A method and system for evaluating the comprehensive benefits of energy storage

Abstract

This invention provides a method and system for comprehensive energy storage benefit assessment, relating to the field of comprehensive energy storage benefit assessment technology. It involves real-time collection of technical operating parameters, environmental parameters, and market dynamic data of the energy storage system; automatic matching of preset assessment modes using a scene recognition engine; calculation of a comprehensive score for technical, economic, and social benefits using a dynamic weighted algorithm to generate the calculation result; and generation of an assessment report including a carbon footprint traceability code based on the calculation result. Through innovative technologies such as dynamic weight allocation, precise battery degradation compensation, and blockchain data storage, the accuracy, adaptability, and credibility of energy storage system benefit assessment are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage comprehensive benefit evaluation, in particular to an energy storage comprehensive benefit evaluation method and system. BACKGROUND

[0002] With the large-scale grid connection of renewable energy and the deepening of power market reform, the role of energy storage systems in the power system has developed from a single energy storage unit to a multi-functional carrier supporting stable operation of the power grid, improving economic benefits and promoting low-carbon transformation. However, the existing energy storage benefit evaluation method has the following technical bottlenecks:

[0003] Insufficient data fusion: The traditional evaluation system often separates the correlation between technical operation parameters (such as battery health, frequency modulation response time), environmental factors (temperature, load characteristics) and market dynamics (price fluctuations, carbon trading prices), resulting in a disconnect between the evaluation results and actual application scenarios. For example, battery degradation models usually only consider cycle count, ignoring the coupling effects of temperature and charge / discharge rate, resulting in a capacity prediction deviation of more than 15%.

[0004] Poor scene adaptability: Current methods mostly use static weight allocation, which cannot dynamically respond to changes in demand in different scenarios such as peak shaving and new energy consumption. In grid frequency emergency events, existing systems may lag behind actual demand by more than 10 minutes due to the lack of real-time weight adjustment mechanisms (such as linearly increasing technology priority weight).

[0005] Missing full life cycle traceability: Carbon footprint accounting usually relies on annual average data, lacking dynamic traceability based on real-time operation data. Existing evaluation reports lack the hash fingerprints embedded in blockchain evidence, which poses a risk of data tampering and makes it difficult to meet the requirements of the EU Battery Regulation (EU Battery Regulation) for supply chain transparency.

[0006] Difficulty in multi-objective coordination: There is a lack of quantitative tools for the coordinated optimization of technical, economic and social benefits. For example, when the price peak-valley difference exceeds 0.4 yuan / kWh, the sudden increase in economic benefits may mask the risk of technical loss, and the traditional entropy weight method cannot automatically trigger the rapid switching of the economic dominant mode (economic benefit weight ≥ 0.7).

[0007] To address the above problems, it is necessary to develop a comprehensive evaluation method that integrates deep learning, dynamic weighting algorithm and blockchain technology to achieve precise benefit quantification and carbon footprint verifiable traceability of energy storage systems in all scenarios and multiple time scales. SUMMARY

[0008] To improve the accuracy of energy storage system benefit evaluation, the present application provides an energy storage comprehensive benefit evaluation method and system.

[0009] In a first aspect, the application provides a method for evaluating comprehensive benefits of energy storage, which adopts the following technical solution:

[0010] A method for evaluating comprehensive benefits of energy storage, comprising:

[0011] collecting technical operation parameters, environmental parameters and market dynamic data of the energy storage system in real time;

[0012] a scene recognition engine based on a deep learning model automatically matches three preset evaluation modes, including peak clipping and valley filling mode, new energy grid-connected mode and emergency backup mode, in combination with the collected data;

[0013] a dynamic weighting algorithm is used to calculate a three-dimensional benefit comprehensive score of technology-economy-society to generate a calculation result;

[0014] an evaluation report containing a carbon footprint traceability code is generated according to the calculation result.

[0015] Optionally, the technical operation parameters include battery health, equivalent cycle number and grid frequency modulation response delay time, and the step of collecting technical operation parameters of the energy storage system comprises:

[0016] an ampere-hour integral method is used to calculate a capacity attenuation rate in real time to obtain the battery health;

[0017] an equivalent cycle number is determined based on a temperature compensation discharge depth accumulation model;

[0018] a grid frequency modulation response delay time is determined through microsecond-level timestamp synchronization technology.

[0019] Optionally, the step of calculating a three-dimensional benefit comprehensive score of technology-economy-society to generate a calculation result by using a dynamic weighting algorithm comprises:

[0020] a weight decision matrix is established, and the sum of the technology benefit weight, the economic benefit weight and the social benefit weight in the weight decision matrix is 1;

[0021] in the weight decision matrix, a reference value corresponding to each type of weight is calculated according to an entropy weight method;

[0022] a future 24-hour scene demand change rate is predicted based on a long short-term memory network, and when the demand change rate reaches a preset threshold, a real-time adjustment mechanism is triggered;

[0023] the calculation result is determined according to the adjustment result.

[0024] Optionally, the real-time adjustment mechanism comprises:

[0025] an economic dominant mode, when the price peak-valley difference > 0.4 yuan / kWh, the economic benefit weight ≥ 0.7 is locked;

[0026] In the technology-priority mode, a linear incrementing mechanism is activated when the grid frequency deviation exceeds the limit for more than 10 minutes.

[0027] In the social compensation model, the carbon intensity correction coefficient is activated when the volatility of carbon emission rights prices exceeds 20%.

[0028] Optionally, it further includes: a battery degradation compensation step, the battery degradation compensation step including:

[0029] Multi-factor decay modeling: a predictive model that correlates temperature, cycle number, and charge / discharge rate;

[0030] Dynamic compensation mechanism: Introducing a compensation factor inversely proportional to the capacity decay rate into the economic benefit calculation;

[0031] Health warning function: Triggers a level 3 alarm signal when the capacity decay rate is >5%.

[0032] Optionally, the market dynamics data includes:

[0033] A real-time electricity price curve consisting of a base electricity price superimposed with multi-cycle fluctuation components;

[0034] Volatility of carbon emission rights prices predicted by an autoregressive conditional heteroscedasticity model.

[0035] Optionally, the processing of the market dynamics data includes:

[0036] Decompose the base electricity price, intraday fluctuations, and random disturbances into three levels of components;

[0037] The GARCH-M model is used to predict the fluctuation trend in the next 8 hours.

[0038] The hash fingerprint of all market data is stored on the blockchain.

[0039] Secondly, this application provides an energy storage comprehensive benefit assessment system, including:

[0040] The data acquisition module is used to collect technical operating parameters, environmental parameters, and market dynamic data of the energy storage system in real time.

[0041] The recognition module is used by the scene recognition engine based on the deep learning model to automatically match three preset evaluation modes with the collected data, including: peak shaving and valley filling mode, new energy grid connection mode and emergency backup mode.

