPE≤5%), ensuring that the model can maintain stable prediction performance under different operating scenarios and sample sizes.

[0230] (3) Short-term prediction model (based on LSTM + attention mechanism):

[0231] Short-term forecasting requires accurately capturing the long-term dependencies and trend characteristics of time-series data. Long Short-Term Memory (LSTM) networks, by introducing three gating mechanisms—the forget gate, the input gate, and the output gate—effectively address the gradient vanishing and gradient exploding problems of traditional Recurrent Neural Networks (RNNs), thus accurately capturing the long-term dependencies in time-series data. Attention mechanisms can adaptively assign different weights to historical data at different times, focusing on key features that have a greater impact on the prediction results, thereby further improving the model's prediction accuracy. This invention organically combines LSTM with attention mechanisms to construct an LSTM+attention mechanism short-term forecasting model, fully leveraging the advantages of both to achieve high-precision short-term forecasts of photovoltaic output and power load.

[0232] 1) Core formula for LSTM layer:

[0233] LSTM layers utilize a gating mechanism to select and retain historical information. The core of the LSTM layer comprises four parts: the forget gate, the input gate, the cell state, and the output gate. The update formula for each moment is:

[0234]

[0235]

[0236]

[0237]

[0238]

[0239]

[0240] In the formula:

[0241] For the first The input feature vector at any given time includes historical power data from the previous 24 hours, predicted temperature, solar irradiance, and temporal features, with dimensions of [missing information]. 1;

[0242] These are the outputs of the input gate, forget gate, and output gate, respectively, with values ​​ranging from [0,1], to filter information.

[0243] Candidate cell state, For the first Cellular state at any given moment For the first The cell state at any given moment enables the memory and updating of historical information;

[0244] For the first The LSTM hidden state at time step 1 is the extracted temporal feature vector with dimension 1. 1 ( (for hidden layer dimensions)

[0245] Here is the model weight matrix, with dimensions as follows: ;

[0246] The model bias vector has dimensions of . 1;

[0247] for Activation function The hyperbolic tangent activation function is used. This is the element-wise multiplication operation.

[0248] 2) Core formula of the attention mechanism layer:

[0249] The attention mechanism layer receives the hidden state sequence from the LSTM layer output. We perform weighted processing, assigning higher weights to hidden states that have a greater impact on the prediction results. The core formula is:

[0250]

[0251]

[0252]

[0253] In the formula:

[0254] For the first Attention score while constantly hiding the state, dimensionless;

[0255] The attention weight vector has dimensions of . 1; The weight matrix has dimensions of . Let be the bias vector, with dimension . 1;

[0256] For the first Attention weights that are always hidden from view, satisfying Dimensionless;

[0257] The output layer weight matrix has dimensions of . This is the output layer bias term, in kW;

[0258] For the first family The short-term forecast of photovoltaic power output at any given time, in kW, is the final output of the model.

[0259] The short-term electricity load forecasting model and the photovoltaic power output forecasting model have completely identical structures, except that the input feature vectors are different. Replace it with a vector that includes historical load data, user behavior characteristics, ambient temperature, time characteristics, etc., to ensure that the model accurately captures the overall trend and time-specific characteristics of electricity load changes.

[0260] (4) Dual-scale prediction fusion strategy:

[0261] To fully leverage the advantages of both the ultra-short-term forecasting model (KELM) and the short-term forecasting model (LSTM+Attention), this invention designs a dynamic weighted dual-scale forecasting fusion strategy. It adopts the approach of "short-term forecasting to determine the trend and ultra-short-term forecasting to correct fluctuations" to fuse the forecast results of the two models and obtain the final photovoltaic power output / electricity load forecast.

[0262] The core formula of the fusion strategy is:

[0263]

[0264] The fusion weights must satisfy the normalization constraint:

[0265]

[0266] And the weighting coefficients are non-negative.

[0267] In the formula:

[0268] for The final fusion prediction value at any given time, in kW (photovoltaic output or electrical load).

