A Distributed Energy Storage Dispatch Method and System Based on Multi-Energy Collaboration

Abstract

This application relates to the field of energy dispatching, and in particular to a method and system for distributed energy storage dispatching based on multi-energy collaboration. The method includes: acquiring a set of factory order information; based on this, extracting and fusing multi-dimensional production value features to generate a dynamic production value assessment information set; based on this, combined with preset energy cost constraints, calculating dynamic electricity value thresholds and making global energy dispatching strategy decisions to generate an energy dispatching strategy information set; acquiring a set of energy supply influencing factors; based on the energy dispatching strategy information set and the energy supply influencing factors set, performing multi-timescale rolling optimization and global preset energy allocation to generate an energy allocation execution instruction set; based on this, issuing and executing dispatching instructions and dynamic production compensation for production lines and energy storage layers, generating and outputting a multi-energy collaborative dispatching report. This application can improve the economic efficiency of factories during energy dispatching.

Description

Technical Field

[0001] This application relates to the field of energy dispatch, and in particular to a method and system for distributed energy storage dispatch based on multi-energy coordination. Background Technology

[0002] In the fields of energy management and intelligent manufacturing, building a green and flexible industrial energy system dominated by a high proportion of renewable energy is the core path to achieving the "dual carbon" goal. Its operational stability and economic efficiency are directly related to the energy costs of enterprises, the reliable execution of production plans, and the overall carbon emission reduction effect. It is a key infrastructure of the modern green and intelligent manufacturing system.

[0003] However, existing industrial energy dispatch methods lack multi-dimensional, real-time value coordination and global optimization mechanisms when facing the inherent strong volatility and randomness of wind and solar power generation. This not only makes it difficult to effectively smooth out power imbalances on a minute / second time scale, but may also cause production interruptions, energy waste or grid penalties under extreme power shortage or surplus conditions, thus constituting a dual bottleneck in system economy and reliability. Summary of the Invention

[0004] This application provides a distributed energy storage scheduling method and system based on multi-energy collaboration to solve the above-mentioned technical problems.

[0005] Firstly, this application provides a distributed energy storage scheduling method based on multi-energy collaboration. The method includes: acquiring a set of factory order information; performing multi-dimensional production value feature extraction and fusion analysis based on the set of factory order information to generate a dynamic production value assessment information set; calculating dynamic electricity value thresholds and making global energy scheduling strategy decisions based on the dynamic production value assessment information set and preset energy cost constraints to generate an energy scheduling strategy information set; acquiring a set of energy supply influencing factors; performing multi-timescale rolling optimization and global preset energy allocation based on the energy scheduling strategy information set and the energy supply influencing factors to generate an energy allocation execution instruction set; and issuing and executing scheduling instructions and dynamic production compensation for production lines and energy storage layers based on the energy allocation execution instruction set to generate and output a multi-energy collaborative scheduling report.

[0006] The above technical solutions enable precise matching of production value and energy utilization, improving factory economic efficiency; dynamically controlling energy costs and reducing waste; enhancing the ability to cope with uncertainties and ensuring stable energy supply; maximizing the synergistic efficiency of production lines and energy storage layers; providing data support for management decisions, promoting the transformation of energy management towards data-driven approaches, and solving various pain points in traditional dispatching.

[0007] Optionally, generating the dynamic production value assessment information set includes: the factory order information set includes the single-item marginal profit, total energy consumption required for production, delivery deadline, and current production progress for each order; dividing the single-item marginal profit by the total energy consumption required for production to obtain the unit energy consumption profit for each order; calculating the reciprocal of the ratio of the remaining production time to the standard production line hours based on the delivery deadline and current production progress as the production urgency; and integrating the unit energy consumption profit and the production urgency of all orders to generate the dynamic production value assessment information set.

[0008] Optionally, generating the energy dispatch strategy information set includes: constructing a comprehensive evaluation index for order production priority based on the unit energy consumption profit and production urgency of each order, as the weighted value contribution of each order; sorting the weighted value contributions of all pending orders, and calculating a unified dynamic power value threshold for the entire plant based on the sorting results, wherein the dynamic power value threshold represents the upper limit of unit energy cost acceptable for maintaining production at the current moment; comparing the dynamic power value threshold with the real-time electricity price of the power grid, executing a global energy dispatch strategy decision based on the comparison results, and generating the energy dispatch strategy information set.

[0009] Optionally, the step of executing a global energy dispatch strategy decision based on the comparison result includes: when the dynamic electricity value threshold is greater than the real-time grid price, determining it to be a value-first mode and generating a production-guarantee dispatch strategy: prioritizing the dispatch of real-time renewable energy for production; when real-time renewable energy is insufficient, dispatching energy storage layer power to supplement it; when the energy storage layer power is still insufficient, purchasing power from the grid to ensure production continuity; when the dynamic electricity value threshold is less than or equal to the real-time grid price, determining it to be a cost-first mode and generating an economic optimization dispatch strategy: restricting production activities to use only real-time renewable energy; when there is a surplus of real-time renewable energy, storing it in the energy storage layer; when the grid price is in a valley, dispatching the energy storage layer power or purchasing power from the grid for production; when the grid price is in a peak period, controlling the energy storage layer to discharge to the grid to realize revenue.

[0010] Optionally, the generation of the energy allocation execution instruction set includes: the energy supply influencing factor set includes weather forecast information, historical energy data, and grid electricity price prediction data within a future preset time window; based on the weather forecast information and the historical energy data, photovoltaic power generation prediction sequences and wind power generation prediction sequences are generated respectively through physical equation calculation and power characteristic curve query; using the photovoltaic power generation prediction sequence, the wind power generation prediction sequence, and the grid electricity price prediction data as dynamic boundary conditions, the energy dispatch strategy information set is coupled to perform multi-time-scale rolling optimization calculations; based on the rolling optimization calculation results, the energy allocation execution instruction set is generated for controlling the charging and discharging of the energy storage layer, power allocation of the production line, and grid interaction in different future time periods.

[0011] Optionally, generating the photovoltaic power generation prediction sequence and the wind power generation prediction sequence respectively includes: calculating an initial photovoltaic power generation sequence based on the solar irradiance and ambient temperature in the weather forecast information using the photoelectric conversion physical equation; simultaneously, correcting the initial photovoltaic power generation sequence by coupling temperature loss and actual equipment attenuation coefficient to generate the photovoltaic power generation prediction sequence; obtaining a reference wind power generation sequence by querying the inherent power characteristic curve of the corresponding wind turbine model based on the wind speed in the weather forecast information; and simultaneously, correcting the wake loss of the reference wind power generation sequence by combining wind direction to generate the wind power generation prediction sequence.

[0012] Optionally, the multi-timescale rolling optimization calculation includes: constructing a three-layer collaborative optimization architecture comprising day-ahead optimization, intraday rolling optimization, and real-time adjustment; at the day-ahead optimization layer, based on the foreseeable future renewable energy generation and electricity price forecast data, coupled with the energy dispatch strategy information set, to formulate a baseline energy allocation plan with the goal of minimizing total energy costs and maximizing production value; at the intraday rolling layer, performing optimization on a rolling basis with fixed time windows, dynamically correcting the charging and discharging strategies and production power allocation for near-term periods in the baseline energy allocation plan based on the latest meteorological and electricity price data, and generating an adjusted energy allocation plan; at the real-time adjustment layer, detecting power deviations based on minute-level monitoring data, and making rapid compensation decisions for the adjusted energy allocation plan based on the dynamic electricity value threshold, thereby achieving the self-evolution of the strategy.

[0013] Optionally, generating and outputting the multi-energy collaborative scheduling report includes: parsing the energy allocation execution instruction set into specific equipment control instructions and distributing them to the production execution layer and the energy management layer respectively; monitoring the total production power balance in real time, and if an energy supply and demand imbalance is detected, initiating a preset compensation strategy based on the dynamic electricity value threshold, recording the deviation data and compensation effect, and feeding it back to the global energy scheduling strategy decision-making step; integrating the specific equipment control instructions and the deviation data and compensation effect to generate and output the multi-energy collaborative scheduling report.

[0014] Optionally, if an energy supply-demand imbalance is detected, a preset compensation strategy is initiated based on the dynamic electricity value threshold, including: when the deviation between the actual output of renewable energy and the predicted sequence exceeds a preset disturbance threshold, real-time rescheduling is immediately triggered; based on the dynamic electricity value threshold, all currently executing production tasks are divided into guaranteed tasks and optimized tasks; for guaranteed tasks, their energy supply is maintained, and the resulting power deficit is compensated by energy storage discharge or grid purchase; for optimized tasks, according to their value contribution ranking, the power of tasks ranked lower is dynamically reduced or they are temporarily suspended to adapt to the currently available energy, and the surplus energy is redistributed.

