A zero-carbon park energy supply optimization method and system based on big data analysis

Abstract

The present invention discloses a method and system for optimizing energy supply in a zero-carbon park based on big data analysis, including: S1. constructing a unified park energy operation data set; S2. utilizing big data processing technology to extract features and fuse data from the park energy operation data set to generate energy supply and demand forecast results for the future; S3. constructing a distributed energy storage control optimization model for a zero-carbon park; S4. utilizing an improved hybrid optimization algorithm to solve the distributed energy storage control optimization model for the zero-carbon park to generate multiple candidate solutions for distributed energy storage control strategies; S5. evaluating the candidate solutions for the distributed energy storage control strategies, selecting the optimal distributed energy storage control strategy that best matches the current park energy supply and demand status based on a pre-set fitness evaluation index, and using the optimal strategy as the control scheme. The present invention can effectively improve energy utilization efficiency, reduce energy storage loss and carbon emissions in a complex dynamic environment, and achieve the goal of intelligent energy supply optimization for a zero-carbon park.

Description

Technical Field

[0001] The present invention relates to the technical field of zero-carbon parks, and in particular to a method and system for optimizing energy supply in zero-carbon parks based on big data analysis. Background Art

[0002] As the global energy structure develops towards low-carbon and intelligent directions, zero-carbon parks, as an important carrier of future green energy utilization, have important application value in renewable energy consumption, distributed energy storage optimization and energy scheduling management. Zero-carbon parks usually integrate solar power generation, wind power generation, traditional power grid supply and distributed energy storage equipment to meet the dynamic energy needs of park users.

[0003] Currently, energy regulation methods for zero-carbon parks primarily include rule-based fixed scheduling strategies and energy management systems based on optimization algorithms. Rule-based scheduling methods rely on preset energy management rules, such as scheduling charging and discharging equipment during fixed time periods or using simple peak-valley electricity pricing strategies to adjust energy storage charging and discharging behavior. However, these methods struggle to adapt to the complex and ever-changing energy supply and demand conditions of parks and are unable to dynamically optimize the operating status of energy storage equipment. This results in low energy utilization efficiency, high charging and discharging losses, and difficulty in effectively reducing carbon emissions.

[0004] On the other hand, in recent years, some research has proposed energy storage control methods based on optimization algorithms, such as using genetic algorithms and particle swarm optimization algorithms to optimize the scheduling of energy storage equipment. These methods can improve the flexibility of energy scheduling to a certain extent, but the following problems still exist: First, the search capabilities of traditional optimization algorithms cannot guarantee global optimality in complex, multi-constrained campus energy management scenarios and are prone to falling into local optimality; second, existing optimization algorithms lack the ability to dynamically adjust to the uncertainties of energy storage systems, resulting in poor adaptability of optimization strategies; and finally, most optimization methods fail to combine big data analysis to accurately predict campus energy demand, resulting in lagging optimization control strategies and an inability to fully utilize real-time energy data for intelligent decision-making.

[0005] In view of the shortcomings of existing technologies, it is urgent to propose a distributed energy storage optimization method that combines big data analysis, intelligent optimization algorithms and energy control technology to achieve efficient coordination and dynamic scheduling of multiple energy systems in zero-carbon parks. Summary of the Invention

[0006] One purpose of the present invention is to propose a zero-carbon park energy supply optimization method and system based on big data analysis. The present invention can effectively improve energy utilization efficiency, reduce energy storage loss and carbon emissions in a complex dynamic environment, and achieve the goal of intelligent energy supply optimization in a zero-carbon park.

[0007] A method for optimizing energy supply in a zero-carbon park based on big data analysis according to an embodiment of the present invention includes the following steps:

[0008] S1. Collect and preprocess energy operation data to build a unified campus energy operation dataset.

[0009] S2. Utilize big data processing technology to extract features and fuse data from the campus energy operation dataset, construct an energy demand forecasting model based on the historical campus energy operation dataset, and generate future energy supply and demand forecasts.

[0010] S3. Construct a zero-carbon park distributed energy storage control optimization model based on the energy supply and demand forecast results and the real-time park energy operation data set of distributed energy storage equipment;

[0011] S4. Solve the zero-carbon park distributed energy storage control optimization model using an improved hybrid optimization algorithm consisting of a firefly optimization algorithm and a variable-step-size random walk to generate multiple candidate solutions for distributed energy storage control strategies;

[0012] S5. Evaluate the candidate distributed energy storage control strategy solutions, select the optimal distributed energy storage control strategy that best matches the current energy supply and demand status of the park based on a pre-set fitness evaluation index, and use the optimal strategy as the control solution.

[0013] Optionally, the S1 includes the following steps:

[0014] S11. Collect operating status data of distributed energy storage equipment, solar power generation equipment, wind power generation equipment, and traditional energy equipment within the zero-carbon park. The operating status data includes the power generation power, charge and discharge status, battery state of charge, equipment operating temperature, and operating fault status of each device. Set a data collection time window and construct an initial energy equipment operating data set:

[0015] ;

[0016] in, Run datasets for energy devices, Representation device In time The collected operating data, is the total number of energy equipment in the park, is the number of data collections within the time window, For equipment In time The power generation capacity, For equipment In time The charge and discharge state, For devices In time The battery state of charge, For devices In time The operating temperature, For devices In time Operational fault status;

[0017] S12. Collect environmental monitoring data within the zero-carbon park, including park meteorological parameters and environmental change factors, and construct an environmental monitoring data set:

[0018] ;

[0019] in, is an environmental monitoring dataset, Indicates time Collected environmental monitoring data, For time ambient temperature, For time wind speed, For time The solar radiation intensity, For time of air humidity, For time atmospheric pressure;

[0020] S13. Collect energy consumption data from users in the zero-carbon park. The user energy consumption data includes real-time power load, historical power consumption habits, and equipment energy consumption characteristics, and construct a user energy consumption dataset:

[0021] ;

[0022] in, is the user energy consumption dataset, Represents a user In time Collected energy consumption data, is the total number of users in the park, For users In time The real-time power load, For users Historical electricity usage parameters, For users In time Energy consumption characteristics of the equipment.

