A Multi-Dimensional Collaborative Community-Shared Energy Storage Optimization and Scheduling Method

Abstract

This invention belongs to the field of power system optimization and energy storage operation technology, specifically involving a multi-dimensional collaborative community-shared energy storage optimization scheduling method. It includes: establishing a dual-constraint scheduling model for a community-shared energy storage system based on relative diversity factors, combining a hard upper limit on total subscriptions with a soft data-driven threshold based on relative diversity factors, and using the non-overlapping nature of user loads to define the boundary of overlapping demand; constructing a two-layer optimization scheduling architecture of outer-layer price setting and inner-layer capacity allocation, introducing a piecewise linear proxy model to achieve fair capacity allocation with the goal of maximizing the minimum user benefit-cost ratio; and establishing an evaluation method that maps user differences to expected shortage costs. This invention provides a multi-dimensional collaborative community-shared energy storage optimization scheduling method that integrates user diversity constraints, supply-demand benefit balance, and risk closed-loop control, filling a gap in existing technology and promoting the large-scale application of shared energy storage in community scenarios.

Description

Technical Field

[0001] This invention belongs to the field of power system optimization and energy storage operation technology, specifically involving a multi-dimensional collaborative community-shared energy storage optimization scheduling method. Background Technology

[0002] With the acceleration of the global energy transition, building energy consumption accounts for about 30% of total energy consumption, making efficient energy management in the building sector a key aspect of energy conservation and emission reduction. Energy storage systems (ESS), as core equipment for peak shaving and valley filling and improving energy utilization efficiency, can store electrical energy during off-peak hours and release it during peak hours, effectively reducing the pressure on the power grid and users' electricity costs, while enhancing the reliability and flexibility of energy supply.

[0003] However, traditional building energy storage often adopts an independent configuration model of "one storage per household," which has significant drawbacks: on the one hand, the peak electricity consumption of a single user is intermittent and discontinuous, resulting in the energy storage equipment being idle most of the time, with the utilization rate of physical assets generally being less than 50%, causing resource waste and low return on investment; on the other hand, the initial investment, installation and maintenance costs of independent energy storage are high, and it is limited by urban building space constraints, making it difficult to promote on a large scale among small and medium-sized user groups.

[0004] To address these issues, the community-shared energy storage model has emerged. This model involves multiple users sharing a single physical energy storage device, leveraging the diversity of user electricity consumption patterns to improve energy storage utilization and reduce the cost burden on individual users. However, existing shared energy storage dispatching technologies still face key bottlenecks: First, overbooking strategies lack scientific constraints, often employing a fixed multiple hard cap model that fails to fully consider the non-overlapping characteristics of user electricity loads. This can easily lead to overlapping demand exceeding physical capacity, triggering grid procurement shortages and associated cost and reliability risks. Second, dispatching schemes fail to balance the interests of both supply and demand sides, either prioritizing service provider revenue maximization while neglecting user fairness, or overemphasizing fairness resulting in insufficient energy storage utilization. Third, risk assessment mechanisms are inadequate, lacking quantitative analysis and closed-loop feedback for overbooking risks, making it difficult to effectively control shortage risks while improving utilization. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multi-dimensional collaborative community shared energy storage optimization scheduling method that integrates user diversity constraints, supply and demand interest balance and risk closed-loop control, fills the gap in the prior art, and promotes the large-scale application of shared energy storage in community scenarios.

[0006] The objective of this invention is achieved as follows: a multi-dimensional collaborative community-shared energy storage optimization scheduling method, comprising the following steps:

[0007] A dual-constraint scheduling model for community-shared energy storage systems based on the Relative Diversity Factor (RDF) is established. This model combines a hard upper limit on total subscriptions with a soft data-driven threshold based on the RDF. By utilizing the non-overlapping nature of user loads, the boundary of overlapping demand is defined, thereby allowing over-booking of virtual capacity under limited physical capacity, and achieving a synergistic improvement in energy storage utilization and reliability.

[0008] A two-layer optimized scheduling architecture is constructed, consisting of an outer layer for pricing and an inner layer for capacity allocation. The outer loop adjusts the release price based on the service provider's survivability index, while the inner loop introduces a piecewise linear proxy model to allocate capacity fairly with the goal of maximizing the minimum user revenue-cost ratio, while meeting hard upper limits and differentiation rules.

[0009] By combining the risk diagnosis mechanism of Monte Carlo simulation, diversity is mapped to expected shortage costs, and an assessment method is established to map user differences to expected shortage costs. This enables community-shared energy storage systems to improve utilization while effectively controlling shortage risks and ensuring the economic benefits of users and service providers. At the same time, it takes into account the three core objectives of energy storage utilization, user fairness, and service provider revenue, and achieves multi-dimensional collaborative optimization.

[0010] Furthermore, the establishment of a dual-constraint scheduling model for a community-shared energy storage system based on relative diversity factors includes establishing a service provider model and a community unit model;

[0011] In the service provider model, the service provider provides energy services to units within the building energy community. Sales of virtual reserved capacity (kWh) to provide shared energy storage (SES) services; shared energy storage SES is supported by physical energy storage systems (ESS) with an energy capacity of E (kWh); shared energy storage systems include three types of entities: building energy communities (BECs) coordinated by energy community service providers (ECSPs), energy storage service providers (SPs), and the power grid;

[0012] In the community unit model, community units subscribe to shared energy storage (SES), obtain bill savings through peak shaving and valley filling, and pay for reserved services.

[0013] Furthermore, the establishment of the service provider model includes:

[0014] make The capacity cost model for energy storage service providers (SPs) represents the levelized cost of energy storage (LCOS) per billing cycle.