[0015] Secondly, this application provides a distributed energy storage scheduling system based on multi-energy collaboration. The system includes: a production value assessment module, used to acquire a set of factory order information, and based on the set of factory order information, to extract and fuse multi-dimensional production value features to generate a dynamic production value assessment information set; an energy scheduling strategy module, used to, based on the dynamic production value assessment information set and combined with preset energy cost constraints, to calculate dynamic electricity value thresholds and make global energy scheduling strategy decisions to generate an energy scheduling strategy information set; an energy allocation execution module, used to acquire a set of energy supply influencing factors, and based on the energy scheduling strategy information set and the set of energy supply influencing factors, to perform multi-timescale rolling optimization and global preset energy allocation to generate an energy allocation execution instruction set; and an energy collaborative scheduling module, used to, based on the energy allocation execution instruction set, issue and execute scheduling instructions and dynamic production compensation for production lines and energy storage layers, and generate and output a multi-energy collaborative scheduling report. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application 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 some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1This is a schematic diagram illustrating an application scenario provided in one embodiment of this application; Figure 2 A flowchart illustrating a distributed energy storage scheduling method based on multi-energy coordination provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a distributed energy storage scheduling system based on multi-energy collaboration, provided as an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0020] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0021] Existing industrial energy dispatch methods lack multi-dimensional, real-time value coordination and global optimization mechanisms when facing the inherent strong volatility and randomness of wind and solar power generation. This not only makes it difficult to effectively mitigate power imbalances on a minute / second time scale, but may also lead to production interruptions, energy waste, or grid penalties under extreme power shortages or surpluses, constituting a dual bottleneck in system economy and reliability.

[0022] Based on this, this application provides a distributed energy storage scheduling method and system based on multi-energy collaboration. First, it acquires a set of factory order information, extracts and integrates multi-dimensional production value characteristics to generate a dynamic production value assessment information set. Then, combined with preset energy cost constraints, it calculates a dynamic electricity value threshold, decides on a global energy scheduling strategy, and forms an energy scheduling strategy information set. Next, it acquires a set of energy supply influencing factors, performs multi-timescale rolling optimization and global preset allocation, and generates an energy allocation execution instruction set. Finally, it issues instructions to the production line and energy storage layer to execute dynamic production compensation, records data, generates a multi-energy collaborative scheduling report, and outputs it to factory management personnel. This method achieves precise matching of production value and energy utilization, improving factory economic efficiency; dynamically controls energy costs, reducing waste; enhances the ability to cope with uncertainties, ensuring stable energy supply; maximizes the collaborative efficiency of the production line and energy storage layer; provides data support for management decisions, promotes the transformation of energy management to data-driven approaches, and solves many pain points of traditional scheduling.

[0023] Figure 1 This diagram illustrates an application scenario provided by this application. In the energy dispatching process, the method provided in this application can achieve precise matching between production value and energy utilization, thereby improving the economic efficiency of the factory.

[0024] Specifically, the method of this application is applied to any server that communicates with the factory order management system and the factory energy monitoring equipment. The server obtains the factory order information set provided by the factory order management system and the energy supply influencing factor set provided by the factory energy monitoring equipment. First, it acquires the factory order information set, extracts multi-dimensional production value characteristics, and performs fusion analysis to generate a dynamic production value assessment information set. Then, combined with preset energy cost constraints, it calculates the dynamic electricity value threshold, decides on the global energy dispatch strategy, and forms an energy dispatch strategy information set. Next, it acquires the energy supply influencing factor set, performs multi-timescale rolling optimization and global preset allocation, and generates an energy allocation execution instruction set. Finally, it issues instructions to the production line and energy storage layer to execute dynamic production compensation, records data, generates a multi-energy collaborative dispatch report, and outputs it to factory management personnel. Specific implementation methods can be found in the following embodiments.

[0025] Figure 2 This is a flowchart illustrating a distributed energy storage scheduling method based on multi-energy collaboration, provided as an embodiment of this application. The method of this embodiment can be applied to servers in the above scenarios. For example... Figure 2 As shown, the method includes: S201. Obtain the factory order information set, and based on the factory order information set, perform multi-dimensional production value feature extraction and fusion analysis to generate a dynamic production value assessment information set.

[0026] The factory order information set can be a complete collection of information related to various production orders received by the factory, covering the marginal profit of individual products, total energy consumption required for production, delivery deadlines, and current production progress. The data originates from the Factory Order Management System (OMS). Multi-dimensional production value characteristics can be a set of key indicators characterizing the core value attributes during the order production process, specifically including order priority, economic value, time value, and energy consumption value dimensions. The dynamic production value assessment information set can be a dynamic information set formed by extracting, integrating, and quantitatively analyzing multi-dimensional production value characteristics. It provides value-oriented guidance for energy dispatching strategy formulation and is generated by the production value assessment service after data processing.

[0027] Specifically, against the backdrop of industrial intelligent transformation, factory production models have shifted from traditional large-scale standardized production to multi-variety, small-batch, and customized production. Dynamic changes in orders have become the norm—urgent orders are frequently inserted, order quantities are adjusted, and delivery cycles are changed, directly leading to significant dynamic fluctuations in production value. However, traditional factory energy dispatching models have fundamental flaws: First, the dispatching logic relies on historical production data or fixed capacity planning, treating all production activities as homogeneous demand and adopting a uniform energy allocation standard, completely ignoring the value differences between different orders. This leads to high-value urgent orders potentially being delayed due to insufficient energy supply, while low-value orders consume a large amount of energy resources, resulting in a double loss of production value and energy waste. Second, traditional dispatching lacks the ability to dynamically perceive production value. When the order structure changes (such as the sudden insertion of high-profit urgent orders), it cannot adjust the energy allocation orientation in a timely manner, still executing according to the original static strategy, resulting in a serious disconnect between energy resources and production value objectives. Third, the evaluation dimension of production value is singular. The traditional model only focuses on the quantity or delivery deadline of orders, without considering key factors such as economic value and energy consumption adaptability, resulting in a lack of comprehensive value support for energy allocation decisions. This step obtains order data in real time from the factory order management system, extracts multi-dimensional features such as economic, time, and resource dimensions, and performs weighted fusion analysis through a rule engine or machine learning model to dynamically generate an evaluation information set including priority scores and resource suggestions, providing a data foundation for energy dispatching. Through this step, the factory can transform the production value of orders into quantifiable evaluation indicators, making energy scheduling more aligned with actual production needs, thereby improving resource utilization efficiency.

[0028] S202. Based on the dynamic evaluation information set of production value and combined with the preset energy cost constraints, perform dynamic electricity value threshold calculation and global energy dispatch strategy decision-making to generate an energy dispatch strategy information set.

[0029] Preset energy cost constraints can be a set of constraint rules based on the costs of the entire process of energy procurement, storage, and use in a factory. These include real-time grid price constraints, renewable energy procurement / generation cost constraints, energy storage device charging and discharging cost constraints, energy transmission loss cost constraints, and energy waste penalty cost constraints. Dynamic electricity value thresholds can be critical values ​​for electricity use value dynamically calculated based on a dynamic production value assessment information set and preset energy cost constraints. These thresholds are used to determine the rationality and priority of electricity resource allocation. Global energy dispatch strategy decisions can be comprehensive energy allocation and control strategies formulated for the overall production activities and energy system of the factory. These strategies cover core aspects such as energy supply source allocation, energy allocation priority across production lines, energy storage device charging and discharging scheduling rules, and energy usage time optimization, aiming to achieve the dual goals of maximizing production value and minimizing energy costs. The energy dispatch strategy information set can be a set of executable strategies formed by refining and quantifying the global energy dispatch strategy, providing a clear basis for subsequent energy allocation execution.

[0030] Specifically, the proportion of energy costs in total production costs in factories has been rising year by year, and energy cost control has become one of the core factors affecting corporate profitability. Traditional factory energy dispatching has significant shortcomings in cost control and strategy formulation: First, energy cost constraints are rigid. Traditional models use fixed cost standards (such as setting energy consumption budgets based on monthly average electricity prices) without adjusting for dynamic factors such as real-time market energy price fluctuations and changes in renewable energy output. This results in the allocation of grid power according to the original quota during peak electricity price periods, leading to high costs, or the failure to fully utilize low-cost energy for energy storage during off-peak electricity price periods, missing cost optimization opportunities. Second, there is a lack of dynamic electricity value threshold guidance, making it impossible to determine whether the use of electricity in different production stages and for different orders is economically reasonable. This leads to some low-value production stages using high-priced electricity, resulting in a waste of resources with "high costs and low returns." Third, dispatching strategies are limited to local optimization. Traditional models often formulate dispatching rules for single production lines or single energy sources, failing to take a holistic view of the factory, considering factors such as production value, energy costs, and energy supply capacity. This results in optimization of some stages but poor overall efficiency (e.g., energy consumption of a certain production line is reduced, but high-value orders are delayed, ultimately affecting overall profits). This step, based on an evaluation information set and preset cost constraints (such as electricity price thresholds), uses an optimization engine to calculate the dynamic electricity value critical point. It then formulates a global energy dispatch strategy through decision trees or rule sets, generating a planning information set containing allocation ratios and timing. Through dynamic thresholds and global decision-making, energy dispatch becomes more intelligent and economical, effectively reducing energy costs while ensuring production value. It also optimizes energy storage usage, extends equipment lifespan, improves overall energy efficiency, and ultimately achieves sustainable synergy between production and energy.