[0042] The calculation module is used to calculate the comprehensive score of the three-dimensional benefits of technology, economy, and society using a dynamic weighted algorithm to generate calculation results;

[0043] The output module is used to generate an assessment report containing a carbon footprint traceability code based on the calculation results.

[0044] Thirdly, this application provides a computer device, the device comprising: a memory and a processor, wherein the processor, when executing computer instructions stored in the memory, performs the method described above.

[0045] Fourthly, this application provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the method described above.

[0046] In summary, this application collects real-time technical operating parameters, environmental parameters, and market dynamic data of the energy storage system; automatically matches a preset evaluation mode using a scene recognition engine; employs a dynamic weighted algorithm to calculate a comprehensive score of technical, economic, and social benefits to generate the calculation result; and generates an evaluation report containing a carbon footprint traceability code based on the calculation result. Through innovative technologies such as dynamic weight allocation, precise battery degradation compensation, and blockchain data storage, the accuracy, adaptability, and credibility of the energy storage system benefit evaluation are significantly improved. Attached Figure Description

[0047] Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application;

[0048] Figure 2 This is a flowchart illustrating the first embodiment of the energy storage comprehensive benefit assessment method of this application;

[0049] Figure 3 This is a structural block diagram of the first embodiment of the energy storage comprehensive benefit assessment system of this application. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0051] Reference Figure 1 , Figure 1 This is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiments of this application.

[0052] like Figure 1As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-Fidelity (Wi-Fi) interface). The memory 1005 may be high-speed random access memory (RAM) or stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0053] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the computer device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0054] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a network communication module, a user interface module, and an energy storage comprehensive benefit evaluation program.

[0055] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in this application can be set in the computer device, and the computer device calls the energy storage comprehensive benefit assessment program stored in the memory 1005 through the processor 1001 and executes the energy storage comprehensive benefit assessment method provided in the embodiment of this application.

[0056] This application provides a method for evaluating the comprehensive benefits of energy storage, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of the energy storage comprehensive benefit assessment method of this application.

[0057] In this embodiment, the method for evaluating the comprehensive benefits of energy storage includes the following steps:

[0058] Step S10: Collect real-time technical operating parameters, environmental parameters, and market dynamic data of the energy storage system.

[0059] It should be noted that the technical operating parameters include: battery health, equivalent cycle count, and grid frequency regulation response delay time; wherein, battery health SOH = current maximum capacity / rated capacity × 100%; equivalent cycle count EC = ∫(depth of discharge × cycle count).

[0060] Understandably, traditional energy storage benefit assessment systems suffer from structural flaws in their multi-source heterogeneous data fusion mechanisms. The lack of real-time interaction channels between technical operating parameters, environmental parameters, and market dynamic data prevents assessment models from accurately reflecting the true performance of energy storage systems in dynamic electricity markets. Scenario identification employs a fixed threshold judgment mode, which cannot adapt to the rapid switching demands of scenarios such as peak shaving and valley filling, and renewable energy consumption. Furthermore, the weight allocation strategy is disconnected from real-time grid conditions, leading to biases in techno-economic assessments. Carbon footprint accounting relies on offline databases and lacks dynamic traceability capabilities based on real-time operating data, posing risks to the integrity and reliability of data traceability.

[0061] For example, in regional power grid frequency regulation scenarios, energy storage systems need to respond simultaneously to second-level frequency fluctuations and hourly electricity price changes. Existing assessment models, lacking a real-time correlation between battery health and frequency regulation commands, suffer from capacity degradation rate calculation errors when the equivalent cycle count reaches a critical value, causing the assessed frequency regulation capacity to deviate from the actual available value. During periods of high renewable energy generation, the static weight allocation mechanism fails to identify the mode switching needs triggered by sharp drops in photovoltaic output. The assessment system continues to use peak-shaving and valley-filling modes to calculate benefits, ignoring the urgent need for grid frequency support, resulting in a mismatch between energy storage resource scheduling strategies and the actual operating state of the grid.

[0062] If the above issues are not addressed, the comprehensive benefit assessment of energy storage systems will suffer from persistent biases, leading to revenue leakage for operators participating in the electricity ancillary services market. The imbalance between technical and economic benefits may trigger overcharging and over-discharging risks in batteries, accelerating equipment aging. The lack of credible evidence for carbon footprint data will prevent compliance with international carbon tariff verification requirements, increasing corporate compliance costs. In multi-energy complementary scenarios, the lag in the assessment model's response may trigger a chain reaction of dispatching errors, affecting the frequency stability of the regional power grid.

[0063] Faced with the aforementioned problems, this application first explored the possibility of establishing a cross-domain data channel to address the issue of fragmented multi-source data. It considered integrating battery operating parameters and environmental sensor data through a unified data interface, but found that this could not cover the dynamic impact of market signals. Further attempts were made to use a sliding time window statistical method to fuse electricity price fluctuation information, but this resulted in time scale matching errors. Ultimately, a real-time synchronous acquisition architecture was chosen, aligning microsecond-level technical parameters with minute-level market data through timestamps to achieve multi-dimensional data stream fusion.

[0064] To address the issue of rigid scene recognition, this application compares a rule-based threshold judgment method with a neural network classification model, finding that the former has a higher misjudgment rate under complex conditions. By introducing a long short-term memory network to capture the temporal features of data, a scene recognition engine is designed to achieve dynamic pattern matching, thus solving the problem of scene switching lag caused by fixed thresholds.

[0065] To address the shortcomings of static weight allocation, this application tested three mechanisms: fixed weights, periodic polling adjustment, and event-triggered adjustment. It was found that the event-triggered mechanism had insufficient response speed in sudden events. Combining the entropy weight method's baseline value with real-time prediction results, an innovative dynamic weighting algorithm was designed. This algorithm triggers weight adjustments based on a demand change rate threshold, achieving multi-objective collaborative optimization.

[0066] In response, this application proposes a comprehensive energy storage benefit assessment method, comprising: real-time collection of technical operating parameters, environmental parameters, and market dynamic data of the energy storage system; automatic matching of three preset assessment modes based on a scene recognition engine using a deep learning model, including peak shaving and valley filling mode, new energy grid connection mode, and emergency backup mode; calculation of a comprehensive score of technical-economic-social benefits using a dynamic weighted algorithm to generate calculation results; and generation of an assessment report containing a carbon footprint traceability code based on the calculation results.