[0023] S14. Construct a campus energy operation dataset based on the energy equipment operation dataset, environmental monitoring dataset, and user energy consumption dataset:

[0024] ;

[0025] in, This is the park energy operation dataset, which consists of the energy equipment operation dataset , environmental monitoring datasets and user energy consumption datasets Composition, covering multi-dimensional data on park energy supply, environmental changes and user energy demand;

[0026] S15. Preprocess the data set of the park energy operation, including data cleaning, format unification and abnormal data removal. Data cleaning includes missing value filling and outlier detection. Format unification includes data standardization and time alignment. Abnormal data removal includes removing extreme data points caused by collection errors. Finally, a standardized park energy operation data set is obtained. .

[0027] Optionally, the S2 includes the following steps:

[0028] S21. Based on standardized campus energy operation data set Extract the power change characteristics of energy equipment, the charge and discharge cycle characteristics of energy storage equipment, the characteristics of park environment changes, and the characteristics of user energy consumption patterns to construct a park energy characteristic matrix:

[0029] ;

[0030] in, represents the park energy characteristic matrix, Indicates the Features in time The eigenvector of represents the total number of energy features extracted, is the number of acquisitions within the data acquisition time window, Indicates the energy device at time The power change characteristics are used to reflect the power change trend of the equipment operation. Indicates that the energy storage device is at time The charge and discharge cycle characteristics are used to characterize the periodic charge and discharge laws of energy storage devices. Indicates the park's time The environmental change characteristics are used to characterize the impact of the park environment on energy operation. Indicates that the user is at time Energy consumption pattern characteristics are used to reflect the changing rules of user load demand;

[0031] S22. The energy characteristic matrix of the park Perform data fusion, determine the degree of correlation between different energy characteristics based on feature correlation analysis, and construct a feature correlation matrix :

[0032] ;

[0033] in, Representation characteristics and features Between the time window The dynamic weighted correlation coefficient within is a time weighting factor used to enhance the impact of recent data on energy storage device status and energy demand forecasts. 、 Characteristics 、 The dynamic weighted mean within the time window is used to describe the real-time energy change trend of the park;

[0034] S23. Based on dynamic weighted correlation coefficient Filter the key features of the park energy feature matrix, eliminate redundant features, and form a fused park energy feature matrix ;

[0035] S24. Based on the integrated park energy characteristic matrix , a long short-term memory network is used to build a zero-carbon park energy demand prediction model, and the zero-carbon park energy demand prediction model function is defined as:

[0036] ;

[0037] in, Indicates the current time After the prediction interval The energy demand forecast results of the park after represents the activation function of the long short-term memory network. By introducing the attention mechanism, the sensitivity of the zero-carbon park energy demand prediction model to the park energy fluctuation data is enhanced, and the prediction accuracy is improved. is the input weight matrix, which is used to capture the mapping relationship between energy characteristics and campus energy demand. is the recursive weight matrix used to capture the temporal dependency of energy demand changes, is the hidden state vector at the previous moment, is the bias term;

[0038] S25. Using the zero-carbon park energy demand prediction model to predict the future window The energy supply and demand situation in the future is forecasted by multi-step rolling forecast, and the energy supply and demand forecast result sequence is obtained. :

[0039] ;

[0040] in, The energy supply and demand forecast result sequence for the future time. Indicates that at the current moment Based on the prediction Energy demand at each step.

[0041] Optionally, S3 includes the following steps:

[0042] S31. Based on the energy supply and demand forecast results of the park and real-time campus energy operation datasets , calculate the park in time Energy balance difference:

[0043] ;

[0044] in, It represents the actual energy supply and demand situation obtained after data fusion in the real-time park energy operation data set. The energy balance difference is calculated based on the current available charge and discharge capacity of each distributed energy storage device. Allocate to each device in proportion and define distributed energy storage devices In time Charge and discharge control variables :

[0045] ;

[0046] in, is the total number of distributed energy storage devices in the park, For equipment In time Available charge and discharge capacity:

[0047] ;

[0048] in, For equipment The maximum charge and discharge power, and Equipment In time The maximum and minimum energy storage capacity, For equipment In time The current energy storage status;

[0049] By calculating the energy balance difference And according to the available charge and discharge capacity of each device Proportional distribution to achieve the charging and discharging control variables of distributed energy storage equipment The definition of charge and discharge control variables Representation device In time The discharge should be carried out at this power, and the charge and discharge control variables Representation device In time It should be charged at this power;

[0050] S32. Construct a distributed energy storage control optimization model for a zero-carbon park. The optimization goal of this model is to simultaneously improve energy utilization efficiency, reduce charging and discharging losses, and reduce carbon emissions. The objective function of this model is:

[0051] ;

[0052] in, 、 and are the weight coefficients of energy supply and demand balance, charge and discharge loss, and carbon emissions, respectively. Representation device In time the carbon emission costs;

[0053] S33. Use piecewise function to define and quantify the carbon emission cost of energy storage equipment during the charging and discharging process:

[0054] ;

[0055] in, is the discharge efficiency of the energy storage device, Charging efficiency of energy storage devices;

[0056] S34. Construct the constraint conditions of the distributed energy storage control optimization model of the zero-carbon park, wherein the constraint conditions include energy balance constraints, energy storage equipment capacity constraints and charging and discharging power constraints.