[0015]

[0016] The revenue of energy storage service providers (SPs) is calculated as follows:

[0017]

[0018] in The published service prices, It is a collection of units that participate in the service. ;

[0019] When the total concurrent requests from all users When it exceeds E, the energy storage service provider SP will supply the grid at a price Purchase difference in electricity volume To fill the gap The levelized price for the shortage; the expected shortage cost within a billing cycle is:

[0020]

[0021] in The time interval is The observation period;

[0022] The benefit-cost ratio (B / C) for energy storage service providers (SPs) is:

[0023]

[0024] Furthermore, the establishment of the community unit model includes:

[0025] make For unit In time The baseline load, To meet the electricity demand, For energy prices, the benchmark bill is:

[0026]

[0027] make Representation unit In time Shared energy storage SES net energy change, It is a unit Energy efficiency, using a standard one-way splitting method:

[0028]

[0029] The billing after using the shared energy storage SES is as follows:

[0030]

[0031] The benefit of the unit is cost savings. For the base bill;

[0032]

[0033] In the above formula This represents the original peak electricity load of user i across all time steps t. The above formula includes the reduction of electricity consumption during peak hours and the adjustment of electricity consumption time.

[0034] Under the shared energy storage SES tariff with only available capacity, the cost of the unit is:

[0035] (9)

[0036] The benefit-cost ratio (B / C) at the unit level is:

[0037]

[0038] The total cost of each unit capacity equals the revenue of the energy storage service provider (SP), i.e. .

[0039] Furthermore, the construction of the two-layer optimized scheduling architecture of outer-layer price setting and inner-layer capacity allocation includes establishing a baseline optimization framework without diversity constraints, providing a comparative basis for the subsequent introduction of RDF soft constraints, specifically including:

[0040] Given the published service price and participating units Allocate capacity ;unit The benefit is its cost savings. Its benefit-cost ratio is ;

[0041] make To achieve the target overbooking rate, the benchmark imposes a strict cap on the total number of subscriptions, as shown below:

[0042]

[0043] In equation (11), when Overbooking is not allowed at this time; with The increase in coverage leads to increased coverage, but fairness and reliability are limited.

[0044] While maintaining unit limits and overbooking caps, maximizing the benefit of the smallest unit is directly proportional, as shown below:

[0045]

[0046] The objective function in equation (12) To maximize the minimum unit revenue-cost ratio, where e is the virtual energy storage capacity subscription vector for community users, the constraints are as follows: For the provider's total revenue, For the total cost of the provider, Save money on user i's bill. For user i's subscription cost, Is it accepting pricing? The set of participating users This represents the maximum subscription capacity limit for user i. For the target oversold ratio ( (No overselling), E is the actual capacity of the physical energy storage system.

[0047] Furthermore, the construction of the two-layer optimized scheduling architecture for outer-layer price setting and inner-layer capacity allocation also includes:

[0048] For shared energy storage (SES) planning, based on the user's historical capacity Weight their demand curves and use the Relative Diversity Factor (RDF); from recent history, for each unit Extracting standardized demand curves ,in Given a subscription vector The definition of RDF is as follows:

[0049]

[0050] Where the denominator is approximately The implied worst-case concurrent requests and the observed curve shape. A higher RDF indicates greater multi-tenancy diversity.

[0051] Replace the hard overbooking cap with the following soft overbooking constraint based on the diversity factor (DF):

[0052]

[0053] And add a DF-based soft overbooking constraint:

[0054]

[0055] Furthermore, the outer loop adjusts the release price based on the service provider's viability index, while the inner loop, under the premise of meeting hard caps and differentiation rules, introduces a piecewise linear proxy model to fairly allocate capacity with the goal of maximizing the minimum user revenue-cost ratio, including:

[0056] Given the published service price The inner layer will include subscription volume Assign to users, using the bill saving function and capacity fees To maximize the minimum B / C, as shown below:

[0057]

[0058] User capacity is allocated to maintain a minimum B / C ratio, i.e. As shown below:

[0059]

[0060] Satisfy the unit-by-unit boundary, hard overbooking cap, and DF-based soft overbooking constraints:

[0061]

[0062] Inner layer solution The service provider's benefit-cost ratio (B / C) is calculated as follows:

[0063]

[0064] in Risk costs are quantified using the Monte Carlo method; based on historical days / scenarios. Generate concurrent requests , A scene-specific curve;

[0065] While maintaining a high degree of fairness, Feasibility of updating to the target service provider The following strategy is adopted:

[0066]

[0067] in It is a small step size; if Then improve Otherwise, it can be reduced. To enhance user engagement and .

[0068] Furthermore, the risk diagnosis mechanism combined with Monte Carlo simulation maps diversity to expected shortage costs, establishing an assessment method that maps user differences to expected shortage costs, including:

[0069] Construct a random set of scenarios and calculate the scenario-based physical shortage;

[0070] Assess the expected shortage risk costs and establish a two-tier overbooking allocation and a DF-based inner allocation algorithm.

[0071] Furthermore, the construction of a random scene set and the calculation of scene-based physical shortages include:

[0072] Extract features from the historical activity data of community users to construct a system containing... A collection of independent historical scenes For each scenario in the set Get each user In time Normalized concurrency curve The normalized concurrency curve reflects the load behavior pattern of users at a specific historical moment; by constructing a scenario set through historical data, a data foundation that fits the actual electricity consumption behavior is provided for subsequent shortage simulation;

[0073] Optimal capacity allocation scheme based on inner loop output Simulation calculations in each scenario The concurrent request situation under the following scenario; for the scenario time The physical shortage path is calculated as follows:

[0074]

[0075] In the formula Let h be the physical shortage amount (kWh) under scenario h and time t. Let h represent the concurrent request volume (kWh) of user i under scenario h and time t. The physical capacity of the shared energy storage system is defined as follows: if the aggregated demand is less than the physical capacity, the shortage is 0; if it exceeds the physical capacity, the excess portion is the physical shortage.