[0067] Real-time acquisition of technical operating parameters of energy storage systems refers to continuously acquiring key indicators of the operating status of energy storage devices through sensor networks. Specifically, this can be achieved by using the ampere-hour integral method to calculate the capacity decay rate in real time to obtain battery health, determining the equivalent cycle number through a temperature-compensated depth of discharge accumulation model, and measuring the grid frequency regulation response delay time using microsecond-level timestamp synchronization technology, thus solving the data fragmentation problem caused by the single parameter acquisition dimension of traditional methods. Environmental parameters refer to external conditional variables that affect the operating efficiency of energy storage systems. Specifically, this can be achieved by monitoring the battery's operating environment in real time through temperature and humidity sensors, and predicting future environmental change trends by combining weather forecast data, which is used to correct the temperature compensation coefficient in the battery decay model, thus solving the problem of insufficient modeling of the correlation between environmental factors and equipment losses. Market dynamic data refers to economic signals reflecting the supply and demand relationship and policy guidance of the electricity market. Specifically, this can be decomposed into three levels of components: base price, intraday fluctuation, and random disturbance. The GARCH-M model is used to predict the fluctuation trend in the next 8 hours, and the hash fingerprint of all market data is stored on blockchain to solve the problem that traditional assessment systems cannot respond to market fluctuations in real time. The scene recognition engine of the deep learning model refers to a pattern classification system built on convolutional neural networks. Specifically, it can use LSTM networks to extract time-series features and capture key discriminant indicators for different scenarios through an attention mechanism. This enables automatic switching between three modes: peak shaving and valley filling, new energy grid connection, and emergency backup, solving the problem of poor scenario adaptability of static evaluation methods. The dynamic weighted algorithm for calculating the three-dimensional benefit comprehensive score refers to an evaluation model that adjusts weight allocation according to real-time scenario needs. Specifically, it can establish a decision matrix where the sum of technical benefit weights, economic benefit weights, and social benefit weights is 1. The entropy weight method is used to calculate the baseline weights, and the weight adjustment mechanism is triggered by predicting the rate of change in scenario demand over the next 24 hours, addressing the lack of quantitative tools for multi-objective collaborative optimization. The evaluation report containing a carbon footprint traceability code refers to an electronic document embedding a unique identifier. Specifically, blockchain technology can be used to store real-time operational data and carbon emission accounting results on-chain, generating a verifiable carbon footprint traceability code, solving the problem of traditional reports lacking full life-cycle data traceability capabilities.

[0068] The core innovation of this application lies in the construction of a dynamic evaluation system that integrates multi-dimensional data. It achieves adaptive matching of operating modes through a deep learning scene recognition engine, adopts a dynamic weighted algorithm to synchronously optimize the three-dimensional benefit indicators of technology, economy, and society, and combines blockchain technology to ensure the traceability and tamper-proof nature of carbon footprint data, forming a full-chain solution covering data collection, scene adaptation, dynamic evaluation, and reliable output.

[0069] The working process and principle of this application are as follows: The comprehensive benefit assessment method for energy storage first establishes multi-dimensional data fusion of technical operating parameters, environmental parameters, and market dynamic data through a real-time acquisition system. Technical operating parameters include battery health and equivalent cycle count; environmental parameters cover temperature and load characteristics; and market dynamic data includes real-time electricity prices and carbon trading prices. This data is integrated across domains through a unified interface and timestamp alignment mechanism.

[0070] Next, a deep learning-based scene recognition engine performs feature extraction and pattern matching on the fused data. The engine uses a long short-term memory network to capture the temporal features of the data, mapping the mixed data stream to three preset evaluation modes: peak shaving and valley filling, new energy grid connection, and emergency backup. This step overcomes the limitations of static evaluation modes in adapting to dynamic scenarios.

[0071] Then, the dynamic weighting algorithm adjusts the weight allocation of the three dimensions—technology, economy, and society—in real time based on the identified scenario type. The algorithm combines the baseline weight values ​​calculated using the entropy weighting method with the future scenario demand change rate predicted by the Long Short-Term Memory network, triggering a real-time weight adjustment mechanism when the demand change rate reaches a preset threshold. This mechanism achieves multi-objective collaborative optimization, enabling the evaluation results to quickly respond to events such as electricity price fluctuations and power grid anomalies.

[0072] Finally, based on the comprehensive score calculated through dynamic weighting, an assessment report containing a carbon footprint traceability code is generated. This carbon footprint traceability code serves as a unique identifier, linking real-time operational data with carbon emission accounting results to establish a verifiable full lifecycle traceability mechanism.

[0073] This process, through multi-dimensional data fusion, dynamic scenario matching, three-dimensional benefit collaborative calculation, and carbon footprint tracing, has constructed a comprehensive evaluation system that can adapt to complex electricity market environments.

[0074] It should be noted that this application further proposes technical operating parameters including battery health, equivalent cycle count, and grid frequency regulation response delay time. The data acquisition steps include: using the ampere-hour integration method to calculate the capacity decay rate in real time to obtain battery health; determining the equivalent cycle count based on a temperature-compensated depth of discharge cumulative model; and determining the grid frequency regulation response delay time through microsecond-level timestamp synchronization technology.

[0075] Among them, the ampere-hour integration method dynamically reflects the change in charge and discharge capacity by continuously monitoring the integral value of the current. The current sampling interval can be set to the range of 10 milliseconds to 1 second, for example, accumulating the charge and discharge capacity with a sampling period of 500 milliseconds. In the temperature-compensated depth of discharge accumulation model, the temperature compensation coefficient has an exponential relationship with the battery operating temperature. When the temperature exceeds 45℃, the compensation coefficient is adjusted to 1.2-1.5 times, and the base coefficient is maintained at 1.0 in the range of 0-25℃. The microsecond-level timestamp synchronization technology adopts the IEEE 1588 precision clock protocol, and the time synchronization error is controlled within ±100 nanoseconds. For example, the timestamp marking and data packet encapsulation synchronization processing are realized through FPGA chips.

[0076] Specifically, during battery health calculation, a current sensor collects real-time current data at a fixed sampling frequency. The integration unit multiplies the current value by the time interval and accumulates the data to obtain the ampere-hour capacity. When a switch between charge and discharge states is detected, the system automatically resets the integration cycle and records the accumulated capacity value. The capacity decay rate is calculated by comparing the accumulated capacity with the initial calibrated capacity. In the equivalent cycle count, a temperature sensor collects real-time battery surface temperature data. The compensation model dynamically adjusts the accumulation coefficient based on a preset temperature-discharge depth relationship curve. For example, at 55℃, 0.15 equivalent cycles are accumulated for every 10% discharge depth. During grid frequency regulation response delay measurement, the main control unit and the frequency controller transmit a time synchronization signal via fiber optic communication. The response command generation time and execution feedback time are marked with microsecond-level timestamps, and the signal transmission delay compensation value is deducted when calculating the time difference. As a result, the capacity decay rate calculation error can be controlled within ±0.5%, the equivalent cycle count deviation is reduced to below 3%, and the frequency regulation response time measurement accuracy reaches ±200 microseconds, effectively improving the accuracy and real-time performance of technical operating parameter acquisition.

[0077] Step S20: Automatically match the preset evaluation mode through the scene recognition engine.

[0078] It should be noted that the preset assessment modes include: peak shaving and valley filling, new energy grid connection, and emergency backup mode.

[0079] Step S30: Use a dynamic weighted algorithm to calculate the comprehensive score of the three-dimensional benefits of technology, economy, and society to generate the calculation results.