[0031] S203. Obtain the set of energy supply influencing factors. Based on the energy dispatch strategy information set and the energy supply influencing factor set, perform multi-time-scale rolling optimization and global energy preset allocation to generate an energy allocation execution instruction set.

[0032] The energy supply influencing factor set can be a collection of data on various factors affecting the stability and availability of a factory's energy supply, including weather forecast information, historical energy data, and grid electricity price forecast data. This data originates from the factory's energy monitoring equipment (power monitoring instruments, energy storage status monitoring devices). Multi-timescale rolling optimization can be a phased, dynamically adjusted optimization method. Based on short-term, medium-term, and long-term timescales, and considering changes in production value, fluctuations in energy supply influencing factors, and changes in energy costs, it periodically adjusts and optimizes the energy dispatch strategy. Global energy pre-allocation can be the process of planning and allocating the energy supply, sources, and usage methods of each production line and energy storage device in the factory at different time periods, based on the rolled-optimized energy dispatch strategy. The energy allocation execution instruction set can be a set of specific executable instructions that transform the global energy pre-allocation results, ensuring that each execution unit (production line, energy storage device) can respond accurately.

[0033] Specifically, the factory's energy supply system is affected by multiple internal and external factors, resulting in significant uncertainties: externally, the power grid may experience peak load curtailment, voltage fluctuations, and sudden changes in electricity prices, while changes in weather conditions can lead to drastic fluctuations in the output of self-provided photovoltaic and wind power (such as a sharp drop in photovoltaic output on rainy days); internally, energy storage devices may experience capacity decay and reduced charging and discharging efficiency, production equipment may experience sudden failures leading to fluctuations in energy consumption, and energy transmission lines may also experience problems such as changes in loss rates. Traditional energy dispatching models struggle to address these uncertainties: First, they employ static allocation schemes, which remain unchanged for extended periods after strategy formulation. When energy supply influencing factors undergo sudden changes, the dispatching strategy cannot adapt in time, leading to an imbalance between energy supply and production demand (e.g., a sudden drop in photovoltaic output without timely grid power replenishment causing production line outages). Second, they lack a multi-timescale optimization perspective, focusing only on real-time energy consumption in the short term while neglecting long-term resource allocation coordination, resulting in a disconnect between short-term responses and long-term planning (e.g., purchasing large quantities of high-priced grid electricity in the short term to ensure production leads to long-term cost overruns). Third, pre-planned allocation is lacking; traditional models often rely on "real-time, on-demand allocation," failing to plan energy supply for different time periods in advance, resulting in delayed energy supply connections. For example, if energy storage devices are not charged in advance during off-peak hours, they cannot meet discharge demands during peak periods. This step acquires real-time energy supply influencing factors (such as weather and grid status), combines this with a strategy information set for multi-timescale rolling optimization, adjusts the allocation scheme, and generates a specific execution instruction set, including control commands and energy storage parameters, ensuring comprehensive allocation coverage. By optimizing across multiple time scales and allocating resources globally, the scheduling system's ability to cope with uncertainties is significantly improved, the risk of energy outages is reduced, and system stability and production efficiency are enhanced, ultimately achieving dynamic balance and efficient operation of energy scheduling.

[0034] S204. Based on the energy allocation execution instruction set, issue and execute scheduling instructions and dynamic production compensation for production lines and energy storage layers, and generate and output multi-energy collaborative scheduling reports.

[0035] The energy storage layer can be an energy storage system used by a factory to store electrical energy, including self-contained energy storage equipment (lithium battery energy storage system), distributed energy storage units, and energy storage management and control modules. Its core function is to balance energy supply and demand, and achieve peak shaving and valley filling, as well as emergency power supply. The multi-energy coordinated dispatch report can be a report formed by summarizing and analyzing the entire dispatch process, results, and key indicators, covering specific equipment control commands, deviation data, and compensation effects.

[0036] Specifically, in distributed energy storage scheduling, command execution and compensation mechanisms are crucial to ensuring effective scheduling. However, traditional methods often lack dynamic compensation and report generation, leading to uncorrected execution deviations. For example, energy allocation commands may fail due to equipment malfunctions or require compensation due to production changes, but the system cannot automatically adjust. This step executes commands issued by the system to production lines and energy storage devices, monitors operational status and triggers dynamic compensation (such as resource reallocation), integrates log data to generate a multi-energy collaborative scheduling report, including energy efficiency indicators and execution summaries, and outputs it to factory management. Through this step, factories can deeply integrate order production value with energy scheduling, achieving adaptive and efficient distributed energy storage management. The overall approach, through multi-step collaboration, solves the problem of production and energy disconnect in traditional scheduling, improves energy utilization, production efficiency, and system reliability, and provides a feasible path for energy transformation in the industrial sector.

[0037] The method provided in this embodiment first acquires a set of factory order information, extracts and integrates multi-dimensional production value characteristics to generate a dynamic production value assessment information set; then, combined with preset energy cost constraints, it calculates a dynamic electricity value threshold, decides on a global energy dispatch strategy, and forms an energy dispatch strategy information set; next, it acquires a set of energy supply influencing factors, performs multi-timescale rolling optimization and global preset allocation, and generates an energy allocation execution instruction set; finally, it issues instructions to the production line and energy storage layer to execute dynamic production compensation, records data, generates a multi-energy collaborative dispatch report, and outputs it to factory management personnel. This method achieves precise matching of production value and energy utilization, improving factory economic efficiency; dynamically controls energy costs and reduces waste; enhances the ability to cope with uncertainties and ensures stable energy supply; maximizes the collaborative efficiency of the production line and energy storage layer; provides data support for management decisions, promotes the transformation of energy management to data-driven approaches, and solves many pain points of traditional dispatching.

[0038] In some embodiments, the factory order information set includes the unit product marginal profit, total energy consumption required for production, delivery deadline, and current production progress for each order; the unit product marginal profit is obtained by dividing the unit product marginal profit by the total energy consumption required for production; the reciprocal of the ratio of the remaining production time to the standard production line hours is calculated as the production urgency based on the delivery deadline and current production progress; and the unit energy consumption profit and production urgency of all orders are integrated to generate a dynamic production value assessment information set.

[0039] Single-product marginal profit refers to the profit obtained by a single product during its production and sales process after deducting variable costs such as direct materials and direct labor. It excludes fixed costs such as fixed asset depreciation and administrative expenses, accurately reflecting the immediate profitability of a single product. Total energy consumption required for production refers to the total energy consumed in all stages of completing the entire production process for a given order, from raw material processing and semi-finished product assembly to finished product testing. This includes various energy sources such as electricity and wind power, and is uniformly converted into standard energy consumption units for measurement. Delivery deadline is the final delivery date of the product as clearly agreed upon by the factory and the customer in the order contract. Current production progress refers to the proportion of production work completed for a given order at a specific point in time to the total production work for that order. This can be characterized by production process completion rate, the percentage of products produced, etc. Unit energy consumption profit is a core evaluation indicator calculated by combining single-product marginal profit and total energy consumption required for production, used to characterize the marginal profit created per unit of energy consumption. Production urgency can be a time-based assessment indicator calculated based on the order delivery deadline and the current production progress. Specifically, it is the reciprocal of the ratio of remaining production time to standard production line hours, used to accurately quantify the urgency of orders that need to be prioritized for production.

[0040] Specifically, traditional production value assessment models have significant flaws. They often use total profit per unit as the sole core indicator, completely severing the connection between order profitability and production energy consumption. This leads to the priority of orders with high marginal profits but extremely high energy consumption (e.g., an order with a profit of 200 yuan per unit but requiring 100kWh of electricity), resulting in the excessive use of limited energy resources, driving up overall energy costs and violating green production requirements. Meanwhile, orders with low energy consumption and high unit efficiency are easily marginalized. The judgment of order urgency also relies on manual experience or only on the number of days remaining for delivery, without considering the current production progress and standard production line hours. For example, two orders with 5 days remaining for delivery, one 80% complete and the other only 30% complete, are considered equally urgent, resulting in a misallocation of production resources. To address the above issues, this step first involves collaboratively collecting data through the Factory Order Management System (OMS), Energy Management System (EMS), and Manufacturing Execution System (MES). The OMS retrieves the marginal profit per unit for a given order (e.g., 180 RMB / unit) and the delivery deadline (e.g., November 20, 2024). The EMS collects the energy consumption for the entire production process of that order (e.g., total energy consumption for processing, assembly, and testing, 60 kWh, uniformly converted to standard electrical units). Simultaneously, the MES obtains the current production progress (e.g., 50% completed as of November 12th, meaning 50 units have been produced). After collection, a data verification mechanism is used for verification. For example, the energy consumption data from the EMS is compared with the smart meters on the production equipment (showing a cumulative energy consumption of 62 kWh), and corrected to 61 kWh to ensure accuracy. Next, the unit energy consumption profit is calculated by dividing the marginal profit per unit of 180 RMB by the corrected total production energy consumption of 61 kWh, resulting in approximately 2.95 RMB / kWh (if there are 3 batches of production, each batch...). For energy consumption of 20kWh, 21kWh, and 20kWh, the unit energy consumption profit per batch is calculated to be 9 yuan / kWh, 8.57 yuan / kWh, and 9 yuan / kWh, respectively. Next, the production urgency is calculated. As of November 12th, there are 8 days of remaining production time. The standard production line time (12 days to complete 100 units) is 8 / 12 ≈ 0.67. Taking the reciprocal, the urgency is approximately 1.5. Finally, all order data is integrated, and dynamic weights are set according to the factory's current green production goals (e.g., unit energy consumption profit weight 0.6, urgency weight 0.4). A comprehensive score for a given order is calculated (e.g., 2.95 yuan / kWh × 0.6 + 1.5 × 0.4 ≈ 2.37 points). This generates a dynamic assessment information set of production value, including order number, unit energy consumption profit, urgency, and comprehensive score. If the energy consumption of this order subsequently increases to 65kWh, the system updates the unit energy consumption profit to approximately 2.77 yuan / kWh in real time and adjusts the comprehensive score accordingly.