[0057] Optionally, the energy balance constraint:

[0058] ;

[0059] in, Representation device In time The energy storage state, is the time step;

[0060] The energy storage device capacity constraints:

[0061] ;

[0062] in, and Equipment Minimum and maximum energy storage capacity;

[0063] The charge and discharge power constraints:

[0064] ;

[0065] in, Represents distributed energy storage equipment In time The current energy storage status, Represents distributed energy storage equipment In time The current energy storage status, represents the time step, represents the discharge efficiency of the energy storage device, represents the charging efficiency of the energy storage device, and Respectively represent devices The minimum and maximum energy storage capacity, Representation device Maximum charge and discharge power.

[0066] Optionally, the S4 includes the following steps:

[0067] S41. Initialize the parameters of the zero-carbon park firefly optimization algorithm and set the firefly population size , maximum number of iterations , initial random perturbation step size , light intensity attraction coefficient and light intensity attenuation factor , the firefly position vector is defined based on the zero-carbon park distributed energy storage control optimization model:

[0068] ;

[0069] in, Indicates the Firefly in The distributed energy storage control strategy corresponding to the iteration is: Representation device At the moment The corresponding charge and discharge control variables, N is the total number of distributed energy storage devices in the park, Optimize the time window length for regulation;

[0070] S42. Counting Fireflies The brightness value of the corresponding distributed energy storage control strategy The brightness value is based on the objective function of the zero-carbon park distributed energy storage control optimization model Characterization:

[0071] ;

[0072] in, Firefly In the number of iterations Brightness value;

[0073] S43. Dynamically adjust the random perturbation step size based on the real-time energy operation data and prediction error of the zero-carbon park, and define the dynamic adjustment strategy of the perturbation step size as follows:

[0074] ;

[0075] in, For fireflies In the The random perturbation step size at the iteration, is the step size adjustment coefficient, For fireflies In the The error between the first iteration of the park energy supply and demand forecast and the real-time data is is the maximum value of the prediction error within the current iteration population;

[0076] S44. Based on the energy regulation needs of zero-carbon parks, an improved position update method is constructed by integrating variable-step random walk and firefly brightness adaptation:

[0077] ;

[0078] in, For fireflies In the The updated position vector after iterations is The brightness value in the current iteration is better than that of fireflies Firefly The position vector of Firefly With fireflies The location distance reflects the degree of difference between distributed energy storage control strategies. is the dynamic perturbation step length, and Respectively The maximum and minimum brightness values of fireflies in the population in the iteration, is a random variable that satisfies the Lévy distribution, 、 Control the scale and stability of the Lévy distribution respectively, is the weight coefficient;

[0079] S45. Through the firefly position adaptive update process of steps S41 to S44, a plurality of candidate solution sets of distributed energy storage control strategies adapted to the energy supply and demand of the park are generated:

[0080] .

[0081] A zero-carbon park energy supply optimization system based on big data analysis is used to implement a zero-carbon park energy supply optimization method based on big data analysis, including the following modules:

[0082] The data acquisition module is used to collect energy operation data, pre-process the collected energy operation data, and build a unified campus energy operation data set;

[0083] A data processing and analysis module, configured to pre-process the data collected by the data collection module and construct a park energy characteristic matrix;

[0084] Energy demand forecasting module, which is used to forecast energy demand based on the park energy characteristic matrix and adopt the long short-term memory network model to generate energy demand forecast results;

[0085] The energy storage control optimization module is used to build a distributed energy storage control optimization model for a zero-carbon campus based on energy demand forecast results and real-time campus energy operation data sets;

[0086] An optimization solution module, configured to solve the energy storage control optimization model based on an improved hybrid optimization algorithm, and generate multiple candidate solutions for distributed energy storage control strategies through iterative optimization;

[0087] The strategy execution and feedback module is used to evaluate multiple candidate solutions for distributed energy storage control strategies and select the optimal control strategy based on preset fitness indicators.

[0088] The beneficial effects of the present invention are:

[0089] (1) The present invention proposes an improved hybrid optimization algorithm that combines the global search capability of the firefly optimization algorithm with the local search advantage of the variable step-size random walk, and improves the convergence speed of the optimization solution by adaptively adjusting the step-size. In the firefly optimization process, a step-size adjustment mechanism based on the energy prediction error of the park is introduced, so that the optimization algorithm has a strong exploration capability in the initial stage, and can more accurately fine-tune the charging and discharging strategy of the energy storage equipment in the later stage of optimization. The Lévy random walk strategy is used to enhance the algorithm's ability to escape from the local optimum, so that the energy storage control scheme can adapt to the complex energy environment of the park, ensuring that efficient and stable optimization effects can be maintained in various dynamic scenarios.