[0076] Furthermore, the assessment of anticipated shortage risk costs and the establishment of a two-tier overbooking allocation and a DF-based inner-layer allocation algorithm include:

[0077] For scene set Statistical analysis was performed on the physical shortages across all scenarios, and diagnostic indicators were calculated for each time step:

[0078]

[0079] Combined with the real-time electricity purchase price of the power grid or the preset power shortage penalty price Calculate the expected risk cost over the entire scheduling cycle, transforming the abstract shortage risk into a concrete and calculable economic cost:

[0080]

[0081] The two-level overbooking allocation and the DF-based inner allocation algorithm are as follows:

[0082] First, enter the physical capacity. Overbooking price DF curve set Piecewise linear net benefit function Breakpoints and slopes, maximum booking volume SP Costs The set of levelized prices in shortage Maximum number of iterations Step size sequence set Initial price And let the number of iterations be ;

[0083] Then when When entering the main loop, first determine the set of participating units. Units that satisfy the condition that cost savings exceed expenses are then optimized using an inner-level binary search method. :set up upper and lower boundaries and In the above formula, It is the initial upper bound of the binary search. For all users and its segmentation Take the maximum value. Corresponding user Corresponding piecewise linear net benefit function The Slope, It is the first The system in the next outer iteration charges users a unit price for pre-booked capacity; when the difference between the upper and lower bounds is greater than the tolerance. When, take the median value. And solve the linear programming LP feasibility problem, with constraints including the user benefit-cost ratio not being less than Unit capacity is between 0 and the upper limit, and total subscriptions do not exceed the hard oversubscription limit. In the worst-case scenario, overlapping demand should not exceed physical capacity; if LP (Limited Limit) is feasible, then update. And store candidate capacity Otherwise update ,until Convergence, yielding the optimal capacity. and minimum benefit cost ratio ;

[0084] Then, the Monte Carlo method was used for risk diagnosis, calculating the overlapping demand, shortage quantity, and expected shortage cost at each time point based on the scenario set. Then calculate the service provider's cost-benefit ratio. and according to step size Updated prices If the price change is less than the tolerance The loop terminates if the loop terminates, otherwise the iteration count is incremented by 1; finally, the optimal price is returned. Capacity allocation Minimum benefit-cost ratio and the cost-benefit ratio of service providers By dynamically adjusting service prices through risk cost feedback, the system ensures that service providers meet cost-benefit ratio targets while avoiding a surge in risk costs due to excessive overbooking, thus achieving a dynamic balance between utilization, fairness, and reliability.

[0085] The beneficial effects of this invention are as follows: This invention provides a multi-dimensional collaborative community-shared energy storage optimization scheduling method. It introduces a relative diversity factor (RDF) to construct a dual constraint mechanism of "hard capacity upper limit + soft diversity threshold," quantifying the non-overlapping characteristics of user electricity consumption into adjustable risk constraints, ensuring that worst-case overlapping demand is always controlled within physical capacity. While improving energy storage utilization, it reduces the expected shortage risk, avoiding the problem of soaring grid procurement costs caused by simply pursuing utilization, thus achieving improved utilization without sacrificing reliability.

[0086] This invention designs a two-layer optimization architecture of "outer layer for price selection + inner layer for maximum-minimum allocation" to achieve a synergistic balance between service provider revenue and user fairness. The collaborative architecture of "outer layer for price selection + inner layer for maximum-minimum allocation" overcomes the bottleneck of independent pricing and allocation in existing technologies. The inner layer reduces the user B / C dispersion (standard deviation) by maximizing the minimum user benefit-cost ratio (B / C), ensuring a balanced benefit for all participating users. The outer layer dynamically adjusts service prices to match the service provider's target B / C, avoiding supply-demand imbalances caused by a single-goal orientation and increasing user participation. This satisfies the service provider's profit needs while increasing user participation through flexible pricing. User participation is higher than traditional solutions in scenarios with limited capacity, resolving the contradiction in existing technologies where either service providers struggle to profit or user churn rates are high.

[0087] This invention integrates Monte Carlo risk diagnosis with a piecewise linear billing-saving proxy model, constructing a closed-loop process of "constraint modeling - optimized scheduling - risk diagnosis." By simulating overlapping demand and physical shortages at various times using a historical electricity consumption scenario database, it accurately quantifies expected shortage costs and feeds this information back to the price adjustment process. The piecewise linear proxy model avoids the complex parameter assumptions of traditional optimization methods, requiring only historical user load data for deployment, thus improving solution efficiency and adapting to different scale scenarios such as building energy communities and park-level microgrids. Simultaneously, the model is compatible with various tariff types, including peak-valley time-of-use pricing and demand pricing, improving the average user B / C and service provider B / C compared to traditional solutions, lowering the technical threshold for commercial application.

[0088] This invention features strong data-driven adaptability and covers optimization across all capacity scenarios. The solution does not rely on complex parameter calibration; it dynamically adjusts constraint thresholds and scheduling strategies based solely on historical user load data, adapting to different ESS capacity scenarios. When ESS capacity is sufficient (≥200kWh), performance is on par with traditional optimal solutions. When capacity is tight, a diversity-priority allocation logic accommodates more users with complementary electricity consumption characteristics, ensuring overall system efficiency while expanding the coverage of shared energy storage services. This provides more flexible technical support for the large-scale promotion of community shared energy storage. Attached Figure Description

[0089] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0090] Figure 1 This is a schematic diagram of the overall structure of a multi-dimensional collaborative community-shared energy storage optimization scheduling method.