[0269] This is the short-term prediction value of the LSTM+Attention model, in kW, used to determine the overall trend of the prediction results. This is the ultra-short-term forecast value of the KELM model, in kW, used to correct for high-frequency random fluctuations in the forecast results;

[0270] , for The dynamic weighting coefficients for each time point are determined based on the time interval between the predicted time point and the current time point. Dynamic adjustment, the adjustment rules are as follows:

[0271]

[0272] In the formula: In ultra-short-term forecasting, the time interval between the predicted time and the current time is used to calculate the time difference. ∈[0.25,1]h,

[0273] In short-term forecasts ∈[1,24]h; For the maximum prediction time interval, take =24h.

[0274] This adjustment rule ensures that: the closer the predicted time is to the current time ( The smaller the value, the greater the weight of ultra-short-term forecasting. The larger the value, the further the predicted time is from the current time. The larger the value, the greater the weight of short-term forecasts. The larger the scale, the better the accuracy of the fused prediction results compared to the single-scale model, while simultaneously meeting the dual data requirements for real-time system control and scheduling planning.

[0275] The dual-scale fusion prediction results serve as important input variables for subsequent community energy sharing decision-making models. (Photovoltaic output prediction values) Compared with the predicted electricity load It can be used to calculate the energy surplus or deficit for each household in future scheduling cycles, thereby calculating the household's energy status. .

[0276] in Indicates the first In the future, families The predicted energy difference at each moment. This parameter will serve as an important basis for community-shared game theory and as key input data for the subsequent evolutionary game payoff function.

[0277] (III) Design and Equilibrium Analysis of Shared Decision-Making Model Based on Evolutionary Game Theory:

[0278] A key aspect of community energy sharing is balancing the benefits for individual households with the overall benefits for the community. Fair and efficient implementation strategies should be developed to enhance user participation. Existing mechanisms often employ fixed rules, neglecting the interests of individual households and the maximization of benefits, resulting in low user participation and low utilization rates of shared energy storage.

[0279] Based on evolutionary game theory, this invention fully considers the balance between individual household benefits and overall community benefits, achieving a win-win situation for both individual household benefits and overall community benefits. At the same time, it designs a dynamic electricity price incentive mechanism to guide households to actively participate in community energy sharing, which can effectively improve user participation and the utilization rate of shared energy storage.

[0280] (1) Definition of game participants and strategy set:

[0281] The evolutionary game participants in this invention are those who join the community for sharing. There are a limited number of households, each with independent decision-making capabilities, aiming to maximize their own benefits. Household decisions are influenced by various factors such as installed capacity, electricity costs, and shared benefits.

[0282] Each household can choose to participate in sharing (S1) or not participate in sharing (S2), which constitutes the basic strategy set of the evolutionary game.

[0283] To more accurately describe the decision-making hesitation and dynamic adjustment process of families, a mixed strategy is introduced, transforming the family's strategy selection into a probability distribution: Let the... The mixed strategy of each family is in ∈[0,1], ∈[0,1]

[0284] In the formula: For the first The probability that a family will choose the "participation and sharing" strategy. The introduction of mixed strategies to determine the probability of choosing the "not participating in sharing" strategy makes the game model more closely resemble real-world family decision-making scenarios.

[0285] (2) Construction of the revenue function and the payoff matrix:

[0286] 1) Profit function:

[0287] Household strategy selection aims to maximize their own gains. This invention divides gains into three parts: basic gains, shared additional gains, and shared gains. Furthermore, based on pairwise games between any two households, a game payoff matrix is ​​constructed to calculate the gains under different strategies, serving as the basis for subsequent solutions.

[0288] No. The core payoff function for a family under the hybrid strategy is:

[0289]

[0290] Definitions and calculation formulas of each component in the formula:

[0291] Basic income :family The net income when not participating in community energy sharing, i.e., the income under the single-household independent operation mode, consists of two parts: income from grid connection of surplus photovoltaic power and cost of purchasing electricity from the distribution network. The core formula is:

[0292]

[0293] In the formula:

[0294] For family The average daily surplus of photovoltaic power, in kWh;

[0295] The electricity purchase price is expressed in yuan / kWh. For family The average daily electricity purchased from the power grid, in kWh; This refers to the grid electricity price, expressed in yuan / kWh. (Sharing additional revenue.) :family The additional net benefit gained from participating in community energy sharing is the core incentive for households to participate. It consists of two parts: the benefit from the shared electricity price difference and the benefit from the community sharing subsidy. The core formula is:

[0296]