[0041] The method provided in this embodiment can improve energy utilization efficiency by prioritizing energy allocation to high-energy-efficiency orders through unit energy consumption profit, thereby maximizing the return on energy investment; reduce the risk of production delays by ensuring timely processing of urgent orders through production urgency, thereby improving order delivery timeliness and customer satisfaction.

[0042] In some embodiments, a comprehensive evaluation index for prioritizing order production is constructed based on the unit energy consumption profit and production urgency of each order, serving as the weighted value contribution of each order; the weighted value contributions of all pending orders are ranked, and a unified dynamic electricity value threshold for the entire plant is calculated based on the ranking results. The dynamic electricity value threshold represents the upper limit of unit energy cost acceptable for maintaining production at the current moment; the dynamic electricity value threshold is compared with the real-time electricity price of the power grid, and a global energy dispatch strategy decision is executed based on the comparison results, generating an energy dispatch strategy information set.

[0043] Weighted value contribution is a quantifiable value obtained by weighting the unit energy consumption profit and production urgency of an order using comprehensive evaluation indicators. It is a core parameter characterizing the value proportion of a single order in global production and energy allocation. Real-time grid electricity price is the unit electricity price published in real time by the grid company based on the current electricity supply and demand situation. It includes peak-hour prices, off-peak-hour prices, and flat-hour prices, and in some regions, it also includes renewable energy surcharges.

[0044] Specifically, traditional energy dispatching strategies have significant flaws: fragmented consideration of order value, prioritizing orders solely based on production urgency leading to high-price energy consumption for low-energy-consumption, low-profit orders, or focusing only on profit per unit of energy consumption while neglecting urgent order delivery, lacking a unified evaluation standard; static and rigid energy cost thresholds, with long-term unchanging thresholds failing to adapt to real-time fluctuations in grid electricity prices (such as peak-valley price differences of up to 0.8 yuan / kWh) and changes in order structure, easily missing low-cost energy purchase opportunities or exceeding cost limits; dispatching decisions are limited to local areas, with each production line purchasing its own energy, resulting in both wasted self-supplied energy and high-priced electricity purchases; and order value and energy cost are viewed in isolation, lacking a scientific correlation mechanism. To address the above issues, this step first extracts unit energy consumption profit and production urgency from the dynamic production value assessment information set to construct a comprehensive evaluation index. Weights are assigned based on the factory's current goals (e.g., for the current month, with profitability as the core, unit energy consumption profit has a weight of 0.6, and production urgency is 0.4). The weighted value contribution is calculated by weighted summation (e.g., for an order with unit energy consumption profit of 6 yuan / kWh and urgency of 1.5, the weighted contribution = 6 × 0.6 + 1.5 × 0.4 = 4.2). Next, pending orders (including 30 queued orders and 5 urgent orders) are retrieved from the Order Management System (OMS) and sorted from highest to lowest weighted contribution (the top 10 are high-value orders). This is then considered in conjunction with the factory's monthly production capacity of 5000 units and total energy consumption requirements. Given a target of 800 MWh and an expected profit of 2 million yuan, a dynamic electricity value threshold of 1.4 yuan / kWh is calculated (to ensure high-value orders while controlling costs). Finally, the real-time electricity price is obtained through the Energy Management System (EMS) (e.g., peak price of 1.6 yuan / kWh at 10:00, which is higher than the threshold). The decision prioritizes the use of the self-owned photovoltaic power station (real-time output of 60kW) and energy storage equipment (discharge power of 80kW) for power supply. Low-value orders are shifted to off-peak hours (electricity price of 0.8 yuan / kWh). This decision is further refined into the energy source ratio of each production line (production line A: 70% photovoltaic power supply, 30% energy storage) and the energy storage charging and discharging plan (charging from 23:00 to 6:00), generating an energy dispatch strategy information set.

[0045] The method provided in this embodiment effectively solves the problems of the disconnect between production value and energy cost, and the lack of overall coordination in scheduling decisions; it achieves scientific order priority by weighted value contribution, ensuring energy supply for high-value orders; it adapts dynamic electricity value thresholds to electricity price fluctuations, reducing energy costs; and the global scheduling strategy enables multiple energy sources to work together, improving energy utilization efficiency and providing precise guidance for subsequent energy allocation and execution.

[0046] In some embodiments, when the dynamic electricity value threshold is greater than the real-time grid electricity price, a value-first mode is determined, and a production-guarantee dispatch strategy is generated: priority is given to dispatching real-time renewable energy for production; when real-time renewable energy is insufficient, energy storage power is dispatched to supplement it; when energy storage power is still insufficient, electricity is purchased from the grid to ensure production continuity. When the dynamic electricity value threshold is less than or equal to the real-time grid electricity price, a cost-first mode is determined, and an economic optimization dispatch strategy is generated: production activities are restricted to using only real-time renewable energy; when there is a surplus of real-time renewable energy, it is stored in the energy storage layer; when the grid electricity price is in a valley, energy storage power is dispatched or electricity is purchased from the grid for production; when the grid electricity price is in a peak period, the energy storage layer is controlled to discharge to the grid to realize revenue.

[0047] Value-first mode is a dispatching mode triggered when the dynamic electricity value threshold is greater than the real-time grid price. Its core objective is to ensure the continuity of order production and value realization, prioritizing energy needs for production and appropriately relaxing short-term energy cost controls. Production-guarantee dispatching strategies can be specific dispatching schemes executed under the value-first mode, with the core objective of "uninterrupted production and prioritizing high-value orders." Cost-first mode is a dispatching mode triggered when the dynamic electricity value threshold is less than or equal to the real-time grid price. Its core objective is to strictly control energy costs, maximize the use of low-cost energy, reasonably avoid high-priced grid electricity, and adjust the pace of production energy consumption when necessary. Economic optimization dispatching strategies can be specific dispatching schemes executed under the cost-first mode, with the goal of "minimizing energy costs and maximizing energy storage benefits." Grid price off-peak periods are the times when grid prices are at their lowest throughout the day, typically during periods of low electricity load (e.g., 00:00-06:00). During these periods, electricity purchase costs are low, making them suitable for energy storage charging or production energy consumption. Peak electricity prices can be the period when grid electricity prices are at a relatively high level throughout the day, usually during peak electricity load periods (such as 09:00-12:00 in the morning and 17:00-20:00 in the afternoon). Energy storage discharge can generate higher returns during this period.

[0048] Specifically, the traditional dispatching model has significant drawbacks. First, it is a rigid and singular model, allocating energy according to a fixed logic regardless of grid electricity prices. For example, long-term reliance on grid purchases leads to underutilization of low-priced electricity for production or energy storage when prices are below cost thresholds, and continued high-priced purchases when prices are above thresholds, pushing up costs. Second, there is insufficient coordination between renewable energy and energy storage. Renewable energy is used only when available and discarded when not, resulting in wasted surplus electricity. During shortages, the grid is relied upon directly, with energy storage only used for emergency backup and not contributing to revenue generation. Third, there is an imbalance between production and costs, either by purchasing high-priced electricity to maintain production or by excessively rationing electricity to control costs, leading to order delays. Fourth, peak-valley electricity pricing is not fully utilized, resulting in missed arbitrage opportunities. To address the above issues, this step of the energy dispatch strategy decision service first retrieves the dynamic electricity value threshold (e.g., 1.5 yuan / kWh) and the real-time grid price pushed by the grid dispatch center (e.g., 1.2 yuan / kWh). After unit and time synchronization verification, it determines to activate the value-first mode and synchronizes the instruction to the energy management system and production execution system. The real-time output of the factory's photovoltaic power station (e.g., 80kW) is obtained through the renewable energy monitoring system and fully allocated to the high-value order production line (requiring 75kW). If the output drops to 60kW due to reduced sunlight, the energy storage management system immediately controls the lithium battery energy storage equipment to discharge at 15kW to supplement it. If this is still insufficient (requiring 75kW, 75kW is just enough, but if 80kW is needed...), the output will be significantly lower. If the power consumption is 5kW, the grid's dedicated interface will be automatically triggered to purchase electricity (5kW) to ensure continuous production. If the threshold is 1.2 yuan / kWh and the electricity price is 1.5 yuan / kWh, the cost-priority mode will be activated, limiting production to use only real-time wind power output (e.g., 45kW). When the output is 10kW in surplus, the energy storage management system will control the energy storage device to charge at 10kW. In conjunction with the grid's peak and valley periods (valley period 00:00-06:00, electricity price 0.6 yuan / kWh; peak period 17:00-20:00, electricity price 1.8 yuan / kWh), the energy storage will be dispatched to discharge at 25kW or purchase electricity (20kW) to supply the production line during the valley period, and the energy storage will be controlled to discharge to the grid at 20kW to generate revenue during the peak period.