[0090] (2) The present invention introduces a real-time feedback control mechanism in the process of executing the energy storage control strategy. By continuously monitoring the operating status of the park's energy storage equipment, changes in energy supply and demand, and environmental parameters, the optimization model parameters are dynamically corrected and the charging and discharging strategies of the energy storage equipment are adaptively adjusted. At the same time, a multi-objective optimization framework is used to comprehensively weigh the key indicators of energy supply and demand balance, energy storage equipment life, and carbon emissions to ensure that the control strategy can achieve optimal adjustment in different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0092] Figure 1 This is a flow chart of a zero-carbon park energy supply optimization method based on big data analysis proposed by the present invention. DETAILED DESCRIPTION

[0093] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0094] refer to Figure 1 A method for optimizing energy supply in a zero-carbon park based on big data analysis includes the following steps:

[0095] S1. Collect and preprocess energy operation data to build a unified campus energy operation dataset.

[0096] S2. Utilize big data processing technology to extract features and fuse data from the campus energy operation dataset. Build an energy demand forecasting model based on the historical campus energy operation dataset and generate future energy supply and demand forecasts.

[0097] S3. Build a zero-carbon campus distributed energy storage control optimization model based on energy supply and demand forecasts and the real-time campus energy operation dataset of distributed energy storage devices.

[0098] S4. Utilize an improved hybrid optimization algorithm, combining the firefly optimization algorithm with a variable-step-size random walk, to solve the distributed energy storage control optimization model for a zero-carbon park, generating multiple candidate solutions for distributed energy storage control strategies.

[0099] S5. Evaluate candidate distributed energy storage control strategy solutions. Based on pre-defined fitness evaluation indicators, select the optimal distributed energy storage control strategy that best matches the current energy supply and demand status of the park, and use this optimal strategy as the control solution.

[0100] In this embodiment, S1 includes the following steps:

[0101] S11. Collect operating status data of distributed energy storage equipment, solar power generation equipment, wind power generation equipment, and traditional energy equipment within the zero-carbon park. The operating status data includes the power generation power, charge and discharge status, battery state of charge, equipment operating temperature, and operating fault status of each device. Set the data collection time window and construct the initial energy equipment operation data set:

[0102] ;

[0103] in, Run datasets for energy devices, Representation device In time The collected operating data, is the total number of energy equipment in the park, is the number of data collections within the time window, For equipment In time The power generation capacity, For equipment In time The charge and discharge state, For equipment In time The battery state of charge, For equipment In time The operating temperature, For equipment In time Operational fault status;

[0104] S12. Collect environmental monitoring data within the zero-carbon park. The environmental monitoring data includes park meteorological parameters and environmental change factors, and construct an environmental monitoring data set:

[0105] ;

[0106] in, For environmental monitoring datasets, Indicates time Collected environmental monitoring data, For time ambient temperature, For time wind speed, For time The solar radiation intensity, For time of air humidity, For time atmospheric pressure;

[0107] S13. Collect user energy consumption data within the zero-carbon park. User energy consumption data includes real-time power load, historical power usage habits, and equipment energy consumption characteristics. Build a user energy consumption dataset:

[0108] ;

[0109] in, is the user energy consumption dataset, Represents a user In time Collected energy consumption data, is the total number of users in the park, For users In time The real-time power load, For users Historical electricity usage parameters, For users In time Energy consumption characteristics of the equipment.

[0110] S14. Construct the campus energy operation dataset based on the energy equipment operation dataset, environmental monitoring dataset, and user energy consumption dataset:

[0111] ;

[0112] in, This is the park energy operation dataset, which consists of the energy equipment operation dataset , environmental monitoring datasets and user energy consumption datasets Composition, covering multi-dimensional data on park energy supply, environmental changes and user energy demand;

[0113] S15. Preprocess the park energy operation dataset, including data cleaning, format unification, and abnormal data removal. Data cleaning includes missing value filling and outlier detection. Format unification includes data standardization and time alignment. Abnormal data removal includes removing extreme data points caused by collection errors. Finally, a standardized park energy operation dataset is obtained. .

[0114] In this embodiment, S2 includes the following steps:

[0115] S21. Based on standardized campus energy operation data set Extract the power change characteristics of energy equipment, the charge and discharge cycle characteristics of energy storage equipment, the characteristics of park environment changes, and the characteristics of user energy consumption patterns to construct a park energy characteristic matrix:

[0116] ;

[0117] in, represents the park energy characteristic matrix, Indicates the Features in time The eigenvector of represents the total number of energy features extracted, is the number of acquisitions within the data acquisition time window, Indicates the energy device at time The power change characteristics are used to reflect the power change trend of the equipment operation. Indicates that the energy storage device is at time The charge and discharge cycle characteristics are used to characterize the periodic charge and discharge laws of energy storage devices. Indicates the park's time The environmental change characteristics are used to characterize the impact of the park environment on energy operation. Indicates that the user is at time Energy consumption pattern characteristics are used to reflect the changing rules of user load demand.

[0118] S22. Park Energy Characteristics Matrix Perform data fusion, determine the degree of correlation between different energy characteristics based on feature correlation analysis, and construct a feature correlation matrix :

[0119] ;

[0120] in, Representation characteristics and features Between the time window The dynamic weighted correlation coefficient within is a time weighting factor used to enhance the impact of recent data on energy storage device status and energy demand forecasts. 、 Characteristics 、 The dynamic weighted mean within the time window is used to describe the real-time energy change trend of the park;

[0121] S23. Based on dynamic weighted correlation coefficient Filter the key features of the park energy feature matrix, eliminate redundant features, and form a fused park energy feature matrix ;

[0122] S24. Based on the integrated park energy characteristic matrix , a long short-term memory network is used to build a zero-carbon park energy demand prediction model, and the zero-carbon park energy demand prediction model function is defined as:

[0123] ;

[0124] in, Indicates the current time After the prediction interval The energy demand forecast results of the park after represents the activation function of the long short-term memory network. By introducing the attention mechanism, the sensitivity of the zero-carbon park energy demand prediction model to the park energy fluctuation data is enhanced, and the prediction accuracy is improved. is the input weight matrix, which is used to capture the mapping relationship between energy characteristics and campus energy demand. is the recursive weight matrix used to capture the temporal dependency of energy demand changes, is the hidden state vector at the previous moment, is the bias term;

[0125] S25. Using the zero-carbon park energy demand forecasting model to predict future windows The energy supply and demand situation in the future is forecasted by multi-step rolling forecast, and the energy supply and demand forecast result sequence is obtained. :

[0126] ;

[0127] in, The energy supply and demand forecast result sequence for the future time. Indicates that at the current moment Based on the prediction Energy demand at each step.

[0128] In this embodiment, S3 includes the following steps:

[0129] S31. Based on the energy supply and demand forecast results of the park and real-time campus energy operation datasets , calculate the park in time Energy balance difference:

[0130] ;

[0131] in, It represents the actual energy supply and demand situation obtained after data fusion in the real-time park energy operation data set. The energy balance difference is calculated based on the current available charge and discharge capacity of each distributed energy storage device. Allocate to each device in proportion and define distributed energy storage devices In time Charge and discharge control variables :

[0132] ;

[0133] in, is the total number of distributed energy storage devices in the park, For equipment In time Available charge and discharge capacity:

[0134] ;

[0135] in, For equipment The maximum charge and discharge power, and Equipment In time The maximum and minimum energy storage capacity, For equipment In time The current energy storage status;

[0136] By calculating the energy balance difference And according to the available charge and discharge capacity of each device Proportional distribution to achieve the charging and discharging control variables of distributed energy storage equipment The definition of charge and discharge control variables Representation device In time The discharge should be carried out at this power, and the charge and discharge control variables Representation device In time It should be charged at this power;

[0137] S32. Construct a distributed energy storage control optimization model for a zero-carbon park. The optimization goal of this model is to simultaneously improve energy utilization efficiency, reduce charging and discharging losses, and reduce carbon emissions. The objective function of this model is:

[0138] ;

[0139] in, 、 and are the weight coefficients of energy supply and demand balance, charge and discharge loss, and carbon emissions, respectively. Representation device In time the carbon emission costs;

[0140] S33. Use piecewise function to define and quantify the carbon emission cost of energy storage equipment during the charging and discharging process:

[0141] ;

[0142] in, is the discharge efficiency of the energy storage device, Charging efficiency of energy storage devices;

[0143] S34. Construct the constraints of the distributed energy storage control optimization model for zero-carbon parks. The constraints include energy balance constraints, energy storage equipment capacity constraints, and charging and discharging power constraints.

[0144] In this embodiment, the energy balance constraint is:

[0145] ;

[0146] in, Representation device In time The energy storage state, is the time step;

[0147] Energy storage equipment capacity constraints:

[0148] ;

[0149] in, and Equipment Minimum and maximum energy storage capacity;

[0150] Charge and discharge power constraints:

[0151] ;

[0152] in, Represents distributed energy storage equipment In time The current energy storage status, Represents distributed energy storage equipment In time The current energy storage status, represents the time step, represents the discharge efficiency of the energy storage device, represents the charging efficiency of the energy storage device, and Respectively represent devices The minimum and maximum energy storage capacity, Representation device Maximum charge and discharge power.

[0153] In this embodiment, S4 includes the following steps:

[0154] S41. Initialize the parameters of the zero-carbon park firefly optimization algorithm and set the firefly population size , maximum number of iterations , initial random perturbation step size , light intensity attraction coefficient and light intensity attenuation factor , the firefly position vector is defined based on the zero-carbon park distributed energy storage control optimization model:

[0155] ;

[0156] in, Indicates the Firefly in The distributed energy storage control strategy corresponding to the iteration is: Representation device At the moment The corresponding charge and discharge control variables, N is the total number of distributed energy storage devices in the park, Optimize the time window length for regulation;

[0157] S42. Counting Fireflies The brightness value of the corresponding distributed energy storage control strategy The brightness value is based on the objective function of the zero-carbon park distributed energy storage control optimization model Characterization

[0158] ;

[0159] in, Firefly In the number of iterations Brightness value;

[0160] S43. Dynamically adjust the random perturbation step size based on the real-time energy operation data and prediction error of the zero-carbon park, and define the dynamic adjustment strategy of the perturbation step size as follows:

[0161] ;

[0162] in, For fireflies In the The random perturbation step size at the iteration, is the step size adjustment coefficient, For fireflies In the The error between the first iteration of the park energy supply and demand forecast and the real-time data is is the maximum value of the prediction error within the current iteration population;

[0163] S44. Based on the energy regulation needs of zero-carbon parks, an improved position update method is constructed by integrating variable-step random walk and firefly brightness adaptation:

[0164] ;

[0165] in, For fireflies In the The updated position vector after iterations is The brightness value in the current iteration is better than that of fireflies Firefly The position vector of Firefly With fireflies The location distance reflects the degree of difference between distributed energy storage control strategies. is the dynamic perturbation step length, and Respectively The maximum and minimum brightness values of fireflies in the population in the iteration, is a random variable that satisfies the Lévy distribution, 、 Control the scale and stability of the Lévy distribution respectively, is the weight coefficient;

[0166] S45. Through the firefly position adaptive update process of steps S41 to S44, a plurality of candidate solution sets of distributed energy storage control strategies adapted to the energy supply and demand of the park are generated:

[0167] .