[0091] Figure 2 This is a schematic diagram of the structure after each borehole is separated, representing a multi-dimensional collaborative community-shared energy storage optimization scheduling method. Detailed Implementation

[0092] The present invention will now be further described with reference to the accompanying drawings.

[0093] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0094] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0095] like Figure 1 As shown, the present invention provides a multi-dimensional collaborative community-shared energy storage optimization scheduling method, comprising the following steps:

[0096] Step S1: Establish a dual-constraint scheduling model for community-shared energy storage systems based on the Relative Diversity Factor (RDF). Combine the hard upper limit of total subscriptions with the soft data-driven threshold based on the RDF. Utilize the non-overlapping nature of user loads to define the boundary of overlapping demands, thereby allowing over-booking of virtual capacity under limited physical capacity, and achieving a synergistic improvement in energy storage utilization and reliability.

[0097] Furthermore, in one embodiment, establishing a dual-constraint scheduling model for a community-shared energy storage system based on relative diversity factors includes establishing a service provider model and a community unit model;

[0098] In the service provider model, the service provider provides energy services to units within the building energy community. Sales of virtual reserved capacity (kWh) to provide shared energy storage (SES) services; shared energy storage SES is supported by physical energy storage systems (ESS), with an energy capacity of E (kWh); such as Figure 2 A shared energy storage (SES) system model for a Building Energy Community (BEC) is presented. The shared energy storage system comprises three entities: the Building Energy Community (BEC) coordinated by the Energy Community Service Provider (ECSP), the Energy Storage Service Provider (SP), and the power grid.

[0099] An energy community (ECSP) is a collection of building units coordinated by an energy community service provider (ECSP). The ECSP acts as a demand aggregator and energy regulator, coordinating energy transactions between community units and external resources (shared energy storage services, the public grid). After the pricing and terms of the shared energy storage service are published, each unit can subscribe to a certain amount of virtual reserved capacity. The Energy Community Service Provider (ECSP) is responsible for maintaining subscription registration information and calculating the operational intentions (charge / discharge requests) of each unit based on the goals and constraints of shared energy storage. These intentions are then synchronized with energy storage service providers (SPs). The SPs are responsible for deploying and operating physical energy storage systems with energy capacity (E) and charge / discharge power limitations, and providing virtual shared energy storage services to the building energy community based on service pricing. At each time step, the SPs aggregate the operational intentions of each unit obtained from the energy community service provider (ECSP) and transform them into feasible energy storage dispatch schemes that comply with equipment operating constraints (power, state of charge, efficiency). The public grid is responsible for maintaining system energy balance and providing the electricity pricing framework.

[0100] In the community unit model, community units subscribe to shared energy storage (SES), obtain bill savings through peak shaving and valley filling, and pay for reserved services.

[0101] Furthermore, in one embodiment, establishing the service provider model includes:

[0102] make (USD / kWh) The term "cycle" represents the levelized cost of energy storage (LCOS) for each billing cycle, calculated from the energy storage costs including hardware (power conversion system, storage media) and software / O&M. The capacity cost model for energy storage service providers (SPs) is as follows:

[0103]

[0104] The revenue of energy storage service providers (SPs) is calculated as follows:

[0105]

[0106] in (USD / kWh) (Period) refers to the published service price. It is a collection of units that participate in the service. ;

[0107] If overbooking is allowed ( Concurrent requests may exceed energy capacity. When the total concurrent requests from all users When it exceeds E, the energy storage service provider SP will supply the grid at a price Purchase difference in electricity volume To fill the gap The levelized price for the shortage; the expected shortage cost within a billing cycle is:

[0108]

[0109] in The time interval is The observation period;

[0110] The benefit-cost ratio (B / C) for energy storage service providers (SPs) is:

[0111]

[0112] Furthermore, in one embodiment, establishing the community unit model includes:

[0113] make For unit In time The baseline load (kW). Electricity demand (USD / kW) The energy price (USD / kW) is the base bill:

[0114]

[0115] make (kW) represents the unit In time Shared energy storage SES net energy change, It is a unit Energy efficiency ( Using the standard one-way splitting method:

[0116]

[0117] unit Shared Energy Storage (SES) Action Subscribed SES capacity limit;

[0118] The billing after using the shared energy storage SES is as follows:

[0119]

[0120] The benefit of the unit is cost savings. For the base bill;

[0121]

[0122] In the above formula This represents the original peak electricity load of user i across all time steps t. The above formula includes the reduction of electricity consumption during peak hours and the adjustment of electricity consumption time.

[0123] Under the shared energy storage SES tariff with only available capacity, the cost of the unit is:

[0124] (9)

[0125] The benefit-cost ratio (B / C) at the unit level is:

[0126]

[0127] The total cost of each unit capacity equals the revenue of the energy storage service provider (SP), i.e. .

[0128] Step S2: Construct a two-layer optimized scheduling architecture with outer-layer pricing and inner-layer capacity allocation. The outer loop adjusts the release price based on the service provider's survivability index. Under the premise of meeting the hard upper limit and differentiation rules, the inner loop introduces a piecewise linear proxy model to allocate capacity fairly with the goal of maximizing the minimum user revenue-cost ratio.

[0129] In traditional SES (Search Engine Sales), the one-to-one capacity configuration per unit often results in idle capacity because not all users need to store simultaneously. This lack of user diversity leads to low utilization of physical assets. Overbooking (i.e., selling virtual capacity exceeding the physical ESS capacity without adding hardware) statistically leverages these non-overlapping demands, improving utilization and SP (Service Provider) revenue. SP revenue is precisely the capacity fee per unit, so more aggressive pricing can support more user participation; therefore, this invention uses... As a primary lever to adjust virtual reserved capacity This section clarifies the logical connection between overbooking and user diversity, explaining that quantifying diversity through RDF can achieve a synergy between improved utilization and controllable risk, providing theoretical support for the subsequent construction of constraint models.