[0049] The method provided in this embodiment adapts to electricity price fluctuations based on differentiated scheduling mode, ensures production continuity and avoids order delays through value-first mode, strictly controls energy costs and maximizes the use of renewable energy through cost-first mode, and activates the multiple values ​​of energy storage layer, increases additional revenue through peak and off-peak operation, and improves the synergistic efficiency of multiple energy sources and the overall benefits of the plant.

[0050] In some embodiments, the energy supply influencing factor set includes weather forecast information, historical energy data, and grid electricity price forecast data within a preset future time window; based on the weather forecast information and historical energy data, photovoltaic power generation forecast sequences and wind power generation forecast sequences are generated respectively through physical equation calculation and power characteristic curve query; using the photovoltaic power generation forecast sequence, wind power generation forecast sequence, and grid electricity price forecast data as dynamic boundary conditions, the energy dispatch strategy information set is coupled to perform rolling optimization calculations at multiple time scales; based on the rolling optimization calculation results, an energy allocation execution instruction set is generated for controlling the charging and discharging of the energy storage layer, power allocation of production lines, and grid interaction in different future time periods.

[0051] The future preset time window can be a predicted time range pre-set by the factory based on production plans and energy dispatch needs. It can be flexibly set to short-term (e.g., 24 hours), medium-term (e.g., 7 days), and long-term (e.g., 30 days) to adapt to energy planning needs at different time scales. Weather forecast information can be meteorological forecast data for the factory's location within the future preset time window. Historical energy data can be a complete historical record of the factory's energy production, consumption, and storage over a past period. Grid electricity price forecast data can be the predicted electricity price information released by the grid company within the future preset time window, including predicted electricity prices for peak, flat, and valley periods, as well as price fluctuation trend indicators. Physical equation calculation can be a professional calculation method used for energy power prediction, by constructing mathematical and physical relationship equations between meteorological factors such as sunlight intensity and wind speed and power generation. Power characteristic curve query can be a reference tool to assist in power prediction, by retrieving the standard power characteristic curves (e.g., photovoltaic module IV curve, wind turbine power curve) provided at the factory for photovoltaic modules and wind turbine generators. A photovoltaic (PV) power generation forecast sequence can be a set of chronologically ordered PV power plant output forecasts, calculated using weather forecasts and historical energy data, dynamically reflecting the PV power generation potential at different future time periods. A wind power generation forecast sequence is similar, consisting of a chronologically ordered set of wind power plant output forecasts. Dynamic boundary conditions can be a set of constraint parameters used in multi-timescale rolling optimization calculations. Rolling optimization calculations can be phased, periodic dynamic optimization methods that continuously adjust and optimize energy allocation schemes based on short-term, medium-term, and long-term timescales, combined with dynamic boundary conditions and energy dispatch strategies.

[0052] Specifically, the traditional energy allocation command generation model has obvious flaws. It ignores the uncertainty of future energy supply and generates commands based solely on real-time data. For example, if the command is still issued based on the current day's situation without anticipating a sudden drop in photovoltaic output due to cloudy weather the next day, it can easily lead to energy shortages and production shutdowns. Renewable energy power forecasts rely solely on historical data and do not incorporate physical equations and power curves, resulting in large errors. For instance, if the predicted wind power is 100kW but the actual output is only 60kW, it can affect production plans. Furthermore, the commands are rigid and lack optimization across multiple time scales, making them unadjustable in the event of sudden changes in weather or electricity prices. They also suffer from vague content and poor executability. To address the above issues, this step first constructs a set of energy supply influencing factors. It then obtains weather forecast information for a future preset time window (e.g., 24 hours) via a meteorological platform interface (e.g., 800W / m² solar irradiance at 9:00 AM, 5m / s wind speed at 2:00 PM). It retrieves six months of historical energy data from the energy management system (e.g., 75kW photovoltaic output at 10:00 AM on a certain day during the same period last year) and obtains electricity price forecast data from the power grid center (e.g., peak electricity price of 1.8 yuan / kWh from 5:00 PM to 8:00 PM). Next, it generates a power prediction sequence, calculates it using the photovoltaic irradiance-power physical equation, and queries it in conjunction with the module IV curve. The photovoltaic power generation prediction sequence is obtained (e.g., 80kW at 9:00, 85kW at 10:00). Similarly, the wind power prediction sequence (e.g., 65kW at 14:00) is generated using the wind speed-power equation and the wind turbine power curve. Then, using these two sequences and electricity price data as dynamic boundary conditions, corresponding strategies are coupled to carry out multi-timescale rolling optimization (updated hourly from 1 to 24 hours in the short term, such as adjusting the energy storage plan at 12:00). Finally, an instruction set is generated, including explicit instructions such as charging energy storage at 120kW at 14:00, setting the energy consumption limit of production line A to 180kW at 15:00, and purchasing 40kW of electricity from the grid at 17:00.

[0053] The instruction set generated by the method provided in this embodiment has strong foresight and accuracy, effectively avoiding the risk of renewable energy fluctuations; multi-timescale rolling optimization allows energy allocation to adapt to the needs of different time periods, ensuring stable production; at the same time, it refines the operating parameters of each link, improves the efficiency of multi-energy synergy, and helps to optimize costs and increase profits.

[0054] In some embodiments, based on solar irradiance and ambient temperature from weather forecast information, an initial photovoltaic power generation sequence is calculated using the photoelectric conversion physical equation to obtain the initial photovoltaic power generation sequence. Simultaneously, the initial photovoltaic power generation sequence is corrected by coupling temperature loss and the actual attenuation coefficient of the equipment to generate a photovoltaic power generation prediction sequence. Based on wind speed from weather forecast information, the inherent power characteristic curve of the corresponding wind turbine model is queried to obtain a reference wind power generation sequence. At the same time, the wake loss of the reference wind power generation sequence is corrected by combining wind direction to generate a wind power generation prediction sequence.

[0055] The photoelectric conversion physical equation can be a specialized equation describing the quantitative relationship between input parameters such as solar irradiance and ambient temperature and the output power of photovoltaic modules. Its core reflects the physical process by which photovoltaic materials convert solar energy into electrical energy. The inherent power characteristic curve refers to the inherent correspondence between wind speed and output power for a specific wind turbine model under standard operating conditions, provided by the wind turbine manufacturer. Wake loss correction refers to the energy loss caused by the airflow wake generated by upstream wind turbines, which leads to reduced wind speed and increased turbulence in the area where downstream wind turbines are located, resulting in a decrease in the power generation of downstream wind turbines. This often occurs in wind turbine arrays.

[0056] Specifically, traditional methods for generating photovoltaic and wind power prediction sequences have significant flaws: photovoltaic predictions rely solely on solar irradiance and ambient temperature to calculate the initial sequence using basic photoelectric conversion equations, completely ignoring temperature losses (e.g., power loss can reach 8% when the ambient temperature exceeds the optimal operating temperature of the modules by 5°C in summer) and the actual degradation coefficient of the equipment (e.g., the degradation coefficient of photovoltaic modules after 6 years of use is approximately 0.88), resulting in predicted values ​​far exceeding actual output. Wind power predictions only use wind speed to query the inherent power characteristic curve of the wind turbine to obtain the baseline sequence, without considering wake losses caused by wind direction (e.g., upstream wind turbines in a wind turbine array reduce the power of downstream wind turbines by 15%-20%), leading to extremely large errors. These crude predictions deprive subsequent multi-timescale rolling optimization calculations of reliable data support, resulting in a severe disconnect between energy allocation execution instructions and actual energy supply. To address the above issues, this step first obtains weather forecast information for a future preset time window (e.g., 24 hours) from a professional meteorological service platform (e.g., solar irradiance 750W / m², ambient temperature 26℃ at 8:00, wind speed 7m / s, wind direction northwest at 15:00). It then retrieves the actual attenuation coefficient of the photovoltaic module equipment (e.g., 0.9, 4 years of use), wind turbine model (e.g., GW160-5.0), and corresponding inherent power characteristic curves from the factory equipment management system. When generating the photovoltaic sequence, the solar irradiance and ambient temperature at 8:00 are substituted into the photoelectric conversion physical equation (e.g., P=G×K×S, where G is irradiance, K is photoelectric conversion efficiency, and S is module area) to calculate the initial photovoltaic power generation sequence (e.g., initial power 85kW at 8:00). This is then corrected to 84.66kW using the temperature loss formula (e.g., power decreases by 0.4% for every 25℃ exceeding the optimal temperature), and further corrected to 76.19kW using the attenuation coefficient of 0.9. The photovoltaic prediction sequence is then generated by processing the data for each time period. When generating the wind power sequence, the GW160-5.0 power curve is queried at a wind speed of 7 m / s at 15:00 to obtain the baseline wind power generation sequence (e.g., the baseline power at 15:00 is 480 kW). Due to the northwest wind direction, the downstream wind turbines suffer from a wake loss of 18%, which is corrected to 393.6 kW. After completing the wake loss correction for each time period, the wind power prediction sequence is generated. Finally, the two sequences are compared with the historical actual output of the same period in recent times (with the error controlled within 6%) to complete the verification.