[0168] A zero-carbon park energy supply optimization system based on big data analysis is used to implement a zero-carbon park energy supply optimization method based on big data analysis, including the following modules:

[0169] The data acquisition module is used to collect energy operation data, pre-process the collected energy operation data, and build a unified campus energy operation data set;

[0170] The data processing and analysis module is used to pre-process the data collected by the data acquisition module and construct the park energy characteristic matrix;

[0171] Energy demand forecasting module, which is used to forecast energy demand based on the park energy characteristic matrix and adopt the long short-term memory network model to generate energy demand forecast results;

[0172] The energy storage control optimization module is used to build a distributed energy storage control optimization model for a zero-carbon campus based on energy demand forecast results and real-time campus energy operation data sets;

[0173] The optimization solution module is used to solve the energy storage control optimization model based on the improved hybrid optimization algorithm and generate multiple candidate solutions for distributed energy storage control strategies through iterative optimization;

[0174] The strategy execution and feedback module is used to evaluate multiple candidate solutions for distributed energy storage control strategies and select the optimal control strategy based on preset fitness indicators.

[0175] Example 1: This example is applied to a smart manufacturing zero-carbon park in East China, with a total area of approximately 3,000 mu. The park is equipped with 10MW photovoltaic power generation, 8MW wind power generation, and a 20MWh distributed energy storage system. An additional 10MW backup power supply is provided by the State Grid. The main energy-consuming units in the park include three high-energy-consuming manufacturing plants, two smart office buildings, five data centers, and multiple laboratories. The park uses a distributed energy management system for real-time energy regulation. The goal is to optimize energy utilization efficiency, reduce energy waste, and reduce carbon emissions through intelligent energy storage management.

[0176] On July 15, 2024, the park encountered extreme weather. Two consecutive days of rain caused a significant drop in photovoltaic power generation, while the instability of wind power generation further exacerbated the imbalance between energy supply and demand. Furthermore, at 4:30 PM that same day, the load on the smart manufacturing production line in Plant A surged from 3.5MW to 5.2MW due to an increase in temporary orders. This led to a sharp increase in overall power demand in the park, posing a serious challenge to energy regulation. Traditional fixed-time scheduling solutions were unable to quickly respond to this sudden load change, causing the park grid load factor to rise from 80% to 95%, approaching the overload threshold and triggering a warning signal from the park's energy management system.

[0177] At 16:35, the park's distributed energy storage system activated the big data-driven zero-carbon park energy storage control algorithm proposed in this invention, automatically analyzing the current park's energy demand and energy storage status, and using the long short-term memory network (LSTM) model to predict the energy supply and demand trends for the next four hours.

[0178] The LSTM model calculated in real time that the park's overall energy gap was expected to reach 3.8MWh within the next two hours. The energy storage system must fill this gap through a reasonable charging and discharging strategy. The system also analyzed electricity costs during different time periods and found that 4:00 PM to 6:00 PM was the peak period for the power grid, with electricity prices reaching 1.2 yuan / kWh. After 6:00 PM, the price gradually fell to 0.7 yuan / kWh. Therefore, the system decided to prioritize energy storage battery discharge to reduce the cost pressure caused by high electricity prices.

[0179] Phase 1 (16:35-17:00): Emergency Discharge Strategy

[0180] The system calls the Firefly optimization algorithm to calculate the optimal discharge strategy. After 200 iterations, it calculates the optimal discharge scheduling plan:

[0181] Energy storage system 1 (capacity 8MWh, current SOC 75%) discharges 2.5MW;

[0182] Energy storage system 2 (6MWh capacity, current SOC 80%) discharges 1.3MW;

[0183] Energy storage system 3 (6MWh capacity, current SOC 65%) discharges 0.8MW;

[0184] At 16:38, the energy storage system officially started discharging, and the load rate of the park power grid dropped from 95% to 80%, successfully avoiding the risk of production interruption caused by tight energy supply.

[0185] Phase 2 (17:00-18:30): Balanced Regulation Strategy

[0186] At 5:00 PM, the load demand at Plant A began to stabilize, and the overall energy gap in the park narrowed to 2.4 MWh. At this point, the variable-step random walk algorithm adjusted the discharge rate of the energy storage device and gradually reduced the discharge power. The specific adjustments are as follows:

[0187] Energy storage system 1: discharge power reduced to 1.8MW;

[0188] Energy storage system 2: discharge power reduced to 1.1MW;

[0189] Energy storage system 3: discharge power maintained at 0.8MW;

[0190] During this stage, the energy storage system still maintains the discharge mode, but the system dynamically adjusts the discharge rate to stabilize the grid load below 82% and reduce the energy storage loss rate.

[0191] Phase 3 (18:30-19:30): Smart Charging Strategy

[0192] After 6:30 PM, wind speeds increased to 8 m / s, and wind power generation increased from 3.2 MW to 5.6 MW. The park's overall energy supply exceeded demand by 2.5 MW, prompting the energy storage system to switch to charging mode. The LSTM model, combined with real-time energy data, calculated the optimal charging strategy, ensuring that energy storage charging was completed by 7:30 PM and that backup energy storage was in place until photovoltaic power generation resumed the following day.