[0297] In the formula: For family The average daily amount of shared electricity delivered to the community is measured in kWh. The electricity price is shared by the community. The unit is yuan / kWh. For family The average daily shared electricity obtained from the community is measured in kWh. The electricity price is shared by the community, and the unit is yuan / kWh; For family Average daily shared electricity volume ( + (), the unit is kWh; The shared subsidy standard for the community is expressed in yuan / kWh. (Sharing cost) :family The total cost incurred when participating in community energy sharing consists of three parts: energy transmission loss cost, energy storage equipment cycle loss cost, and two-way energy routing equipment depreciation cost. The core formula is:

[0298]

[0299] In the formula: For family The average daily energy storage cycle power generated due to sharing is expressed in kWh. The unit cycle loss cost of energy storage batteries is expressed in yuan / kWh. The purchase cost of a home-grade two-way energy routing device is expressed in yuan. The design lifespan of the equipment is given in years.

[0300] 2) Game payoff matrix:

[0301] The payoff matrix of evolutionary game is based on the pairwise game between any two households. According to the strategy combination of the two households (S1-S1, S1-S2, S2-S1, S2-S2), combined with the core payoff function of the households, the pairwise game payoff matrix is ​​constructed. Table 2 is a schematic table of the payoff matrix provided by the optional implementation of the present invention, as shown in Table 2 below.

[0302] Table 2

[0303]

[0304] In the formula:

[0305] :family In the family When choosing to participate in sharing, you choose the benefits of participating in sharing.

[0306] :family In the family The benefits of choosing not to participate in sharing;

[0307] :family In the family When you choose not to participate in sharing, you receive the benefits of not participating in sharing.

[0308] , , Corresponding family Returns under different strategy combinations, defined in relation to the family Consistent.

[0309] Within the community The multiple games of individual families can be expanded through pairwise games to ultimately achieve game equilibrium analysis of the entire community. The payoff matrix clarifies the differences in family income under different strategy combinations and is the core foundation for solving evolutionary stable strategies.

[0310] (3) Solving for Evolutionary Stable Strategy (ESS) and proving Nash equilibrium:

[0311] Evolutionarily stable strategy (ESS) refers to a strategy that remains unchanged even when all individuals in a group adopt the same strategy, indicating its resilience and stability. This invention describes the evolution of household strategies in a community using a replication dynamic equation and solves for the evolutionarily stable strategy. Simultaneously, it rigorously proves the Nash equilibrium of the game, clarifies the optimal decision state for community energy sharing, and designs a dynamic electricity price incentive mechanism to ensure that households converge to the evolutionarily stable strategy of "participation in sharing."

[0312] 1) Replicating the dynamic equations:

[0313] The core idea of ​​the replication dynamic equation is that the higher the reward of a strategy, the greater its proportion will gradually become in the population.

[0314] For the For a given family, the replication dynamic equation for choosing the "participatory sharing" strategy is:

[0315]

[0316] In the formula:

[0317] For family The rate of change of the probability of choosing the “participation and sharing” strategy reflects the speed at which this strategy evolves in family decision-making.

[0318] For family The expected return when choosing the "participation and sharing" strategy is given by the core formula: For family The core formula for the average expected return under a hybrid strategy is: ,in For family Expected benefits when choosing the "do not participate in sharing" strategy.

[0319] Will and Substituting into the replication dynamic equation, it can be simplified to:

[0320]

[0321] 2) Solving for the Evolutionary Stable Strategy (ESS):

[0322] The evolutionary stabilizing strategy is to replicate the asymptotically stable fixed point of the dynamic equation. The solution process is as follows: First, set the first derivative of the replicating dynamic equation to 0, i.e. The fixed point of the equation is obtained by solving the equation. The stability of the fixed point is determined by judging the sign of the second derivative of the dynamic equation at the fixed point. When the second derivative is less than 0, the fixed point is an asymptotically stable point, i.e., an evolutionarily stable strategy (ESS).

[0323] Solving for fixed points: Let We can obtain three potential fixed points:

[0324]

[0325] In the formula: For internal fixed points, This indicates that all households have chosen "not to participate in sharing". This indicates that all families have chosen to "participate in sharing".

[0326] Determine stability: Take the second derivative (Jacobi matrix) of the replicating dynamic equation:

[0327]

[0328] According to the criteria for determining an evolutionarily stable strategy, when When the fixed point is asymptotically stable, it represents an evolutionarily stable strategy.