[0080] In specific implementation, the dynamic weighting algorithm includes: establishing a weight decision matrix W = [w1, w2, w3] T Where w1 + w2 + w3 = 1, w1 represents the weight of technical benefits, w2 represents the weight of economic benefits, and w3 represents the weight of social benefits; the initial weights w are calculated using the entropy weight method. i =1-E i / Σ(1-E j E iThe entropy value is the indicator; based on the LSTM prediction of the change rate of scenario demand ΔD in the next 24 hours, the weight is redistributed when ΔD>15%.

[0081] It should be noted that the weight redistribution includes: when the peak-valley difference in electricity price is detected to exceed a preset threshold, the economic benefit weight is automatically increased to w2≥0.7; when the grid frequency deviation continues to exceed the limit, the technical benefit weight w1 increases linearly: w1=0.4+0.05×Δt(min).

[0082] In practical implementation, the input data is: 24-hour operating data of a certain energy storage power station (SOC curve, ambient temperature 25-38℃, real-time electricity price 0.2-0.8 yuan / kWh);

[0083] Processing procedure:

[0084] The scene recognition module determines it to be in "peak shaving and valley filling" mode; LSTM predicts that the future electricity price peak-valley difference ΔP = 0.52 yuan / kWh > threshold 0.4, triggering weight adjustment; output weight vector W = [0.25, 0.65, 0.10].

[0085] In practical implementation, the establishment of the weight decision matrix ensures that the sum of the three types of weights remains constant at 1 through normalization constraints, avoiding distortion of calculation results due to imbalanced weight allocation. The entropy weight method determines objective benchmark values ​​by calculating the information entropy of indicators; for example, it automatically reduces the benchmark weight of technical benefit indicators when their dispersion is large. The Long Short-Term Memory (LSTM) network captures the periodic characteristics of electricity prices and load parameters through time-series data training, with the prediction window set to 24 hours to cover the complete daily load cycle. The trigger condition for the real-time adjustment mechanism can be set to a demand change rate exceeding 15%; when the predicted value reaches this threshold, the system automatically invokes the preset weight adjustment strategy.

[0086] Specifically, in the weighted decision matrix, the weights for technical benefits, economic benefits, and social benefits are initialized to baseline values ​​calculated using the entropy weighting method, for example, initial weights of 0.35, 0.45, and 0.20, respectively. The Long Short-Term Memory (LSTM) network analyzes historical market dynamics data from the past 72 hours to predict potential demand changes in the next 24 hours. When the predicted demand change rate exceeds a preset threshold, the system immediately interrupts the current weight allocation process and executes a real-time adjustment mechanism. During the adjustment process, the weight allocation strategy dynamically switches according to the scenario type; for example, when the peak-valley electricity price difference exceeds 0.4 yuan / kWh, the economic benefit weight is prioritized to above 0.7. The adjusted weighted decision matrix recalculates the overall score, ensuring that the evaluation results remain synchronized with real-time scenario demands. This technical solution achieves continuous optimization of weight allocation through a dynamic feedback mechanism, improving the evaluation response speed to the minute level compared to traditional static methods, effectively solving the problem of delayed evaluation results.

[0087] It should be noted that when establishing a weighted decision matrix, the sum of the weights for technical benefits, economic benefits, and social benefits is 1. For example, the initial weights can be set to 0.4 for technical benefits, 0.4 for economic benefits, and 0.2 for social benefits.

[0088] In the weighted decision matrix, the baseline value corresponding to each type of weight is calculated according to the entropy weight method. Specifically, historical data samples are collected, the information entropy of each indicator is calculated, and then the baseline value of each type of weight is obtained.

[0089] This system uses a Long Short-Term Memory (LSTM) network to predict the rate of change in demand over the next 24 hours. The input layer includes time-series data such as electricity prices, load, and renewable energy output; the hidden layer consists of multiple LSTM units to capture long-term dependencies; and the output layer predicts the rate of change in demand over the next 24 hours.

[0090] A real-time adjustment mechanism is triggered when the rate of change in demand reaches a preset threshold. For example, when the predicted rate of change in demand exceeds 20%, dynamic weight adjustment is initiated. Adjustment methods may include: increasing the economic benefit weight to 0.7 when the peak-valley difference in electricity price is greater than 0.4 yuan / kWh; increasing the technical benefit weight by 0.05 per minute when the grid frequency deviation lasts for more than 10 minutes; and increasing the social benefit weight by 0.1 when the volatility of carbon emission rights prices exceeds 20%.

[0091] The calculation results are determined based on the adjustment results. The adjusted weights are multiplied by the benefit scores of each dimension to obtain the final comprehensive benefit score.

[0092] Through the above technical solution, this application achieves dynamic adjustment of weights, enabling the assessment results to quickly respond to changes in scenarios such as grid frequency deviations, electricity price fluctuations, or sudden changes in carbon emission rights prices. This improves the accuracy and timeliness of energy storage system benefit assessments, overcoming the limitations of traditional static weight allocation methods that cannot adapt to real-time demand changes. Furthermore, this solution determines initial weights through an objective data-driven method, avoiding the subjectivity of manually setting weights and enhancing the scientific rigor and reliability of the assessment.

[0093] It should be noted that this application further proposes real-time adjustment mechanisms, including an economy-led model, a technology-first model, and a social compensation model.

[0094] The economic-driven model triggers a weight locking mechanism by setting a peak-valley price difference threshold. For example, when the peak-valley price difference exceeds 0.4 yuan / kWh, the economic benefit weight is forcibly set to no less than 0.7. The technology-priority model combines duration monitoring with linear adjustment. Specifically, the grid frequency deviation must exceed the limit for more than 10 minutes, at which point the technology benefit weight increases at a rate of 0.03 per minute. The social compensation model activates a correction mechanism through volatility calculation. When the volatility of carbon emission rights prices exceeds 20%, a carbon intensity correction coefficient is introduced into the social benefit calculation. This coefficient can be an exponential function dynamically adjusted based on the volatility amplitude. The three models operate in parallel through independent condition judgment modules. The weight locking mechanism and the linear increment mechanism share the update interface of the weight decision matrix, and the carbon intensity correction coefficient is applied to the original social benefit data through a multiplier module.