[0193] The charging strategy finally calculated by the system is:

[0194] Energy storage system 1: charging 2.2MW to 85% SOC;

[0195] Energy storage system 2: charging 1.5MW to 90% SOC;

[0196] Energy storage system 3: charging 0.8MW to 75% SOC;

[0197] At 19:30, the energy storage system completed charging and the entire park returned to normal energy supply.

[0198] In order to evaluate the actual effect of the method of the present invention, this example records the comparative data of the traditional fixed scheduling scheme and the method of the present invention in the same scenario:

[0199]

[0200] Comparing the data reveals that the method significantly shortens energy storage response time, increasing the discharge utilization rate of energy storage equipment to 94.2%. The time required to fill an emergency energy supply gap has been reduced from 30 minutes to 8 minutes, effectively avoiding the risk of production disruptions. Furthermore, the method utilizes LSTM energy forecasting to improve energy supply and demand matching and reduce charging and discharging losses by 40.3%. Furthermore, the carbon emission reduction rate has increased by 117.7% compared to traditional methods.

[0201] This example verifies the feasibility and superiority of the method of the present invention through practical application in a real industrial zero-carbon park. First, by introducing LSTM energy prediction and intelligent optimization scheduling, the present invention can quickly respond to extreme load changes and adjust the energy storage strategy within minutes. Secondly, by combining the firefly optimization algorithm with the variable step-size random walk algorithm, the energy storage control process has more global optimization capabilities, improves the discharge utilization rate and reduces energy storage loss. Finally, experimental results show that the method of the present invention can effectively reduce carbon emissions, enhance the energy autonomous regulation capability of the zero-carbon park, and provide technical support for future smart park energy management.

[0202] The present invention proposes an improved hybrid optimization algorithm, which combines the global search capability of the firefly optimization algorithm with the local search advantage of the variable step-size random walk, and improves the convergence speed of the optimization solution by adaptively adjusting the step size. A step-size adjustment mechanism based on the park energy prediction error is introduced into the firefly optimization process, so that the optimization algorithm has a strong exploration capability in the initial stage, and can more accurately fine-tune the charging and discharging strategy of the energy storage equipment in the later stage of optimization. The Lévy random walk strategy is used to enhance the algorithm's ability to escape from the local optimum, so that the energy storage control scheme can adapt to the complex park energy environment and ensure that efficient and stable optimization effects can be maintained in various dynamic scenarios.

[0203] The present invention introduces a real-time feedback control mechanism during the execution of the energy storage regulation strategy. By continuously monitoring the operating status of the park's energy storage equipment, changes in energy supply and demand, and environmental parameters, it dynamically corrects and optimizes the model parameters and adaptively adjusts the charging and discharging strategies of the energy storage equipment. At the same time, it uses a multi-objective optimization framework to comprehensively weigh key indicators such as energy supply and demand balance, energy storage equipment life, and carbon emissions to ensure that the regulation strategy can achieve optimal adjustment in different scenarios.

[0204] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A zero-carbon park energy supply optimization method based on big data analysis, characterized in that: The steps include: S1. Collect and pre-process energy operation data to build a unified campus energy operation data set; S2. Utilize big data processing technology to extract features and fuse data from the park's energy operation dataset, construct an energy demand forecasting model based on the historical park energy operation dataset, and generate future energy supply and demand forecasts; S3. Construct a zero-carbon park distributed energy storage control optimization model based on the energy supply and demand forecast results and the real-time park energy operation data set of the distributed energy storage equipment; The S3 includes the following steps: S31. Based on the energy supply and demand forecast results of the park and real-time campus energy operation dataset D final , calculate the energy balance difference of the park at time t: in, It represents the actual energy supply and demand situation obtained after data fusion in the constructed real-time park energy operation data set. The energy balance difference Δ is calculated based on the current available charge and discharge capacity of each distributed energy storage device. t Allocate to each device in proportion and define the charge and discharge control variable x of distributed energy storage device i at time t i,t : Where N is the total number of distributed energy storage devices in the park, is the available charge and discharge capacity of device i at time t: in, is the maximum charge and discharge power of device i, and are the maximum and minimum energy storage capacities of device i at time t, E i,t is the current energy storage state of device i at time t; By calculating the energy balance difference Δ t And according to the available charge and discharge capacity of each device The proportional distribution realizes the charging and discharging control variable x of the distributed energy storage device i,t Definition of charge and discharge control variable x i,t >0 means that device i should discharge at this power at time t, and the charge and discharge control variable x i,t <0 means that device i should be charged at this power at time t; S32. Construct a distributed energy storage control optimization model for a zero-carbon park. The optimization goal of this model is to simultaneously improve energy utilization efficiency, reduce charging and discharging losses, and reduce carbon emissions. The objective function of this model is: Among them, α, β and γ are the weight coefficients of energy supply and demand balance, charge and discharge loss and carbon emission respectively, ξ i,t represents the carbon emission cost of equipment i at time t; S33. Use piecewise function to define and quantify the carbon emission cost of energy storage equipment during the charging and discharging process: Among them, η dis is the discharge efficiency of the energy storage device, η ch Charging efficiency of energy storage devices; S34. Constructing constraints for a distributed energy storage control optimization model for a zero-carbon park, the constraints including energy balance constraints, energy storage device capacity constraints, and charge and discharge power constraints; S4. Solve the zero-carbon park distributed energy storage control optimization model using an improved hybrid optimization algorithm consisting of a combination of a firefly optimization algorithm and a variable step-size random walk to generate multiple candidate solutions for distributed energy storage control strategies; S5. Evaluate the candidate solutions of the distributed energy storage control strategy, select the optimal distributed energy storage control strategy that best matches the current energy supply and demand status of the park based on a pre-set fitness evaluation index, and use the optimal distributed energy storage control strategy as the control solution.