[0329] when hour, Place ,at this time =1 represents an evolutionarily stable strategy;

[0330] when hour, Place ,at this time =0 represents an evolutionarily stable strategy.

[0331] 3) Nash equilibrium proof and dynamic electricity price incentive mechanism:

[0332] Nash equilibrium proof: The core condition is satisfied when the expected payoff for a household choosing "participation in sharing" is greater than or equal to the expected payoff for "not participating in sharing."

[0333] hour, The globally unique evolutionarily stable strategy (ESS) in the community energy-sharing evolutionary game is defined as follows: All households in the community choose the "participate in sharing" strategy, the system reaches Nash equilibrium, and the overall energy-sharing benefit of the community is maximized. Substituting the expected benefit into the core conditions, it can be simplified to the sharing benefit incentive conditions:

[0334]

[0335] In other words, when the additional benefits that a family receives from participating in sharing are greater than or equal to the cost of sharing, the family will actively choose to participate in the sharing strategy.

[0336] To ensure this core condition remains in effect, guide families towards To address the convergence of evolutionary stabilization strategies, this invention designs a community dynamic electricity price incentive mechanism. The core of this mechanism is to adjust the community shared electricity price in real time based on factors such as the community's photovoltaic surplus / shortage status, shared energy storage SOC status, and distribution network electricity price. and shared electricity pricing The adjustment rules are as follows:

[0337]

[0338] In the formula: , This is the electricity price adjustment coefficient, taken as... ∈[0.05,0.2] yuan / kWh; Shared energy storage for the community The SOC value at time t. The shared power reference value can be calculated based on the above game theory results. .

[0339]

[0340] This value will be used as the tracking target for subsequent control power, thereby achieving reasonable allocation and optimized utilization within the community.

[0341] (iv) Design and stability verification of fractional-order cooperative control algorithm:

[0342] Existing photovoltaic-storage systems commonly use traditional PID control algorithms, which are only suitable for linear, time-invariant systems and cannot accurately describe the nonlinear and time-delay characteristics of photovoltaic-storage sharing systems. This invention, based on Caputo fractional calculus theory, designs a home-community two-layer fractional-order collaborative control algorithm to replace traditional integer-order PID control. Simultaneously, it rigorously verifies the system's stability based on Lyapunov stability theory and optimizes control parameters using particle swarm optimization (PSO) algorithm, thereby improving the system's operational stability and control accuracy.

[0343] (1) Definition of Caputo fractional calculus:

[0344] For continuous-time functions (t) The core definition of Caputo fractional calculus is:

[0345] Fractional order integral operator (order) ∈(0,1)):

[0346] d

[0347] Fractional differential operators (order) ∈(0,1)):

[0348] d

[0349] In the formula: The gamma function is a core mathematical tool in fractional calculus.

[0350] Defined as:

[0351] The gamma function satisfies the recurrence relation ,when When it is a positive integer, In other words, integer calculus is a special case of fractional calculus, while fractional calculus is more general.

[0352] Compared with traditional integer-order calculus (order 1 or 0), Caputo fractional-order calculus can accurately characterize the memory and time delay of the optical-storage sharing system, theoretically breaking through the limitations of traditional integer-order control and laying the mathematical foundation for the design of high-precision collaborative control algorithms.

[0353] (2) Design of a two-level fractional-order collaborative control law for family and community:

[0354] 1) Control law design:

[0355] The fractional-order cooperative control law designed in this invention replaces the first-order integral and first-order derivative with Caputo fractional-order integral and fractional-order derivative, based on traditional integer-order PID control. It introduces two adjustable parameters, the integral order α and the derivative order β, to improve the degree of freedom and adaptability of the control algorithm. The core control law formula is as follows:

[0356]

[0357] The parameters in the formula are defined as follows:

[0358] The control output at any time is the charging and discharging power command of the energy storage device, in kW. The household level is the charging and discharging power command of the household energy storage, and the community level is the charging and discharging power command of the shared energy storage.

[0359] The control error at any given time, i.e., the difference between the target control value and the actual value, is dimensionless or has the same dimension as the target value. The core formula is: (Power Point Tracking) or (SOC stability control), where Indicates a reference value. Indicates the actual value;

[0360] : Proportional coefficient, dimensionless or inversely proportional to the control error, used for rapid response to control errors;

[0361] Integral coefficient, dimensionless or inversely proportional to the control error, used to eliminate steady-state error;

[0362] Differential coefficients are dimensionless or inversely proportional to the control error, used to suppress system overshoot and improve stability.