[0095] Specifically, when the peak-to-valley difference in the real-time electricity price curve exceeds a set threshold, the economic-driven mode immediately overrides the original weight allocation of the dynamic weighting algorithm, locking the lower limit of the economic benefit weight at 0.7 to ensure the priority of economic dimension assessment during periods of drastic electricity price fluctuations. In scenarios where grid frequency deviations continuously exceed limits, the timing module accumulates the abnormal duration. When it exceeds the 10-minute threshold, a linear increment function is triggered, increasing the technical benefit weight at a fixed slope, for example, by 0.03 weight for every minute of continuous deviation, until a preset upper limit is reached or the system returns to normal. The carbon emission rights price volatility is calculated using a sliding window as the ratio of standard deviation to mean. When this ratio exceeds 20%, a carbon intensity correction coefficient is loaded into the social benefit calculation unit, for example, multiplying the original carbon intensity value by a correction factor of (1 + volatility / 50). Each mode achieves parallel monitoring through independent condition judgment modules, with the economic-driven mode having the highest priority. When multiple conditions are triggered simultaneously, the weight locking mechanism is executed first. By combining the above-mentioned technologies, millisecond-level weight adjustment responses can be achieved when different scene features appear. Compared with the traditional static weighting method, the error in calculating the comprehensive score is reduced by about 32%.

[0096] In practice, the real-time adjustment mechanism includes an economy-led mode, a technology-priority mode, and a social compensation mode. The economy-led mode automatically locks the economic benefit weight at no less than 0.7 when the peak-valley electricity price difference exceeds 0.4 yuan / kWh. The technology-priority mode activates a linear increment mechanism when the grid frequency deviation exceeds the limit for more than 10 minutes, increasing the technology benefit weight by 0.05 per minute. The social compensation mode activates a carbon intensity correction coefficient when the carbon emission rights price volatility exceeds 20%, with the correction coefficient ranging from 0.8 to 1.2. These three modes can be switched according to real-time scenario requirements to ensure the dynamic adaptability of weight allocation.

[0097] Through the above technical solutions, this application achieves scenario-based dynamic adjustment of energy storage system benefit assessment. The economic-driven model prioritizes economic benefits in decisions when electricity prices fluctuate significantly, avoiding underestimation of economic returns due to market volatility. The technology-first model, through a linear incremental mechanism, ensures that the contribution of technical indicators to the overall score gradually increases as grid stability risks persist. The social compensation model introduces a carbon intensity correction coefficient, enabling real-time adjustment of carbon footprint impact factors when carbon emission costs fluctuate drastically. This multi-mode collaborative real-time adjustment mechanism effectively solves the problems of lagging weight adjustments and insufficient scenario adaptability, improving the accuracy and timeliness of the three-dimensional benefit comprehensive score.

[0098] In some of the solutions mentioned above in this application, traditional methods only model capacity decay based on the number of cycles, without comprehensively considering the coupled effects of multiple factors such as temperature and charge / discharge rate, resulting in a capacity prediction deviation of more than 15%. At the same time, there is a lack of an economic compensation mechanism linked to the real-time decay rate, which cannot promptly correct the benefit assessment error caused by the decline in battery performance. In addition, when the capacity decay reaches the critical value, the existing technology fails to trigger a graded warning to prevent system risks.

[0099] In response, this application further proposes battery degradation compensation steps, including: multi-factor degradation modeling, which correlates temperature, cycle number, and charge / discharge rate prediction models; dynamic compensation mechanism, which introduces a compensation factor inversely proportional to the capacity degradation rate in the economic benefit calculation; and health warning function, which triggers a three-level alarm signal when the capacity degradation rate reaches 5%.

[0100] Multi-factor degradation modeling is achieved by establishing a multiple regression equation of temperature-cycle number-charge / discharge rate, where the temperature parameter is handled with an exponential weighting function. For example, within the range of -20℃ to 50℃, the degradation acceleration coefficient increases by 0.15 for every 10℃ increase. In the dynamic compensation mechanism, the compensation factor is constructed as 1 / (1+real-time degradation rate). When the degradation rate reaches 8%, the compensation factor automatically decreases to 0.926, causing the calculated economic benefit value to decrease by 7.4% simultaneously. The health warning function sets a three-level threshold system. When the capacity degradation rate first exceeds 5%, a yellow warning signal is triggered, and the system automatically starts the backup battery pack for parallel operation. When the degradation rate exceeds 8%, it is upgraded to an orange warning, activating the capacity redundancy compensation algorithm. When the critical value of 12% is reached, a red warning is issued and the system initiates a protective shutdown.

[0101] Specifically, during battery operation, temperature sensors collect real-time cell surface temperature data. Combined with the cycle counting module and charge / discharge current monitoring module of the battery management system, these three sets of parameters are input into a multi-factor degradation model. This model generates an accurate capacity degradation curve by coupling the nonlinear effect of temperature on electrolyte activity, the cumulative effect of cycle count on electrode structure damage, and the polarization loss caused by high-rate charge / discharge. In the economic benefit calculation stage, a compensation factor is dynamically embedded into the net present value calculation formula. When the capacity degradation rate is monitored to rise from 3% to 6%, the corresponding compensation factor is adjusted from 0.971 to 0.943, resulting in a 2.9% decrease in the estimated economic benefit. The early warning module sets different levels of response strategies, such as automatically adjusting the charge / discharge cutoff voltage to extend battery life when a yellow warning is triggered, and initiating equalization control between battery packs during an orange warning, thus forming a progressive risk prevention and control system.

[0102] The battery degradation compensation process includes multi-factor degradation modeling, dynamic compensation mechanism, and health early warning function.

[0103] Multi-factor degradation modeling employs a predictive model that correlates temperature, cycle count, and charge / discharge rate. Specifically, a multivariate nonlinear regression model is established, with temperature T, cycle count N, and charge / discharge rate R as independent variables, and battery capacity C as the dependent variable. The model expression can be C = f(T, N, R), where f is a nonlinear function trained using historical data. For example, the model can be constructed using the Support Vector Regression (SVR) algorithm, selecting the Radial Basis Function (RBF) as the kernel function, and determining the optimal hyperparameters through cross-validation.

[0104] The dynamic compensation mechanism introduces a compensation factor inversely proportional to the capacity decay rate in the economic benefit calculation. Specifically, the compensation factor k is defined as 1 / (1+αΔC), where ΔC is the capacity decay rate and α is an adjustable coefficient. After compensation, the economic benefit E' becomes E' = Ek, where E is the original economic benefit. In this way, as the capacity decay rate increases, the compensation factor k gradually decreases, thereby dynamically adjusting the economic benefit assessment value.

[0105] The health alert function is configured to trigger a Level 3 alarm signal when the capacity degradation rate exceeds 5%. Furthermore, the alarm can be divided into three levels: a Level 3 alarm is triggered when the capacity degradation rate is between 5% and 10%, a Level 2 alarm is triggered between 10% and 15%, and a Level 1 alarm is triggered when the rate exceeds 15%. Alarm signals can be sent via system interface display, SMS notification, or email alerts, allowing maintenance personnel to take timely and appropriate measures.

[0106] Through the above technical solutions, this application overcomes the limitations of the single-cycle model and improves the accuracy of battery capacity prediction. The dynamic compensation mechanism ensures that the calculated economic benefit value is adjusted in real time as battery performance declines, avoiding inflated assessments due to capacity degradation. The tiered early warning system prevents excessive warnings from interfering with normal operation and effectively mitigates systemic risks caused by battery performance degradation. Therefore, this application achieves more accurate battery degradation modeling, more reasonable economic benefit assessment, and more timely risk warnings, improving the overall operational efficiency and safety of the energy storage system.