[0057] The method provided in this embodiment improves the accuracy of power prediction sequences through targeted corrections, effectively avoiding deviations between ideal calculations and actual operation; it adapts to the effects of photovoltaic module aging and temperature, corrects wind turbine wake losses, provides reliable data support for subsequent scheduling, ensures reasonable and efficient energy allocation, and enhances the synergistic efficiency of multiple energy sources.

[0058] In some embodiments, a three-layer collaborative optimization architecture is constructed, comprising day-ahead optimization, intraday rolling optimization, and real-time adjustment. At the day-ahead optimization layer, based on predictable future renewable energy generation and electricity price forecasts, an energy dispatch strategy information set is coupled to formulate a baseline energy allocation plan with the goal of minimizing total energy costs and maximizing production value. At the intraday rolling optimization layer, optimization is performed on a rolling basis with fixed time windows. Based on the latest meteorological and electricity price data, the charging and discharging strategies and production power allocation for near-term periods in the baseline energy allocation plan are dynamically corrected to generate an adjusted energy allocation plan. At the real-time adjustment layer, power deviations are detected based on minute-level monitoring data, and rapid compensation decisions are made for the adjusted energy allocation plan based on dynamic electricity value thresholds, enabling the strategy to evolve self-evolving.

[0059] The day-ahead optimization layer can be considered the long-term planning layer in the three-layer architecture, focusing on energy allocation planning for a full day (or longer period) in the future. Its core responsibility is to formulate a baseline energy allocation plan. This baseline plan can be the initial energy allocation scheme formulated by the day-ahead optimization layer, with the core objectives of minimizing total energy costs and maximizing production value. The intraday rolling layer can be considered the medium-term adjustment layer in the three-layer architecture, dynamically revising the plan over fixed time windows, bridging day-ahead planning and real-time execution. The adjusted energy allocation plan can be a dynamic scheme formed by the intraday rolling layer after revising the baseline energy allocation plan, focusing on optimizing dispatch strategies for the immediate period, adapting to the latest weather and electricity price changes, and is updated and generated by the dynamic dispatch service. The real-time adjustment layer can be considered the short-term response layer in the three-layer architecture, quickly processing power deviations based on minute-level high-frequency monitoring data to ensure the real-time adaptability of dispatch strategies. The rapid compensation decision can be the immediate dispatch decision made by the real-time adjustment layer in response to power deviations, based on dynamic electricity value thresholds, quickly adjusting the charging and discharging power of energy storage or the energy allocation of production lines, and is automatically generated by the real-time response service.

[0060] Specifically, traditional multi-timescale optimization calculations have significant drawbacks. First, their architecture is simplistic, often consisting of day-ahead static planning or a simple two-layer structure, lacking three-layer coordination and unable to cope with sudden intraday weather changes (such as a sudden midday downpour causing a 30% drop in photovoltaic output) and electricity price fluctuations. Second, their optimization objectives are one-sided, focusing only on minimizing costs or ensuring production, failing to balance the dual needs of minimizing total energy costs and maximizing production value. Third, their response cycle is on the hourly level, unable to handle minute-level power deviations (such as a 15% fluctuation in wind turbine output due to instantaneous wind speed changes). Fourth, their strategies are rigid and lack self-evolution capabilities, unable to adapt to long-term production and equipment changes. To address these issues, this step first establishes a three-layer collaborative optimization architecture including day-ahead optimization, intraday rolling, and real-time adjustment. The day-ahead layer is responsible for formulating the baseline plan, the intraday layer makes rolling corrections with a fixed 1-hour time window, and the real-time layer adjusts based on 5-minute data. Each layer achieves data interoperability through an energy management system (EMS). The daytime layer retrieves the next day's photovoltaic power generation forecast sequence (e.g., 85kW at 9:00, 70kW at 14:00), wind power generation forecast sequence (e.g., 60kW at 10:00, 75kW at 16:00), and grid electricity price forecast data (peak period 17:00-20:00 1.8 yuan / kWh) generated by weight 5. Coupled with the cost-priority strategy of weight 4, a baseline energy allocation plan is formulated with the goal of minimizing total energy cost and maximizing production value (e.g., energy storage charging at 120kW from 9:00-11:00). The intraday layer retrieves the latest meteorological data every hour (e.g., updating the 11:00-12:00 photovoltaic output forecast to 65kW at 10:30), revising the baseline plan and reducing the energy storage charging power to 100kW from 11:00-12:00, generating an adjustment plan. Data is collected every 5 minutes in real time. If the actual wind power output at 11:05 is 72kW (12kW over the plan), the power deviation is calculated and the energy storage is instructed to charge at an incremental rate of 12kW based on the dynamic electricity value threshold (1.5 yuan / kWh). At the same time, the cause of the deviation and the compensation effect are recorded to optimize the subsequent intraday rolling window (such as adjusting the window to 45 minutes) and realize the self-evolution of the strategy.

[0061] The approach provided in this embodiment significantly enhances the system's adaptability to fluctuations in renewable energy and changes in market prices, effectively reduces the risk of increased operating costs and production interruptions caused by forecasting errors, continuously optimizes decision-making quality through ongoing strategy self-evolution, and ultimately achieves a synergistic improvement in energy utilization efficiency and economic benefits.

[0062] In some embodiments, the energy allocation execution instruction set is parsed into specific equipment control instructions and sent to the production execution layer and the energy management layer respectively; the total production power balance is monitored in real time, and if an energy supply and demand imbalance is detected, a preset compensation strategy is initiated based on a dynamic electricity value threshold, the deviation data and compensation effect are recorded, and the feedback is sent to the global energy dispatch strategy decision-making steps; the specific equipment control instructions, deviation data and compensation effect are integrated to generate and output a multi-energy collaborative dispatch report.

[0063] Equipment control commands can be specific operational commands that can be directly executed by production and energy equipment, formed by breaking down the energy allocation execution command set. Total production power balance refers to the dynamic balance between total energy supply and total consumption in real time during the factory's production process. Total supply includes renewable energy output, energy storage discharge, and grid-purchased electricity; total consumption is the sum of energy consumption of each production line and auxiliary equipment. Energy supply-demand imbalance refers to a state where the total production power balance is disrupted, manifested as total energy supply exceeding total consumption (oversupply) or total supply falling short of total consumption (undersupply), exceeding the normal fluctuation range. Deviation data can be the difference between actual energy supply and consumption and planned values ​​when energy supply-demand imbalance occurs, as well as parameter changes before and after the implementation of compensation strategies. Compensation effect can be the improvement in energy supply-demand imbalance and the overall impact after the implementation of a preset compensation strategy.

[0064] Specifically, the generation of dispatch reports usually has obvious defects. The report content is fragmented, only recording energy consumption or production data, without integrating equipment control commands, deviation data and compensation effects, making it difficult to trace the entire dispatch chain. When energy supply and demand are unbalanced, the response is lagging, relying on manual adjustments and lacking closed-loop feedback, and similar problems recur. The report focuses on archiving results, lacks analysis of the causes of imbalance and the effectiveness of compensation, has low guidance value, and separates production and energy data, lacking a collaborative perspective. To address the above issues, this step initiates an instruction parsing service, breaking down the energy allocation execution instruction set into equipment control instructions. For example, an instruction is issued to the production execution layer to limit the energy consumption of production line A to 180kW from 10:00 to 12:00 and prioritize the use of photovoltaic power. An instruction is issued to the energy management layer to discharge energy at 120kW from 14:00 to 16:00. These instructions are then transmitted to the MES and EMS via the Industrial Control System (ICS) communication bus. Smart monitoring instruments collect data in real time. If an energy supply-demand imbalance (power deviation -30kW) is detected at 11:00, a preset compensation strategy is triggered. Based on a dynamic electricity value threshold of 1.5 yuan / kWh, 25kW of energy storage is discharged and 5kW of grid power is purchased to fill the gap. The deviation data (imbalance duration 15 minutes) and the compensation effect (balance restored at 11:15) are recorded. Finally, the instruction details, deviation data, and compensation effect are integrated to generate a report, which is then pushed to the management layer and archived. For example, the report may show a photovoltaic absorption rate of 85% and an imbalance compensation success rate of 100%.