[0363] ∈(0,1): Caputo fractional order integral order, dimensionless, used to adapt the integral characteristics of the system;

[0364] ∈(0,1): Caputo fractional derivative order, dimensionless, used to adapt the differential properties of the system.

[0365] when , At this time, the fractional-order cooperative control law degenerates into the traditional integer-order PID control law:

[0366]

[0367] Therefore, traditional integer-order PID control is a special case of the fractional-order collaborative control of this invention, and the control algorithm of this invention is more general and adaptable.

[0368] 2) Control parameter optimization based on PSO:

[0369] Fractional-order cooperative control laws include , , , , Five parameters need to be optimized, and their values ​​directly determine the performance of the control algorithm. Traditional manual tuning methods are inefficient and inaccurate, and cannot achieve globally optimal parameter configuration. This invention uses the Particle Swarm Optimization (PSO) algorithm to globally optimize the five control parameters, taking overshoot, settling time, and steady-state error as optimization objectives, constructing a multi-objective optimization function, and solving for the optimal parameter combination.

[0370] Optimize the objective function:

[0371]

[0372] In the formula: To control overshoot (%). The steady-state time (s) This is the steady-state error; , , The maximum allowable value for each indicator is taken as follows: , =5s、 ; , , For the weighting coefficients, satisfying According to control requirements, the present invention takes , , Prioritize ensuring overshoot and settling time.

[0373] Parameter constraints:

[0374] ,

[0375] ,

[0376] ,

[0377] ∈(0,1)

[0378] ∈(0,1)

[0379] 3) PSO optimization:

[0380] Step 1: Initialize the particle swarm, with a particle dimension of 5 (corresponding to...) , , , , The number of particles is set to 50, the number of iterations is set to 100, and the inertia weight is set to... Linear decrease ( (∈[0.4,0.9]), learning factor ;

[0381] Step 2: Calculate the fitness value of each particle (i.e., the optimization objective function). );

[0382] Step 3: Update the individual optimal position of each particle. and the optimal position of the group ;

[0383] Step 4: Adjust particle positions according to the PSO velocity and position update formula:

[0384]

[0385] In the formula: , For the first Particle velocity and position in the next iteration , It is a random number;

[0386] Step 5: Repeat steps 2-4 until the maximum number of iterations is reached, and output the optimal position of the population. This is the optimal combination of control parameters.

[0387] The optimized control parameters obtained through the PSO algorithm enable the fractional-order cooperative control law to achieve optimal performance and adapt to the complex dynamic characteristics of the optical-storage sharing system.

[0388] (3) System stability verification:

[0389] System stability is a core requirement for control algorithm design. Only by ensuring that the system reaches asymptotic stability under the control algorithm can the safe and stable operation of the system be guaranteed. This invention, based on Lyapunov stability theory, rigorously verifies the stability of a shared optical-storage micro-circulation system using a fractional-order cooperative control algorithm, theoretically proving the asymptotic stability of the system. Simultaneously, simulation verification and engineering experiments are used to verify the control performance of the algorithm.

[0390] 1) Proof of theoretical stability:

[0391] The dynamic characteristics of the photovoltaic-storage shared microcirculation system can be described by fractional-order nonlinear state equations:

[0392]

[0393] In the formula: Let the order of the fractional system be denoted by . It is the system state vector (including energy storage SOC, shared power, distribution network interaction power, etc.). For control input (fractional order cooperative control law output). It is a nonlinear function that describes the evolution of the system state.

[0394] The control objective is: when At that time, the system state Converging to the reference state That is, control error

[0395] 0.

[0396] Stability proof process:

[0397] Construct positive definite Lyapunov functions:

[0398] In the formula: It is a positive definite symmetric matrix that satisfies the Lyapunov equation. ( (is a positive definite symmetric matrix), ensuring For all Established.