[0130] To link diversity with overbooking, this invention uses a relative diversity factor (RDF), which almost perfectly coincides with the limit of overlapping demand. In this way, user diversity can be mapped to a transparent cap on total subscriptions. This provides a consistently auditable link between utilization and risk, and naturally integrates with maximum-minimum benefit-cost ratio allocation and pricing selection.

[0131] As a benchmark, we consider overbooking implemented with a hard overbooking cap, where total virtual capacity cannot exceed a fixed multiple of physical ESS capacity. This captures the coverage fairness trade-off without considering diversity.

[0132] Furthermore, in one embodiment, the construction of the two-layer optimized scheduling architecture of outer-layer price setting and inner-layer capacity allocation includes establishing a baseline optimization framework without diversity constraints, providing a comparative basis for the subsequent introduction of RDF soft constraints, specifically including:

[0133] Given the published service price and participating units Allocate capacity ;unit The benefit is its cost savings. Its benefit-cost ratio is ;

[0134] make To achieve the target overbooking rate, the benchmark imposes a strict cap on the total number of subscriptions, as shown below:

[0135]

[0136] In equation (11), when Overbooking is not allowed at this time; with The increase in coverage leads to increased coverage, but fairness and reliability are limited.

[0137] While maintaining unit limits and overbooking caps, maximizing the benefit of the smallest unit is directly proportional, as shown below:

[0138]

[0139] The objective function in equation (12) To maximize the minimum unit revenue-cost ratio, where e is the virtual energy storage capacity subscription vector for community users, the constraints are as follows: For the provider's total revenue, For the total cost of the provider, Save money on user i's bill. For user i's subscription cost, Is it accepting pricing? The set of participating users This represents the maximum subscription capacity limit for user i. For the target oversold ratio ( (No overselling), E is the actual capacity of the physical energy storage system; the first and second constraints enforce fairness: guaranteeing the minimum benefit-cost ratio. Implicitly satisfying individual rationality (Right now );limit This reflects the operational / contractual constraints in the third constraint. The fourth constraint limits the overbooking cap.

[0140] Furthermore, in one embodiment, the construction of the two-layer optimized scheduling architecture for outer-layer price setting and inner-layer capacity allocation further includes:

[0141] The baseline model (P0) uses a hard overbooking cap. The total subscription limit is limited; however, it ignores the fact that peak user demand periods rarely coincide. Therefore, a perceived diversity overbooking rule is added, where overbooking increases with the increase in non-overlapping periods. To match hardware with overlapping peaks, for shared energy storage (SES) planning, this is based on the user's historical capacity. Weight their demand curves and use the Relative Diversity Factor (RDF); from recent history, for each unit Extracting standardized demand curves ,in Given a subscription vector The definition of RDF is as follows:

[0142]

[0143] Where the denominator is approximately The implied worst-case concurrent requests and the observed curve shape. Intuitively, a higher RDF indicates greater multi-tenancy diversity.

[0144] The following formula replaces the hard overbooking limit with a soft overbooking constraint based on the diversity factor DF. DF is an indicator used to measure the asynchronous nature of user loads. Its value is equal to the ratio of the sum of the non-overlapping peak loads of all users to the overlapping peak loads of the entire user group during the same period. DF is always greater than 1. The larger the value, the less synchronous the peak electricity consumption of users.

[0145]

[0146] And add a DF-based soft overbooking constraint:

[0147]

[0148] Therefore, only when diversity is sufficient to keep the overlap of the worst-case scenario within acceptable limits... Only when it reaches this point can the total number of subscribers approach [a certain number]. In practice, a conservative threshold is applied. Estimated from the duration of time, and requiring This part breaks through the single constraint of traditional hard capacity limit and constructs a dual constraint of hard capacity limit + soft diversity threshold, ensuring that overbooking is only implemented when diversity is sufficient to control risk, taking into account both utilization and reliability.

[0149] Price announcement through a two-tiered collaborative design process. and subscription distribution The inner layer, under the hard overbooking cap and the DF-based soft overbooking constraint, has a fixed... Capacity is allocated to maximize the minimum unit benefit-cost ratio (B / C). The outer layer, while maintaining fairness, adjusts the B / C feasibility based on SP. This part implements a two-layer optimization architecture: an outer layer for price selection and an inner layer for maximum and minimum allocation, clearly defining the collaborative logic between the two layers. This ensures fairness for users while also meeting the profit needs of service providers, avoiding an imbalance between supply and demand.

[0150] Furthermore, in one embodiment, the outer loop adjusts the release price based on the service provider's viability index, while the inner loop, under the premise of satisfying hard caps and differentiation rules, introduces a piecewise linear proxy model to fairly allocate capacity with the goal of maximizing the minimum user revenue-cost ratio, including:

[0151] Given the published service price The inner layer will include subscription volume Assign to users, using the bill saving function and capacity fees To maximize the minimum B / C, as shown below:

[0152]

[0153] In equation (16), the condition for user participation is ;make In practice, It exhibits diminishing returns because as As the amount increases, marginal bill savings decrease; therefore, we use a piecewise linear convex approximation constructed from the bill simulation. To ensure continuity.

[0154] User capacity is allocated to maintain a minimum B / C ratio, i.e. As shown below:

[0155]

[0156] Satisfy the unit-by-unit boundary, hard overbooking cap, and DF-based soft overbooking constraints:

[0157]

[0158] Due to the first constraint and It's a multiplicative relationship; the model (P1) is not a direct linear programming (LP) problem. Through... Solve using a simple one-dimensional search (such as binary search); fix Then, the constraints in the model (P1) will be transformed into a small LP; if If feasible, increase Otherwise reduce Default settings This satisfies the individual rationality (IR) condition of equation (16). This part refines the max-min allocation inner ring of step 2, clarifying the capacity allocation algorithm and constraints under a fixed price.