[0065] The method provided in this embodiment constructs a closed loop of scheduling command execution and strategy optimization, and real-time compensation ensures stable energy supply and demand and continuous production; the report integrates data from the entire process, providing management with clear decision-making basis, helping to continuously optimize scheduling strategies and improve the synergistic efficiency of multiple energy sources.

[0066] In some embodiments, when the deviation between the actual output of renewable energy and the predicted sequence exceeds a preset disturbance threshold, real-time rescheduling is immediately triggered: based on a dynamic electricity value threshold, all currently executing production tasks are divided into guaranteed tasks and optimized tasks; for guaranteed tasks, their energy supply is maintained, and the resulting power deficit is compensated by energy storage discharge or grid purchase; for optimized tasks, according to their value contribution ranking, the power of tasks ranked lower is dynamically reduced or they are suspended to adapt to the currently available energy, and the surplus energy is redistributed.

[0067] The preset disturbance threshold can be a pre-defined critical value used to determine whether fluctuations in renewable energy output constitute significant interference. When the deviation between actual output and predicted value exceeds this threshold, the system will activate an emergency compensation mechanism. Guarantee-level tasks can be production tasks that must prioritize energy supply during real-time rescheduling. Optimization-level tasks can be production tasks that can appropriately adjust energy allocation to adapt to overall resource availability during real-time rescheduling.

[0068] Specifically, traditional solutions for addressing renewable energy output deviations have significant flaws: they do not prioritize production tasks, and when deviations exceed limits, they either blindly reduce energy consumption for all tasks, leading to production interruptions for high-value orders; or they use energy storage or the grid to make up the shortfall, causing a surge in energy costs; and they do not dynamically allocate surplus energy, resulting in an inefficient approach. To address the above issues, this step first sets a preset disturbance threshold (e.g., 5%). When the actual output of renewable energy (e.g., 47kW of photovoltaic power) deviates from the predicted sequence (50kW) by 6% and exceeds the threshold, real-time rescheduling is immediately triggered. Based on the dynamic electricity value threshold (e.g., 1.6 yuan / kWh), production tasks are divided into a guarantee level (e.g., high-value order production lines, with a unit energy consumption profit of 2 yuan / kWh) and an optimization level (e.g., low-value spare parts production, with a unit energy consumption profit of 1.2 yuan / kWh). For the guarantee level, energy supply is maintained by discharging 3kW of energy storage to fill the gap, and if insufficient, electricity is purchased from the grid. For the optimization level, the power of the last three tasks is reduced (from 80kW to 60kW) according to value, and the remaining 20kW is redistributed to the guarantee level to ensure that high-value tasks are not affected.

[0069] The system provided in this embodiment can respond precisely and quickly to sudden energy fluctuations, ensuring that high-value production activities are not disrupted and effectively avoiding economic losses caused by "one-size-fits-all" production restrictions. At the same time, this mechanism enhances the resilience and adaptability of the entire energy system, enabling continuous optimization of global energy allocation efficiency while ensuring production continuity.

[0070] Figure 3 This is a schematic diagram of the structure of a distributed energy storage scheduling system based on multi-energy coordination provided in an embodiment of this application, as shown below. Figure 3 As shown, the distributed energy storage scheduling system 300 based on multi-energy collaboration in this embodiment includes: a production value assessment module 301, an energy scheduling strategy module 302, an energy allocation execution module 303, and an energy collaborative scheduling module 304. The production value assessment module 301 is used to acquire a set of factory order information, and based on the set of factory order information, to perform multi-dimensional production value feature extraction and fusion analysis to generate a dynamic production value assessment information set. The energy dispatch strategy module 302 is used to perform dynamic electricity value threshold calculation and global energy dispatch strategy decision-making based on the production value dynamic evaluation information set and in combination with preset energy cost constraints, and generate an energy dispatch strategy information set. The energy allocation execution module 303 is used to obtain the energy supply influencing factor set, perform multi-time-scale rolling optimization and global energy preset allocation based on the energy scheduling strategy information set and the energy supply influencing factor set, and generate an energy allocation execution instruction set. The energy collaborative scheduling module 304 is used to issue and execute scheduling instructions and dynamic production compensation for production lines and energy storage layers based on the energy allocation execution instruction set, and generate and output a multi-energy collaborative scheduling report.

[0071] Optionally, the production value assessment module 301 is specifically used for: the factory order information set including the single-item marginal profit, total energy consumption required for production, delivery deadline, and current production progress of each order; dividing the single-item marginal profit by the total energy consumption required for production to obtain the unit energy consumption profit of each order; calculating the reciprocal of the ratio of the remaining production time to the standard production line hours based on the delivery deadline and current production progress as the production urgency; and integrating the unit energy consumption profit and the production urgency of all orders to generate the dynamic production value assessment information set.

[0072] Optionally, when generating the energy dispatch strategy module 302, it is specifically used to: construct a comprehensive evaluation index for order production priority based on the unit energy consumption profit and production urgency of each order, as the weighted value contribution of each order; sort the weighted value contributions of all pending orders, and calculate and generate a unified dynamic power value threshold for the entire plant based on the sorting result, wherein the dynamic power value threshold represents the upper limit of unit energy cost acceptable for maintaining production at the current moment; compare the dynamic power value threshold with the real-time electricity price of the power grid, execute a global energy dispatch strategy decision based on the comparison result, and generate the energy dispatch strategy information set.

[0073] Optionally, when the energy dispatch strategy module 302 executes the global energy dispatch strategy decision based on the comparison result, it is specifically used for: when the dynamic power value threshold is greater than the real-time grid price, determining it to be a value-first mode and generating a production-guarantee dispatch strategy: prioritizing the dispatch of real-time renewable energy power for production; when real-time renewable energy power is insufficient, dispatching energy storage power to supplement it; when the energy storage power is still insufficient, purchasing power from the grid to ensure production continuity; when the dynamic power value threshold is less than or equal to the real-time grid price, determining it to be a cost-first mode and generating an economic optimization dispatch strategy: restricting production activities to use only real-time renewable energy power; when there is a surplus of real-time renewable energy power, storing it in the energy storage layer; when the grid price is in a valley, dispatching the energy storage layer power or purchasing power from the grid for production; when the grid price is in a peak period, controlling the energy storage layer to discharge to the grid to realize revenue.

[0074] Optionally, when generating the energy allocation execution instruction set, the energy allocation execution module 303 is specifically used for: the energy supply influencing factor set including weather forecast information, historical energy data, and grid electricity price prediction data within a future preset time window; based on the weather forecast information and the historical energy data, generating photovoltaic power generation prediction sequences and wind power generation prediction sequences respectively through physical equation calculation and power characteristic curve query; using the photovoltaic power generation prediction sequence, the wind power generation prediction sequence, and the grid electricity price prediction data as dynamic boundary conditions, coupling the energy dispatch strategy information set, and performing multi-time-scale rolling optimization calculations; and based on the rolling optimization calculation results, generating the energy allocation execution instruction set for controlling the charging and discharging of the energy storage layer, production line power allocation, and grid interaction in different future time periods.

[0075] Optionally, when the energy allocation execution module 303 generates the photovoltaic power generation prediction sequence and the wind power generation prediction sequence respectively, it specifically performs the following steps: based on the solar irradiance and ambient temperature in the weather forecast information, it calculates the initial photovoltaic power generation sequence using the photoelectric conversion physical equation to obtain the initial photovoltaic power generation sequence; simultaneously, it corrects the initial photovoltaic power generation sequence by coupling temperature loss and the actual attenuation coefficient of the equipment to generate the photovoltaic power generation prediction sequence; based on the wind speed in the weather forecast information, it queries the inherent power characteristic curve of the corresponding wind turbine model to obtain the reference wind power generation sequence; simultaneously, it corrects the wake loss of the reference wind power generation sequence by combining the wind direction to generate the wind power generation prediction sequence.

[0076] Optionally, when performing multi-timescale rolling optimization calculations, the energy allocation execution module 303 is specifically used to: construct a three-layer collaborative optimization architecture including day-ahead optimization, intraday rolling, and real-time adjustment; at the day-ahead optimization layer, based on the foreseeable future renewable energy generation and electricity price forecast data, coupled with the energy dispatch strategy information set, to formulate a baseline energy allocation plan with the goal of minimizing total energy costs and maximizing production value; at the intraday rolling layer, to perform rolling optimization with a fixed time window period, and based on the latest meteorological and electricity price data, to dynamically correct the charging and discharging strategies and production power allocation for the near-term periods in the baseline energy allocation plan, generating an adjusted energy allocation plan; at the real-time adjustment layer, to detect power deviations based on minute-level monitoring data, and to make rapid compensation decisions for the adjusted energy allocation plan based on the dynamic electricity value threshold, thereby realizing the self-evolution of the strategy.