[0399] Find the fractional derivative of the Lyapunov function (based on the chain rule of Caputo fractional derivatives):

[0400]

[0401] Substituting the system state equations into the fractional derivative expression of the control error:

[0402]

[0403] Due to reference state For constants or slowly changing quantities, they can be approximated. ,therefore:

[0404]

[0405] Fractional-order cooperative control law Substituting into the system state equations and considering the constraints satisfied by the parameters after PSO optimization, we can derive:

[0406]

[0407] In the formula: It is a positive definite symmetric matrix, ensuring For all Established.

[0408] According to the Lyapunov stability criterion for fractional-order systems, when a positive definite Lyapunov function exists... And its fractional derivative At that time, the control error of the system Will over time The increase of approaches zero, that is:

[0409]

[0410] Therefore, the shared micro-circulation system of light and energy storage using the fractional-order collaborative control algorithm is asymptotically stable.

[0411] Simulation scene settings: Solar power output suddenly dropped by 30%. A sudden 20% increase in household electricity load; the control target is shared power tracking error. The energy storage SOC is stable in the range of [0.4, 0.6]. Table 3 is a schematic table of simulation results provided by the optional embodiments of the present invention. The simulation results are shown in Table 3 below.

[0412] Table 3

[0413]

[0414] Simulation results show that the fractional-order collaborative control algorithm has significantly better overshoot, settling time, and steady-state error than the traditional integer-order PID control. Furthermore, the energy storage SOC fluctuation range is smaller, which can better cope with disturbances in photovoltaic output and electricity load and meet the real-time control requirements of the system.

[0415] The above optional implementation methods can achieve at least the following beneficial effects:

[0416] This study addresses common problems of single-household solar energy storage systems, such as poor coordination within communities, low dispatch efficiency, and unstable control. A two-tiered, home-community shared solar-storage micro-circulation system was designed, connecting homes and communities and integrating multi-scale forecasting, game theory, and advanced control algorithms into a unified whole. The key lies in precise calculations to accurately track energy flow, utilizing intelligent forecasts of solar energy and electricity consumption to guide sharing decisions. The game theory component ensures fair allocation and increases user participation, while control technology guarantees stable system operation. This gradual approach improves solar energy utilization, makes energy storage more efficient, stabilizes the system, and reduces costs, providing a practical pathway for households to transition to clean energy, with significant practical application potential.

[0417] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0418] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0419] Example 2

[0420] According to embodiments of the present invention, an apparatus for implementing the above-described method for determining household power output within a community is also provided. Figure 6 This is a structural block diagram of a household power output determination device in a community according to an embodiment of the present invention, such as... Figure 6 As shown, the device includes: an acquisition module 602, a prediction module 604, a fusion module 606, a first determination module 608, an adjustment module 610, and a second determination module 612. The device will be described in detail below.

[0421] The acquisition module 602 is used to acquire household electricity consumption data corresponding to multiple households within a predetermined community, wherein the household electricity consumption data includes historical power data, environmental data, and time characteristics; the prediction module 604, connected to the acquisition module 602, is used to perform multi-time-scale prediction operations on the household electricity consumption data to obtain multi-time-scale prediction data, wherein the multi-time-scale prediction data includes photovoltaic output prediction values ​​and electricity load prediction values; the fusion module 606, connected to the prediction module 604, is used to perform weighted fusion operations on the multi-time-scale prediction data according to a preset dynamic weighted fusion strategy to obtain fused prediction data for the multiple households at future times; the first determination module 608, connected to the fusion module 606, is used to determine the predicted energy difference corresponding to the multiple households based on the fused prediction data for the multiple households at future times, wherein the predicted energy difference The difference between the predicted output of the integrated photovoltaic system and the predicted electricity load is denoted as 608. The adjustment module 610, connected to the first determining module 608, is used to adjust the shared electricity price within the community based on the dynamic electricity price adjustment rules and the community shared energy storage state-of-charge value when multiple households are not evenly participating in the sharing process. This is to maintain the target stability condition corresponding to the predetermined community, ensuring that the multiple households are in a state of selective participation in the sharing process. The target stability condition includes that the additional sharing revenue is greater than or equal to the household sharing cost. The second determining module 612, connected to the adjustment module 610, is used to determine the shared power value corresponding to each of the multiple households when all households are participating in the sharing process, based on the community shared energy storage state-of-charge value, the predicted energy difference corresponding to each household, and the household energy storage state-of-charge value. Control is then performed based on the corresponding shared power value.