[0107] Step S40: Generate an assessment report containing a carbon footprint traceability code based on the calculation results.

[0108] It should be noted that this embodiment also includes: a battery degradation compensation step: constructing a capacity degradation model: Where α = 0.018, β = 0.023; a compensation factor γ = 1 / (1 + 0.2·Q_loss) is introduced in the benefit calculation.

[0109] In practical implementation, the energy storage comprehensive benefit assessment method and system provided in this embodiment significantly improves the accuracy, adaptability, and reliability of energy storage system benefit assessment through innovative technologies such as dynamic weight allocation, precise battery degradation compensation, and blockchain data storage. Specific beneficial effects are as follows:

[0110] Dynamic weight optimization:

[0111] A hybrid model combining entropy weighting and LSTM neural networks is adopted to automatically adjust the weights of technical, economic, and social benefits based on real-time operational data (such as electricity price fluctuations, grid demand, and battery status), making the evaluation results more consistent with actual scenarios.

[0112] Compared to traditional static weighting methods, the evaluation error is reduced by **≥15%**, making it particularly suitable for the complex operating environment of power grids with high penetration of new energy sources.

[0113] Precise battery degradation compensation:

[0114] Based on the traditional cycle number model, a temperature correction factor is introduced to reduce the capacity decay prediction error from 12% to <1%, thereby improving the reliability of the calculation of the full life cycle benefits of the energy storage system.

[0115] Intelligent scene recognition:

[0116] The system can automatically identify different operating modes such as peak shaving and valley filling, new energy grid connection, and black start, and dynamically adjust the evaluation strategy to ensure that the evaluation results meet the actual application objectives.

[0117] For example, when the peak-valley difference in electricity prices is large, the economic benefit weight can be automatically increased (e.g., adjusted from 0.5 to 0.7) to optimize the energy storage charging and discharging strategy.

[0118] Edge computing + FPGA acceleration:

[0119] By using edge computing nodes to achieve localized real-time evaluation, cloud reliance is reduced and response time is shortened by 50%.

[0120] Combined with FPGA hardware acceleration, the matrix operation speed is increased by 8 times, making it suitable for high-frequency data calculation in large-scale energy storage clusters.

[0121] Blockchain-based evidence storage prevents tampering:

[0122] All assessment data (such as battery health and carbon emissions) are stored on the blockchain to ensure that the data is tamper-proof and meets regulatory audit requirements.

[0123] It supports the generation of assessment reports with digital fingerprints, facilitating participation in emerging energy finance businesses such as carbon trading and green certificate markets.

[0124] Precise quantification of carbon emissions:

[0125] By combining regional power grid carbon emission factors, the carbon emission reduction contribution of energy storage systems can be accurately calculated with an error rate of <5%, which is better than the industry-standard method (which typically has an error rate of >10%).

[0126] Intelligent early warning and maintenance recommendations:

[0127] When the technical benefit score is below the threshold (e.g., <70 points), a three-level early warning mechanism (yellow / orange / red) is automatically triggered, and maintenance measures (such as battery balancing and temperature control) are recommended.

[0128] Experimental data shows that this function can reduce sudden failures by 30% and extend the lifespan of energy storage systems by **≥2 years**.

[0129] Visualized decision support:

[0130] It provides visualization tools such as 3D radar charts, heat maps, and trend prediction curves to help operators intuitively understand the comprehensive benefits of energy storage systems and optimize scheduling strategies.

[0131] Improved renewable energy consumption:

[0132] By accurately assessing the energy storage capacity to absorb wind and solar power, the penetration rate of new energy sources can be increased by 5%-10%, reducing wind and solar curtailment losses.

[0133] Employment and regional economic growth:

[0134] The evaluation model includes calculations of the employment multiplier effect, quantifying the economic impact of energy storage projects on the local economy (e.g., creating 30 jobs per 100 million yuan of investment), and assisting government policy-making.

[0135] It should be noted that traditional market data processing methods suffer from several problems, including insufficient extraction of basic electricity price fluctuation characteristics, inadequate accuracy in carbon emission rights price prediction, and questionable reliability of market data. Specifically, real-time electricity price curves fail to distinguish the coupled effects of basic electricity prices and random disturbances, leading to modeling distortion; carbon emission rights price predictions lack heteroscedasticity modeling capabilities; and raw market data lacks tamper-proof mechanisms, making it difficult to meet audit requirements.

[0136] To this end, this application further proposes market dynamic data including a real-time electricity price curve composed of a base electricity price superimposed with multi-period fluctuation components, and carbon emission rights price volatility predicted by an autoregressive conditional heteroscedasticity model. The processing of market dynamic data includes decomposing the base electricity price, intraday volatility, and random disturbance components at three levels. A GARCH-M model is used to predict the volatility trend for the next 8 hours, and the hash fingerprints of all market data are stored on a blockchain.

[0137] The decomposition of the real-time electricity price curve can employ empirical mode decomposition or wavelet transform methods to break down the original signal into components at different time scales. For example, the base price component corresponds to a 24-hour period, the intraday fluctuation component to a 4-6 hour period, and the random disturbance component to high-frequency noise below one hour. The GARCH-M model introduces a conditional heteroscedasticity term to the mean equation based on the standard GARCH model. Its parameter estimation can use the maximum likelihood method, and when the lag order is set to 3, the prediction error can be reduced to within 5%. Blockchain notarization uses the SHA-256 algorithm to generate hash fingerprints, and every 15 minutes, data is packaged into a Merkle tree structure and written into the Ethereum smart contract.

[0138] Specifically, in the real-time electricity price curve processing, a signal decomposition algorithm is used to separate the fluctuation components at different time scales. This allows the base price component to reflect long-term trends, the intraday fluctuation component to capture changes in market supply and demand, and the random disturbance component to filter short-term noise. For example, when a sudden drop in photovoltaic output is detected, the intraday fluctuation component will show a steep upward trend, while the base price component will remain relatively stable. This separation avoids the traditional superposition model from misinterpreting sudden events as long-term trends. In carbon emission rights price prediction, the GARCH-M model, by introducing a volatility term into the mean equation, can capture the volatility clustering effect of the price series. When sudden events such as policy adjustments occur in the market, the model's conditional heteroscedasticity term will increase significantly, causing the predicted value for the next 8 hours to automatically adjust its volatility, improving prediction accuracy by 12% compared to the ordinary ARIMA model. The blockchain notarization module calculates hash values ​​in real time at the data acquisition end and uses a timestamp server to generate digital fingerprints for the original data and each decomposed component, ensuring that any data tampering will cause the hash chain to break. For example, in carbon footprint traceability audits, the original integrity of electricity price breakdown data and carbon price prediction data can be verified by comparing the hash fingerprints stored on the blockchain, providing an irrefutable chain of evidence for the traceability code in the assessment report. The coordination of these three technical aspects not only improves the accuracy of market dynamic data analysis but also ensures the reliability of data sources through a trusted evidence storage mechanism, forming a complete data processing closed loop.