[0077] Optionally, when generating and outputting the multi-energy collaborative scheduling report, the energy collaborative scheduling module 304 is specifically used to: parse the energy allocation execution instruction set into specific equipment control instructions and send them to the production execution layer and the energy management layer respectively; monitor the total production power balance in real time, and if an energy supply and demand imbalance is detected, initiate a preset compensation strategy based on the dynamic electricity value threshold, record the deviation data and compensation effect, and feed it back to the global energy scheduling strategy decision-making step; integrate the specific equipment control instructions and the deviation data and compensation effect to generate and output the multi-energy collaborative scheduling report.

[0078] Optionally, when the energy coordinated scheduling module 304 initiates a preset compensation strategy based on the dynamic electricity value threshold if an energy supply-demand imbalance is detected, it is specifically used to: immediately trigger real-time rescheduling when the deviation between the actual output of renewable energy and the predicted sequence exceeds a preset disturbance threshold; divide all currently executing production tasks into guaranteed tasks and optimized tasks based on the dynamic electricity value threshold; maintain the energy supply for guaranteed tasks, and compensate for the resulting power deficit by energy storage discharge or grid purchase; dynamically reduce the power of tasks ranked lower according to their value contribution or postpone them to adapt to the currently available energy, and redistribute the surplus energy.

[0079] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A distributed energy storage scheduling method based on multi-energy coordination, characterized in that, include: Obtain a set of factory order information, and based on the set of factory order information, perform multi-dimensional production value feature extraction and fusion analysis to generate a dynamic production value assessment information set; Based on the aforementioned dynamic production value assessment information set, combined with preset energy cost constraints, dynamic electricity value threshold calculation and global energy dispatch strategy decision-making are performed to generate an energy dispatch strategy information set. Obtain the set of energy supply influencing factors, and based on the energy dispatch strategy information set and the set of energy supply influencing factors, perform multi-time-scale rolling optimization and global energy preset allocation to generate an energy allocation execution instruction set; Based on the energy allocation execution instruction set, scheduling instructions and dynamic production compensation for production lines and energy storage layers are issued and executed, and multi-energy collaborative scheduling reports are generated and output.

2. The method according to claim 1, characterized in that, The generated dynamic assessment information set for production value includes: The factory order information set includes the unit product marginal profit, total energy consumption required for production, delivery deadline, and current production progress for each order; Divide the marginal profit of the single product by the total energy consumption required for production to obtain the unit energy consumption profit for each order; Based on the aforementioned delivery deadline and the current production progress, the reciprocal of the ratio of the remaining production time to the standard production line hours is calculated as the production urgency. By integrating the unit energy consumption profit of all orders with the production urgency, a dynamic evaluation information set of production value is generated.

3. The method according to claim 2, characterized in that, The generated energy dispatch strategy information set includes: Based on the unit energy consumption profit and production urgency of each order, a comprehensive evaluation index for order production priority is constructed as the weighted value contribution of each order. The weighted value contribution of all pending orders is sorted, and a unified dynamic electricity value threshold for the entire plant is calculated based on the sorting results. The dynamic electricity value threshold represents the upper limit of unit energy cost that can be accepted to maintain production at the current moment. The dynamic electricity value threshold is compared with the real-time electricity price of the power grid. Based on the comparison result, a global energy dispatch strategy decision is made, and the energy dispatch strategy information set is generated.

4. The method according to claim 3, characterized in that, The step of making a global energy scheduling strategy decision based on the comparison results includes: When the dynamic electricity value threshold is greater than the real-time electricity price of the power grid, the system is determined to be in value-first mode, and a production-guarantee dispatching strategy is generated. Priority is given to scheduling real-time renewable energy power for production. When real-time renewable energy power is insufficient, energy storage power is scheduled to supplement it. When the energy storage power is still insufficient, power is purchased from the grid to ensure production continuity. When the dynamic electricity value threshold is less than or equal to the real-time electricity price of the power grid, the system is determined to be in cost-first mode, and an economically optimized scheduling strategy is generated. Production activities are restricted to using only real-time renewable energy power, and when there is a surplus of real-time renewable energy power, it is stored in the energy storage layer. When the grid electricity price is in a low period, the energy storage layer can be dispatched to generate electricity or purchased from the grid for production. When the grid electricity price is in a high period, the energy storage layer can be controlled to discharge to the grid to generate revenue.

5. The method according to claim 4, characterized in that, The set of instructions for generating energy allocation includes: The set of energy supply influencing factors includes weather forecast information, historical energy data, and grid electricity price forecast data within a future preset time window. Based on the weather forecast information and the historical energy data, photovoltaic power generation prediction sequences and wind power generation prediction sequences are generated respectively through physical equation calculation and power characteristic curve query. Using the photovoltaic power generation prediction sequence, the wind power generation prediction sequence, and the grid electricity price prediction data as dynamic boundary conditions, coupled with the energy dispatch strategy information set, rolling optimization calculations are performed at multiple time scales. Based on the rolling optimization calculation results, the energy allocation execution instruction set is generated for controlling the charging and discharging of the energy storage layer, the power allocation of the production line, and the grid interaction in different future time periods.

6. The method according to claim 5, characterized in that, The generation of photovoltaic power generation prediction sequences and wind power generation prediction sequences respectively includes: Based on the solar irradiance and ambient temperature in the weather forecast information, the initial photovoltaic power generation sequence is calculated using the photoelectric conversion physical equation to obtain the initial photovoltaic power generation sequence; Simultaneously, the initial photovoltaic power generation sequence is corrected by coupling temperature loss and the actual attenuation coefficient of the equipment to generate the photovoltaic power generation prediction sequence; Based on the wind speed in the weather forecast information, the inherent power characteristic curve of the corresponding wind turbine model is queried to obtain the benchmark wind power generation power sequence; Simultaneously, the wake loss of the benchmark wind power generation sequence is corrected by combining wind direction to generate the wind power generation prediction sequence.

7. The method according to claim 5, characterized in that, The multi-timescale rolling optimization calculation includes: Construct a three-layer collaborative optimization architecture that includes day-ahead optimization, intraday rolling optimization, and real-time adjustment; In the current optimization layer, based on the foreseeable future renewable energy generation and electricity price forecast data, the energy dispatch strategy information set is coupled to formulate a benchmark energy allocation plan with the goal of minimizing total energy costs and maximizing production value. In the intraday rolling layer, optimization is performed on a rolling basis with a fixed time window. Based on the latest meteorological and electricity price data, the charging and discharging strategies and production power allocation in the near-term of the baseline energy allocation plan are dynamically corrected to generate an adjusted energy allocation plan. In the real-time adjustment layer, power deviation is detected based on minute-level monitoring data, and rapid compensation decisions are made for the adjusted energy allocation plan based on the dynamic power value threshold, thereby realizing the self-evolution of the strategy.

8. The method according to claim 7, characterized in that, The generation and output of the multi-energy coordinated scheduling report includes: The energy allocation execution instruction set is parsed into specific equipment control instructions and then sent to the production execution layer and the energy management layer, respectively. Real-time monitoring of total production power balance; if an energy supply and demand imbalance is detected, a preset compensation strategy is initiated based on the dynamic electricity value threshold, deviation data and compensation effect are recorded, and feedback is provided to the global energy dispatch strategy decision-making step. The specific equipment control commands and the deviation data and compensation effects are integrated to generate and output the multi-energy collaborative scheduling report.

9. The method according to claim 8, characterized in that, If an energy supply-demand imbalance is detected, a preset compensation strategy is initiated based on the dynamic electricity value threshold, including: When the deviation between the actual output of renewable energy and the predicted sequence exceeds a preset disturbance threshold, real-time rescheduling is immediately triggered: based on the dynamic power value threshold, all currently executing production tasks are divided into guaranteed tasks and optimized tasks. For the aforementioned safeguard missions, their energy supply is maintained, and the resulting power deficit is compensated by energy storage discharge or grid power purchase; For the optimized tasks, the power of tasks ranked lower in the ranking is dynamically reduced or they are temporarily suspended to adapt to the available energy, and the surplus energy is redistributed.

10. A distributed energy storage dispatch system based on multi-energy coordination, characterized in that, The method applied to any one of claims 1-9 includes: The production value assessment module is used to acquire a set of factory order information, and based on the set of factory order information, to perform multi-dimensional production value feature extraction and fusion analysis to generate a dynamic production value assessment information set. The energy dispatch strategy module is used to perform dynamic electricity value threshold calculation and global energy dispatch strategy decision-making based on the production value dynamic assessment information set and in combination with preset energy cost constraints, and generate an energy dispatch strategy information set. The energy allocation execution module is used to obtain the energy supply influencing factor set, perform multi-time-scale rolling optimization and global energy preset allocation based on the energy scheduling strategy information set and the energy supply influencing factor set, and generate an energy allocation execution instruction set. The energy collaborative scheduling module is used to issue and execute scheduling instructions and dynamic production compensation for production lines and energy storage layers based on the energy allocation execution instruction set, and generate and output multi-energy collaborative scheduling reports.

Images