[0422] It should be noted that the above-mentioned acquisition module 602, prediction module 604, fusion module 606, first determination module 608, adjustment module 610 and second determination module 612 correspond to steps S102 to S112 in the method for determining household power output within the implementation community. The instances and application scenarios implemented by multiple modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiment 1.

[0423] Example 3

[0424] According to another aspect of the present invention, an electronic device is also provided, comprising: a processor; and a memory for storing processor-executable instructions, wherein the processor is configured to execute the instructions to implement the method for determining household power output within a community as described above.

[0425] Example 4

[0426] According to another aspect of the present invention, a computer-readable storage medium is also provided, which, when the instructions in the computer-readable storage medium are executed by a processor of an electronic device, enables the electronic device to perform the community-based household power output determination method described above.

[0427] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0428] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0429] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0430] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0431] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0432] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0433] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for determining a household effort power in a community, characterized by, include: Obtain household electricity consumption data for multiple households within a predetermined community, wherein the household electricity consumption data includes historical power data, environmental data, and time characteristics; The household electricity consumption data is subjected to multi-time-scale prediction operations to obtain multi-time-scale prediction data, wherein the multi-time-scale prediction data includes photovoltaic power output prediction values ​​and electricity load prediction values. Based on a preset dynamic weighted fusion strategy, a weighted fusion operation is performed on the multi-time-scale prediction data to obtain the fused prediction data of the multiple families at future times. Based on the combined predicted data of the multiple households at future times, the predicted energy difference corresponding to the multiple households is determined, wherein the predicted energy difference is the difference between the combined photovoltaic output predicted value and the combined electricity load predicted value; When multiple households are not evenly participating in the sharing state, the sharing price within the community is adjusted according to the dynamic electricity price adjustment rules and the community shared energy storage state of charge value, so as to maintain the target stability condition corresponding to the predetermined community and keep the multiple households in the state of choosing to participate in the sharing. The target stability condition includes that the additional sharing income is greater than or equal to the household sharing cost. When all the households are participating in the sharing state, the shared power value corresponding to each of the multiple households is determined based on the community shared energy storage state of charge value, the predicted energy difference corresponding to each household and the household energy storage state of charge value, so as to carry out control based on the corresponding shared power value.

2. The method according to claim 1, characterized in that According to the dynamic electricity price adjustment rules, the shared electricity price within the community is adjusted based on the state of charge value of the shared energy storage in order to maintain the target stability condition corresponding to the predetermined community, so that the multiple households are in a state of choosing to participate in sharing, including: Determine multiple game payoff matrices under different strategy combinations, wherein the game payoff matrix is ​​a matrix representing the strategy combination corresponding to the shared participation outcome of one household among multiple households and the shared participation outcome chosen by another household; Based on the multiple game payoff matrices under different strategy combinations, determine the payoff values ​​under different strategy combinations; Based on the returns under the different strategy combinations, the expected returns and average expected returns for each of the multiple families choosing to participate in the sharing strategy are determined. Based on the expected and average expected returns of the participating strategies for each of the multiple households, replication dynamic equations are established for each of the multiple households, where the replication dynamic equations represent the rate of evolution of the probability of a household choosing to participate in the sharing strategy over time. Multiple fixed points of the replicated dynamic equations are determined, and stability analysis is performed on each fixed point to obtain the stability determination result of each fixed point. When any household's fixed point is determined to be unstable, the shared electricity price within the community is adjusted according to the dynamic electricity price adjustment rules and the state of charge value of the community shared energy storage until all stability determination results are stable.

3. The method according to claim 2, characterized in that Determine multiple game payoff matrices under different strategy combinations, including: Obtain the basic income corresponding to each of the multiple families, share additional income and family sharing costs; Based on the basic income corresponding to each of the multiple families, the shared additional income, and the family sharing cost, determine the income function corresponding to each of the multiple families; Based on the evolutionary game model, a set of strategies for participating in and not participating in sharing is set for each household, and a game payoff matrix is ​​constructed under different strategy combinations in the set based on the payoff function corresponding to the corresponding household.