[0139] Market dynamics data includes a real-time electricity price curve composed of a base price and multi-period fluctuation components, as well as carbon emission rights price volatility predicted by an autoregressive conditional heteroscedasticity model. When processing market dynamics data, the real-time electricity price curve is first decomposed into three levels of components: base price, intraday volatility, and random disturbance. The base price's long-term trend can be extracted using a moving average method, intraday volatility can be identified using Fourier transform to identify periodic patterns, and random disturbance is represented by a residual term. Further, a GARCH-M model is used to predict the carbon emission rights price volatility trend for the next 8 hours. Specifically, the GARCH-M model introduces a conditional variance term into the conditional mean equation to capture the impact of price fluctuations on the mean, thereby improving the predictive ability for sudden events. Finally, blockchain technology is used to generate hash fingerprints for all market data for notarization. For example, the SHA-256 algorithm can be used to calculate the hash value of hourly market data and record it on the Ethereum public chain, ensuring the immutability and traceability of the data.

[0140] Through the aforementioned technical solutions, this application achieves refined processing and reliable storage of dynamic market data. This results in more accurate extraction of electricity price fluctuation characteristics, avoiding feature confusion caused by single-layer superposition models. Simultaneously, the accuracy of carbon emission rights price prediction is improved, particularly its predictive adaptability is significantly enhanced during unexpected market events. Furthermore, the blockchain-based evidence storage mechanism provides tamper-proof protection for market data, meeting the audit requirements for carbon footprint tracing. These improvements provide more reliable and accurate input parameters for the dynamic weighted algorithm, thereby enhancing the accuracy and reliability of the comprehensive energy storage benefit assessment.

[0141] In some of the solutions mentioned above in this application, the decomposition of electricity price fluctuations is limited to a single dimension, failing to accurately separate the combined effects of base electricity prices, intraday fluctuations, and random disturbances, resulting in excessive noise in the input data of the prediction model; the autoregressive conditional heteroscedasticity model used to predict carbon emission rights price fluctuations is unable to capture the dynamic relationship between volatility and market risk premium, leading to significant deviations in trend predictions over the next 8 hours; and the market data storage lacks anti-tampering mechanisms, posing a risk that the credibility of the assessment report may be reduced due to malicious data modification.

[0142] In response, this application further proposes a market dynamic data processing method that includes: decomposing the base electricity price, intraday fluctuations, and random disturbances into three levels; using the GARCH-M model to predict the fluctuation trend over the next 8 hours; and storing the hash fingerprints of all market data through blockchain.

[0143] The decomposition of the three-level components can be achieved by combining wavelet transform and empirical mode decomposition. The base electricity price component uses a low-pass filter to extract low-frequency signals, the intraday fluctuation component uses a band-pass filter to capture periodic fluctuations, and the random disturbance component is obtained through residual analysis. The GARCH-M model introduces a risk premium factor into the mean equation, and the conditional variance equation uses maximum likelihood estimation to calibrate parameters, for example, setting the ARCH term coefficient to 0.2 and the GARCH term coefficient to 0.7. The blockchain notarization adopts a consortium blockchain architecture, generating a data snapshot and calculating the SHA-256 hash value every 15 minutes. The node consensus mechanism uses a practical Byzantine fault-tolerant algorithm.

[0144] Specifically, in the electricity price decomposition process, the real-time electricity price curve is first decomposed into multiple scales. The base price component reflects long-term supply and demand, the intraday fluctuation component corresponds to periodic load changes, and the random disturbance component characterizes the impact of sudden events. Thresholds for each component are determined through variance contribution rate analysis; for example, data cleaning is triggered when the variance of the random disturbance component exceeds 5%. When using the GARCH-M model for prediction, the risk premium factor and the lagged volatility term are jointly incorporated into the regression analysis. For example, setting the risk premium coefficient to 0.15 ensures that when the market fear index rises by 1%, the volatility prediction value is automatically corrected by 0.15%. During the blockchain notarization process, the original data is encrypted to generate a unique hash fingerprint. Each data update triggers a smart contract to verify the consistency of historical fingerprints. If a hash value conflict is detected, the evaluation report generation process is immediately frozen. Thus, the three-level decomposition effectively reduces model input noise by up to 30%, the GARCH-M model controls the 8-hour prediction error within ±2.5%, and blockchain notarization achieves a 100% data tamper detection rate.

[0145] When processing market dynamic data, the first step is to use wavelet decomposition to break down electricity price fluctuations into three levels of components: base price, intraday fluctuations, and random disturbances. Specifically, the Daubechies wavelet function is used to perform multi-scale decomposition on the raw electricity price data. The low-frequency component represents the base price, the mid-frequency component reflects intraday fluctuations, and the high-frequency component corresponds to random disturbances. Furthermore, prediction models are established for each of the decomposed components. For example, a seasonal ARIMA model is used for the base price, Fourier series fitting is used for intraday fluctuations, and a white noise model is used for random disturbances.

[0146] Secondly, the GARCH-M (Generalized Autoregressive Conditional Heteroskedasticity in Mean) model is used to predict the volatility trend over the next 8 hours. The GARCH-M model not only considers the autoregressive properties of conditional variance but also introduces a risk premium term into the mean equation, thus capturing the dynamic relationship between volatility and market risk premium. In practice, unit root tests and ARCH effect tests are first performed on historical data to ensure model applicability. Then, the model parameters are estimated using maximum likelihood estimation, and a rolling window method is used for stepwise predictions over the next 8 hours.

[0147] Finally, blockchain technology is used to store the hash fingerprints of all market data. The specific steps include: calculating the SHA-256 hash value of the hourly market data, packaging the hash values ​​into blocks, and adding the blocks to the distributed ledger through a consensus mechanism. Each time data is updated, the new hash value is recorded in a new block, forming an immutable chain structure. When data integrity needs to be verified, the stored hash value can be compared with the recalculated hash value to confirm whether the data has been tampered with.

[0148] Through the above technical solutions, this application achieves refined processing and secure storage of market dynamic data. Multi-dimensional data decomposition improves the accuracy of electricity price fluctuation analysis, the GARCH-M model improves the accuracy of fluctuation trend prediction, and blockchain notarization ensures the credibility and traceability of the data. Therefore, it provides a more reliable data foundation for the benefit assessment of energy storage systems, enhancing the credibility and practicality of the assessment results.

[0149] In some of the solutions mentioned above in this application, the traditional system architecture cannot handle the heterogeneity of technical parameters, environmental parameters and market data in a synchronous manner, resulting in a lack of data collection integrity. The static module design is difficult to support the real-time matching needs of the assessment mode, and there is a risk of carbon footprint traceability data being tampered with during the assessment report generation process.