[0159] The outer layer selects a published price to ensure that the service is feasible for the service provider while maintaining fairness; in order to refine the outer ring of price selection in step 2, the basis and strategy for price adjustment are clarified, and the Monte Carlo risk diagnosis in step 3 is also connected.

[0160] Inner layer solution The service provider's benefit-cost ratio (B / C) is calculated as follows:

[0161]

[0162] in Risk costs are quantified using the Monte Carlo method; based on historical days / scenarios. Generate concurrent requests , A scene-specific curve;

[0163] While maintaining a high degree of fairness, Feasibility of updating to the target service provider The following strategy is adopted:

[0164]

[0165] in It is a small step size; if Then improve Otherwise, it can be reduced. To enhance user engagement and .

[0166] Step S3: Combining the risk diagnosis mechanism of Monte Carlo simulation, we map diversity to the expected shortage cost, establish an assessment method that maps user differences to the expected shortage cost, and realize that while improving the utilization rate of the community shared energy storage system, we can effectively control the shortage risk and ensure the economic benefits of users and service providers. At the same time, we can take into account the three core objectives of energy storage utilization rate, user fairness and service provider revenue, and achieve multi-dimensional collaborative optimization.

[0167] Furthermore, in one embodiment, the risk diagnosis mechanism combining Monte Carlo simulation, which maps diversity to anticipated shortage costs, and the assessment method that maps user differences to anticipated shortage costs, includes:

[0168] Construct a random set of scenarios and calculate the scenario-based physical shortage;

[0169] Assess the expected shortage risk costs and establish a two-tier overbooking allocation and a DF-based inner allocation algorithm.

[0170] Furthermore, in one embodiment, constructing a random scene set and calculating the scene-based physical shortage includes:

[0171] Extract features from the historical activity data of community users to construct a system containing... A collection of independent historical scenes For each scenario in the set Get each user In time Normalized concurrency curve The normalized concurrency curve reflects the load behavior pattern of users at a specific historical moment; by constructing a scenario set through historical data, a data foundation that fits the actual electricity consumption behavior is provided for subsequent shortage simulation;

[0172] Optimal capacity allocation scheme based on inner loop output Simulation calculations in each scenario The concurrent request situation under the following scenario; for the scenario time The physical shortage path is calculated as follows:

[0173]

[0174] In the formula Let h be the physical shortage amount (kWh) under scenario h and time t. Let h represent the concurrent request volume (kWh) of user i under scenario h and time t. The physical capacity of the shared energy storage system is defined as follows: if the aggregated request volume is less than the physical capacity, the shortage is 0; if it exceeds the physical capacity, the excess portion is the physical shortage. To address the problem that traditional oversubscription strategies cannot accurately predict overlapping demand gaps, this method quantifies the physical shortage scale for each scenario and at each moment, clarifies the specific shortage level when the total subscription volume exceeds the physical capacity, and provides accurate data support for risk cost accounting.

[0175] Furthermore, in one embodiment, the assessment of anticipated shortage risk costs and the establishment of a two-tier overbooking allocation and a DF-based inner-layer allocation algorithm include:

[0176] For scene set Statistical analysis was performed on the physical shortages across all scenarios, and diagnostic indicators were calculated for each time step:

[0177]

[0178] Combined with the real-time electricity purchase price of the power grid or the preset power shortage penalty price Calculate the expected risk cost over the entire scheduling cycle, transforming the abstract shortage risk into a concrete and calculable economic cost:

[0179]

[0180] The two-level overbooking allocation and the DF-based inner allocation algorithm are as follows:

[0181] First, enter the physical capacity. Overbooking price DF curve set Piecewise linear net benefit function Breakpoints and slopes, maximum booking volume SP Costs The set of levelized prices in shortage Maximum number of iterations Step size sequence set Initial price And let the number of iterations be ;

[0182] Then when When entering the main loop, first determine the set of participating units. Units that satisfy the condition that cost savings exceed expenses are then optimized using an inner-level binary search method. :set up upper and lower boundaries and In the above formula, It is the initial upper bound of the binary search. For all users and its segmentation Take the maximum value. Corresponding user Corresponding piecewise linear net benefit function The Slope, It is the first The system in the next outer iteration charges users a unit price for pre-booked capacity; when the difference between the upper and lower bounds is greater than the tolerance. When, take the median value. And solve the linear programming LP feasibility problem, with constraints including the user benefit-cost ratio not being less than Unit capacity is between 0 and the upper limit, and total subscriptions do not exceed the hard oversubscription limit. In the worst-case scenario, overlapping demand should not exceed physical capacity; if LP (Limited Limit) is feasible, then update. And store candidate capacity Otherwise update ,until Convergence, yielding the optimal capacity. and minimum benefit cost ratio ;

[0183] Then, the Monte Carlo method was used for risk diagnosis, calculating the overlapping demand, shortage quantity, and expected shortage cost at each time point based on the scenario set. Then calculate the service provider's cost-benefit ratio. and according to step size Updated prices If the price change is less than the tolerance The loop terminates if the loop terminates, otherwise the iteration count is incremented by 1; finally, the optimal price is returned. Capacity allocation Minimum benefit-cost ratio and the cost-benefit ratio of service providers By dynamically adjusting service prices through risk cost feedback, the system ensures that service providers meet cost-benefit ratio targets while avoiding a surge in risk costs due to excessive overbooking, thus achieving a dynamic balance between utilization, fairness, and reliability.