4. The method of claim 2, wherein, Obtain the basic income corresponding to each of the multiple families, share additional income and family sharing costs, including: Acquire community electricity data, as well as electricity transaction data and energy loss data for multiple households within the community. The community electricity data includes the community shared electricity sales price, the community shared electricity purchase price, and the shared subsidy standard. The electricity transaction data includes the average daily photovoltaic surplus electricity, the average daily electricity purchased from the grid, the grid purchase price, and the grid sales price. The energy loss data includes energy transmission loss cost, energy storage battery unit cycle loss cost, and routing equipment depreciation cost. Based on the electricity trading data, the basic income of each of the multiple households in independent operation mode is determined; Based on the community electricity data and the predicted energy difference corresponding to each of the multiple households, the shared electricity for each of the multiple households is determined. Based on the energy loss data corresponding to each of the multiple households and the shared electricity consumption of the households, the household sharing cost corresponding to each of the multiple households is determined; Based on the household-shared electricity consumption corresponding to each of the multiple households, the shared additional revenue corresponding to each of the multiple households is determined.

5. The method of claim 1, wherein, After determining the shared power value corresponding to each of the multiple households based on the community-shared energy storage state of charge value, the predicted energy difference corresponding to each of the multiple households, and the household energy storage state of charge value, the method further includes: Obtain the control target value and actual operating value of the energy storage device corresponding to each of the multiple households, and determine the control error corresponding to each of the multiple households, wherein the control target value is the corresponding shared power value; Based on the fractional-order cooperative control law defined by Caputo fractional calculus, proportional, fractional-order integral, and fractional-order derivative operations are performed on the corresponding control error to generate control commands corresponding to the multiple households. The corresponding control commands are used to adjust the charging and discharging power of the corresponding energy storage devices.

6. The method of claim 1, wherein, Before generating control commands corresponding to the multiple households by performing proportional, fractional integral, and fractional derivative operations on the corresponding control errors according to the fractional cooperative control law defined by Caputo fractional calculus, the process further includes: A multi-objective optimization function is invoked, wherein the multi-objective optimization function aims to jointly minimize overshoot, settling time, and steady-state error; Based on the multi-objective optimization function, the optimal combination of control parameters is obtained by global optimization within a preset parameter constraint range using the particle swarm optimization algorithm. The target parameters in the fractional-order cooperative control law are then determined, wherein the target parameters include proportional coefficient, integral coefficient, derivative coefficient, integral order, and derivative order.

7. The method according to any one of claims 1 to 6, characterized in that, include: When the multi-timescale prediction operation includes a first timescale prediction operation and a second timescale prediction operation, the first timescale prediction operation is performed on the household electricity consumption data according to the kernel extreme learning machine model to obtain first timescale prediction data, and the second timescale prediction operation is performed on the household electricity consumption data according to the long short-term memory network model with fused attention mechanism to obtain second timescale prediction data, wherein the first timescale is smaller than the second timescale.

8. A device for determining household power output within a community, characterized in that, include: The acquisition module is used to acquire household electricity consumption data for multiple households within a predetermined community. The household electricity consumption data includes historical power data, environmental data, and time characteristics. The prediction module is used to perform multi-time-scale prediction operations on household electricity consumption data to obtain multi-time-scale prediction data, which includes photovoltaic power output prediction values ​​and electricity load prediction values. The fusion module is used to perform weighted fusion operations on multi-time-scale prediction data according to a preset dynamic weighted fusion strategy, so as to obtain fused prediction data of multiple families at future times. The first determining module is used to determine the predicted energy difference for each of the multiple households based on the integrated prediction data of multiple households at future times. The predicted energy difference is the difference between the integrated photovoltaic output prediction value and the integrated electricity load prediction value. The adjustment module is used to adjust the shared electricity price within the community according to the dynamic electricity price adjustment rules and the community shared energy storage state value when multiple households are not evenly participating in the sharing state, so as to maintain the target stability condition corresponding to the predetermined community and keep multiple households in the state of choosing to participate in the sharing state. The target stability condition includes that the additional sharing income is greater than or equal to the household sharing cost. The second determining module is used to determine the shared power value corresponding to each of the multiple households when multiple households are participating in the sharing state, based on the community shared energy storage state value, the predicted energy difference corresponding to each household and the household energy storage state value, so as to control according to the corresponding shared power value.

9. An electronic device, comprising: include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the method for determining household power output within a community as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of the electronic device, the electronic device is able to perform the method for determining household power output within a community as described in any one of claims 1 to 7.

Images