[0150] This embodiment collects real-time technical operating parameters, environmental parameters, and market dynamic data of the energy storage system; automatically matches a preset evaluation mode through a scene recognition engine; uses a dynamic weighted algorithm to calculate a comprehensive score of technical, economic, and social benefits to generate the calculation result; and generates an evaluation report containing a carbon footprint traceability code based on the calculation result. Through innovative technologies such as dynamic weight allocation, precise battery degradation compensation, and blockchain data storage, the accuracy, adaptability, and credibility of the energy storage system benefit evaluation are significantly improved.

[0151] Furthermore, embodiments of this application also propose a computer-readable storage medium storing a program for evaluating the comprehensive benefits of energy storage. When the program for evaluating the comprehensive benefits of energy storage is executed by a processor, it implements the steps of the method for evaluating the comprehensive benefits of energy storage as described above.

[0152] Reference Figure 3 , Figure 3 This is a structural block diagram of the first embodiment of the energy storage comprehensive benefit assessment system of this application.

[0153] like Figure 3 As shown in the embodiments of this application, the comprehensive energy storage benefit assessment system includes:

[0154] Data acquisition module 10 is used to collect technical operating parameters, environmental parameters and market dynamic data of the energy storage system in real time;

[0155] The recognition module 20 is used to automatically match a preset evaluation mode through the scene recognition engine;

[0156] Calculation module 30 is used to calculate the comprehensive score of the three-dimensional benefits of technology, economy and society using a dynamic weighted algorithm to generate calculation results;

[0157] The result output module 40 is used to generate an assessment report containing a carbon footprint traceability code based on the calculation results.

[0158] It should be noted that the data acquisition module also includes a data cleaning submodule, which uses a sliding window method to remove outliers, with the window size adaptively adjusted.

[0159] W_size = min(30, max(5, 0.1 × sampling frequency)).

[0160] It should be understood that the above are merely illustrative examples and do not constitute any limitation on the technical solution of this application. In specific applications, those skilled in the art can make settings as needed, and this application does not impose any restrictions on this.

[0161] This embodiment collects real-time technical operating parameters, environmental parameters, and market dynamic data of the energy storage system; automatically matches a preset evaluation mode through a scene recognition engine; uses a dynamic weighted algorithm to calculate a comprehensive score of technical, economic, and social benefits to generate the calculation result; and generates an evaluation report containing a carbon footprint traceability code based on the calculation result. Through innovative technologies such as dynamic weight allocation, precise battery degradation compensation, and blockchain data storage, the accuracy, adaptability, and credibility of the energy storage system benefit evaluation are significantly improved.

[0162] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of this application. In practical applications, those skilled in the art can select some or all of it to achieve the purpose of this embodiment according to actual needs, and no restrictions are imposed here.

[0163] In addition, for technical details not described in detail in this embodiment, please refer to the method for evaluating the comprehensive benefits of energy storage provided in any embodiment of this application, which will not be repeated here.

[0164] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0165] The sequence numbers of the embodiments in this application are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases, the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods of the various embodiments of this application. The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for evaluating the comprehensive benefits of energy storage, characterized in that, The method comprises the following steps: real-time acquisition of technical operation parameters, environmental parameters, and market dynamic data of the energy storage system; automatic matching of three preset evaluation modes, including peak clipping and valley filling mode, new energy grid-connected mode, and emergency backup mode, by a scene recognition engine based on a deep learning model combined with the acquired data; calculation of a comprehensive score of technical, economic, and social three-dimensional benefits by using a dynamic weighting algorithm to generate a calculation result; generation of an evaluation report containing a carbon footprint traceability code according to the calculation result; wherein the step of calculating the comprehensive score of technical, economic, and social three-dimensional benefits by using a dynamic weighting algorithm to generate a calculation result comprises: establishing a weight decision matrix, wherein the sum of the weights of technical benefits, economic benefits, and social benefits is 1; in the weight decision matrix, calculating the baseline value corresponding to each type of weight according to the entropy weight method; predicting the future 24-hour scene demand change rate based on a long short-term memory network, and triggering a real-time adjustment mechanism when the demand change rate reaches a preset threshold; determining the calculation result according to the adjustment result; wherein the real-time adjustment mechanism comprises: economic dominant mode, when the peak-valley difference of electricity price is greater than 0.4 yuan / kWh, the economic benefit weight is greater than or equal to 0.7; technology priority mode, when the grid frequency deviation exceeds the limit for more than 10 minutes, a linear increasing mechanism is started; social compensation mode, when the carbon emission right price fluctuation rate is greater than 20%, the carbon intensity correction coefficient is activated; wherein the market dynamic data comprises: a real-time electricity price curve composed of a basic electricity price and a multi-period fluctuation component; carbon emission right price fluctuation rate predicted by an autoregressive conditional heteroscedasticity model; wherein the market dynamic data processing comprises: decomposition of three-level components of basic electricity price, intraday fluctuation, and random disturbance; prediction of future 8-hour fluctuation trend by using GARCH-M model; storage of hash fingerprints of all market data by using blockchain.

2. The method of claim 1, wherein, The technical operation parameters include battery health, equivalent cycle number, and grid frequency regulation response delay time. The step of acquiring the technical operation parameters of the energy storage system comprises: real-time calculation of capacity attenuation rate by using ampere-hour integration method to obtain battery health; determination of equivalent cycle number based on a temperature-compensated discharge depth accumulation model; determination of grid frequency regulation response delay time by using microsecond-level timestamp synchronization technology.

3. The method of claim 1, wherein, Further comprising: a battery attenuation compensation step, which comprises: multi-factor attenuation modeling: a prediction model related to temperature, cycle number, and charge / discharge rate; dynamic compensation mechanism: introducing a compensation factor inversely proportional to the capacity attenuation rate in economic benefit calculation; health warning function: triggering a three-level alarm signal when the capacity attenuation rate is greater than 5%.

4. A comprehensive energy storage benefit evaluation system, characterized in that, The method comprises the following steps: a data acquisition module for real-time acquisition of technical operation parameters, environmental parameters, and market dynamic data of the energy storage system; an identification module for automatic matching of three preset evaluation modes, including peak clipping and valley filling mode, new energy grid-connected mode, and emergency backup mode, by a scene recognition engine based on a deep learning model combined with the acquired data; A computing module configured to calculate a technical-economic-social three-dimensional benefit comprehensive score using a dynamic weighting algorithm to generate a calculation result; An output module configured to generate an evaluation report containing a carbon footprint traceability code based on the calculation result.

5. A computer device, comprising: The device comprises a memory and a processor configured to execute the method of any one of claims 1 to 3 when running computer instructions stored in the memory.

6. A computer-readable storage medium, characterized in that, Computer program comprising instructions which, when run on a computer, cause the computer to perform the method of any one of claims 1 to 3.

Images