[0184] In summary, this invention presents a multi-dimensional collaborative community-shared energy storage optimization scheduling method. It introduces a relative diversity factor (RDF) to construct a dual-constraint mechanism of "hard capacity cap + soft diversity threshold," quantifying the non-overlapping characteristics of user electricity consumption into adjustable risk constraints. This ensures that worst-case overlapping demand is always controlled within physical capacity. While improving energy storage utilization, it reduces the expected shortage risk, avoiding the surge in grid procurement costs caused by simply pursuing utilization rates, thus achieving improved utilization without sacrificing reliability.

[0185] This invention designs a two-layer optimization architecture of "outer layer for price selection + inner layer for maximum-minimum allocation" to achieve a synergistic balance between service provider revenue and user fairness. The collaborative architecture of "outer layer for price selection + inner layer for maximum-minimum allocation" overcomes the bottleneck of independent pricing and allocation in existing technologies. The inner layer reduces the user B / C dispersion (standard deviation) by maximizing the minimum user benefit-cost ratio (B / C), ensuring a balanced benefit for all participating users. The outer layer dynamically adjusts service prices to match the service provider's target B / C, avoiding supply-demand imbalances caused by a single-goal orientation and increasing user participation. This satisfies the service provider's profit needs while increasing user participation through flexible pricing. User participation is higher than traditional solutions in scenarios with limited capacity, resolving the contradiction in existing technologies where either service providers struggle to profit or user churn rates are high.

[0186] This invention integrates Monte Carlo risk diagnosis with a piecewise linear billing-saving proxy model, constructing a closed-loop process of "constraint modeling - optimized scheduling - risk diagnosis." By simulating overlapping demand and physical shortages at various times using a historical electricity consumption scenario database, it accurately quantifies expected shortage costs and feeds this information back to the price adjustment process. The piecewise linear proxy model avoids the complex parameter assumptions of traditional optimization methods, requiring only historical user load data for deployment, thus improving solution efficiency and adapting to different scale scenarios such as building energy communities and park-level microgrids. Simultaneously, the model is compatible with various tariff types, including peak-valley time-of-use pricing and demand pricing, improving the average user B / C and service provider B / C compared to traditional solutions, lowering the technical threshold for commercial application.

[0187] This invention features strong data-driven adaptability and covers optimization across all capacity scenarios. The solution does not rely on complex parameter calibration; it dynamically adjusts constraint thresholds and scheduling strategies based solely on historical user load data, adapting to different ESS capacity scenarios. When ESS capacity is sufficient (≥200kWh), performance is on par with traditional optimal solutions. When capacity is tight, a diversity-priority allocation logic accommodates more users with complementary electricity consumption characteristics, ensuring overall system efficiency while expanding the coverage of shared energy storage services. This provides more flexible technical support for the large-scale promotion of community shared energy storage.

[0188] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0189] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A multi-dimensional collaborative community-shared energy storage optimization scheduling method, characterized in that, Includes the following steps: A dual-constraint scheduling model for community-shared energy storage systems based on relative diversity factors is established, which combines the hard upper limit of total subscriptions with the soft data-driven threshold based on relative diversity factors, and uses the non-overlapping nature of user loads to define the boundary of overlapping demand. A two-layer optimized scheduling architecture is constructed, consisting of an outer layer for pricing and an inner layer for capacity allocation. The outer loop adjusts the release price based on the service provider's survivability index, while the inner loop introduces a piecewise linear proxy model to allocate capacity fairly with the goal of maximizing the minimum user revenue-cost ratio, while meeting hard upper limits and differentiation rules. By combining the risk diagnosis mechanism of Monte Carlo simulation, we map diversity to expected shortage costs and establish an assessment method that maps user differences to expected shortage costs.

2. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 1, characterized in that, The establishment of a dual-constraint scheduling model for a community-shared energy storage system based on relative diversity factors includes establishing a service provider model and a community unit model; In the service provider model, the service provider provides energy services to units within the building energy community. Sales of virtual reserved capacity To provide shared energy storage services; Shared Energy Storage (SES) is supported by physical energy storage systems with an energy capacity of E; Shared Energy Storage Systems include three types of entities: building energy communities coordinated by energy community service providers, energy storage service providers, and the power grid; In the community unit model, community units subscribe to shared energy storage, gaining bill savings through peak shaving and valley filling, while paying for reserved services.

3. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 2, characterized in that, The establishment of the service provider model includes: make The levelized cost of energy storage (LCOE) represents the cost per billing cycle. The capacity cost model for energy storage service providers is as follows: The revenue of energy storage service providers is calculated as follows: in The published service prices, It is a collection of units that participate in the service. ; When the total concurrent requests from all users When it exceeds E, the energy storage service provider SP will supply the grid at a price Purchase difference in electricity volume To fill the gap The levelized price for the shortage; the expected shortage cost within a billing cycle is: in The time interval is The observation period; The cost-benefit ratio of energy storage service providers is: 。 4. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 3, characterized in that, The establishment of the community unit model includes: make For unit In time The baseline load, To meet the electricity demand, For energy prices, the benchmark bill is: make Representation unit In time The net change in shared energy storage It is a unit Energy efficiency, using a standard one-way splitting method: The bill after using shared energy storage is as follows: The benefit of the unit is cost savings. For the base bill; In the above formula This represents the peak value of the original electrical load for user i across all time steps t; Under the shared energy storage SES tariff with only available capacity, the cost of the unit is: (9) The benefit-cost ratio at the unit level is: The total cost of each unit capacity equals the revenue of the energy storage service provider, i.e. .

5. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 1, characterized in that, The construction of the two-layer optimized scheduling architecture, consisting of outer-layer price setting and inner-layer capacity allocation, includes establishing a baseline optimization framework without diversity constraints. This provides a comparative basis for the subsequent introduction of RDF soft constraints, specifically including: Given the published service price and participating units Allocate capacity ;unit The benefit is its cost savings. Its benefit-cost ratio is ; make To achieve the target overbooking rate, the benchmark imposes a strict cap on the total number of subscriptions, as shown below: While maintaining unit limits and overbooking caps, maximizing the benefit of the smallest unit is directly proportional, as shown below: The objective function in equation (12) To maximize the minimum unit revenue-cost ratio, where e is the virtual energy storage capacity subscription vector for community users, the constraints are as follows: For the provider's total revenue, For the total cost of the provider, Save money on user i's bill. For user i's subscription cost, Is it accepting pricing? The set of participating users This represents the maximum subscription capacity limit for user i. The target oversold ratio is E, and the actual capacity of the physical energy storage system is E.

6. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 5, characterized in that, The two-layer optimized scheduling architecture, which combines outer-layer price setting and inner-layer capacity allocation, also includes: For shared energy storage (SES) planning, based on the user's historical capacity Weight their demand curves and use the Relative Diversity Factor (RDF); from recent history, for each unit Extracting standardized demand curves ,in Given a subscription vector The definition of RDF is as follows: The following formula replaces the hard overbooking limit with a soft overbooking constraint based on the diversity factor DF. DF is an indicator used to measure the asynchronous nature of user loads. Its value is equal to the ratio of the sum of the non-overlapping peak loads of all users to the overlapping peak loads of the entire user group during the same period. DF is always greater than 1. The larger the value, the less synchronous the peak electricity consumption of users. And add a DF-based soft overbooking constraint: 。 7. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 6, characterized in that, The outer loop adjusts the release price based on the service provider's viability index. The inner loop, while meeting hard caps and differentiation rules, introduces a piecewise linear proxy model to fairly allocate capacity with the goal of maximizing the minimum user revenue-cost ratio. This includes: Given the published service price The inner layer will include subscription volume. Assign to users, using the bill saving function and capacity fees To maximize the minimum B / C, as shown below: User capacity is allocated to maintain a minimum B / C ratio, i.e. As shown below: Satisfy the unit-by-unit boundary, hard overbooking cap, and DF-based soft overbooking constraints: Inner layer solution The service provider's benefit-cost ratio (B / C) is calculated as follows: in It is the risk cost quantified using the Monte Carlo method; While maintaining a high degree of fairness, Feasibility of updating to the target service provider The following strategy is adopted: in It is a small step size; if Then improve Otherwise, it can be reduced. To enhance user engagement and .

8. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 1, characterized in that, The risk diagnosis mechanism combined with Monte Carlo simulation maps diversity to expected shortage costs, and the assessment method that maps user differences to expected shortage costs includes: Construct a random set of scenarios and calculate the scenario-based physical shortage; Assess the expected shortage risk costs and establish a two-tier overbooking allocation and a DF-based inner allocation algorithm.

9. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 8, characterized in that, The process of constructing a random scene set and calculating the scene-based physical shortage includes: Extract features from the historical activity data of community users to construct a system containing... A collection of independent historical scenes For each scenario in the set Get each user In time Normalized concurrency curve ; Optimal capacity allocation scheme based on inner loop output Simulation calculations in each scenario Concurrent request scenarios; for example time The physical shortage path is calculated as follows: In the formula The physical shortage quantity under scenario h and time t. Let h be the number of concurrent requests from user i under scenario h and time t. The physical capacity of the shared energy storage system is defined as follows: if the aggregated demand is less than the physical capacity, the shortage is 0; if it exceeds the physical capacity, the excess portion is the physical shortage.

10. The multi-dimensional collaborative community-shared energy storage optimization scheduling method as described in claim 9, characterized in that, The assessment of anticipated shortage risk costs and the establishment of a two-tier overbooking allocation and a DF-based inner allocation algorithm include: For scene set Statistical analysis was performed on the physical shortages across all scenarios, and diagnostic indicators were calculated for each time step: Combined with the real-time electricity purchase price of the power grid or the preset power shortage penalty price Calculate the expected risk cost over the entire scheduling cycle, transforming the abstract shortage risk into a concrete and calculable economic cost: The two-level overbooking allocation and the DF-based inner allocation algorithm are as follows: First, enter the physical capacity. Overbooking prices DF curve set Piecewise linear net benefit function Breakpoints and slopes, maximum booking volume SP Costs The set of levelized prices in shortage Maximum number of iterations Step size sequence set Initial price And let the number of iterations be ; Then when When entering the main loop, first determine the set of participating units. Units that satisfy the condition that cost savings exceed expenses are then optimized using an inner-level binary search method. :set up upper and lower boundaries and In the above formula, It is the initial upper bound of the binary search. For all users and its segmentation Take the maximum value. Corresponding user Corresponding piecewise linear net benefit function The Slope, It is the first The system in the next outer iteration charges users a unit price for pre-booked capacity; when the difference between the upper and lower bounds is greater than the tolerance. When, take the median value. And solve the linear programming LP feasibility problem, with constraints including the user benefit-cost ratio not being less than Unit capacity is between 0 and the upper limit, and total subscriptions do not exceed the hard oversubscription limit. In the worst-case scenario, overlapping demand should not exceed physical capacity; if LP (Limited Limit) is feasible, then update. And store candidate capacity Otherwise update ,until Convergence, yielding the optimal capacity. and minimum benefit cost ratio ; Then, the Monte Carlo method was used for risk diagnosis, calculating the overlapping demand, shortage quantity, and expected shortage cost at each time point based on the scenario set. Then calculate the service provider's cost-benefit ratio. and according to step size Updated prices If the price change is less than the tolerance The loop terminates if the loop terminates, otherwise the iteration count is incremented by 1; finally, the optimal price is returned. Capacity allocation Minimum benefit-cost ratio and the cost-benefit ratio of service providers .

Images