2. The method for optimizing energy supply in a zero-carbon park based on big data analysis according to claim 1 is characterized in that: Said S1 comprises the following steps: S11. Collect operating status data of distributed energy storage equipment, solar power generation equipment, wind power generation equipment, and traditional energy equipment within the zero-carbon park. The operating status data includes the power generation power, charge and discharge status, battery state of charge, equipment operating temperature, and operating fault status of each device. Set a data collection time window and construct an initial energy equipment operating data set: Among them, D device Run the dataset for energy devices, d i,t represents the operating data collected by device i at time t, N is the total number of energy devices in the park, T is the number of data collection times in the time window, and P i,t is the power generated by device i at time t, C i,t is the charge and discharge state of device i at time t, S i,t is the battery state of charge of device i at time t, T i,t is the operating temperature of device i at time t, F i,t is the operating fault state of device i at time t; S12. Collect environmental monitoring data within the zero-carbon park, including park meteorological parameters and environmental change factors, and construct an environmental monitoring data set: Among them, D env is the environmental monitoring dataset, e t Represents the environmental monitoring data collected at time t, Temp t is the ambient temperature at time t, Wind t is the wind speed at time t, Solar t is the solar radiation intensity at time t, Hum t is the air humidity at time t, Pressure t is the atmospheric pressure at time t; S13. Collect energy consumption data from users in the zero-carbon park. The user energy consumption data includes real-time power load, historical power consumption habits, and equipment energy consumption characteristics, and construct a user energy consumption dataset: Among them, D user is the user energy consumption dataset, u j,t represents the energy consumption data collected by user j at time t, M is the total number of users in the park, L j,t is the real-time electricity load of user j at time t, H j,t is the historical electricity usage habit parameter of user j, E j,t is the device energy consumption characteristics of user j at time t; S14. Construct a campus energy operation dataset based on the energy equipment operation dataset, environmental monitoring dataset, and user energy consumption dataset: D park =D device ∪D env ∪D user ; Among them, D park The energy operation dataset of the park is composed of the energy equipment operation dataset D device , Environmental Monitoring Dataset D env and user energy consumption dataset D user Composition, covering multi-dimensional data on park energy supply, environmental changes and user energy demand; S15. Preprocess the data set of the park energy operation, including data cleaning, format unification, and abnormal data removal. Data cleaning includes missing value filling and outlier detection. Format unification includes data standardization and time alignment. Abnormal data removal includes removing extreme data points caused by collection errors. Finally, a standardized park energy operation data set D is obtained. final .

3. The method for optimizing energy supply in a zero-carbon park based on big data analysis according to claim 2 is characterized in that: The S2 comprises the following steps: S21. Based on the standardized campus energy operation data set D final Extract the power change characteristics of energy equipment, the charge and discharge cycle characteristics of energy storage equipment, the characteristics of park environment changes, and the characteristics of user energy consumption patterns to construct a park energy characteristic matrix: Among them, F park represents the park energy characteristic matrix, f k,t represents the feature vector of the kth feature at time t, K represents the total number of energy features extracted, T is the number of acquisitions within the data acquisition time window, Indicates the power change characteristics of energy equipment at time t, which is used to reflect the power change trend of equipment operation. Indicates the charge and discharge cycle characteristics of the energy storage device at time t, which is used to characterize the periodic charge and discharge law of the energy storage device. It represents the environmental change characteristics of the park at time t, and is used to describe the impact of the park environment on energy operation. Indicates the energy consumption pattern characteristics of the user at time t, which is used to reflect the changing law of user load demand. S22. The park energy characteristic matrix F park Perform data fusion, determine the degree of correlation between different energy characteristics based on feature correlation analysis, and construct the feature correlation matrix R feature : Among them, r m,n Represents the dynamic weighted correlation coefficient between feature m and feature n in the time window T, ω t is a time weighting factor used to enhance the impact of recent data on energy storage device status and energy demand forecasts. are the dynamic weighted means within the time window of features m and n, respectively, used to characterize the real-time energy change trend of the park; S23. Based on the dynamic weighted correlation coefficient r m,n Filter the key features of the park energy feature matrix, remove redundant features, and form the integrated park energy feature matrix F fusion ; S24. Based on the integrated park energy characteristic matrix F fusion , a long short-term memory network is used to build a zero-carbon park energy demand prediction model, and the zero-carbon park energy demand prediction model function is defined as: in, represents the energy demand forecast result of the park after the prediction interval Δt at the current time t, f LSTM (·) represents the activation function of the long short-term memory network. By introducing the attention mechanism, the sensitivity of the zero-carbon park energy demand prediction model to the park energy fluctuation data is enhanced, and the prediction accuracy is improved. f is the input weight matrix, which is used to capture the mapping relationship between energy characteristics and campus energy demand. r is the recursive weight matrix used to capture the temporal dependency of energy demand changes, h t-1 is the hidden state vector of the previous moment, and b is the bias term; S25. Use the zero-carbon park energy demand forecasting model to perform a multi-step rolling forecast of the energy supply and demand situation within the future forecast window ΔT to obtain a sequence of energy supply and demand forecast results at future moments. in, The energy supply and demand forecast result sequence for the future time. It indicates the energy demand situation at the i-th step based on the current time t.

4. The method for optimizing energy supply in a zero-carbon park based on big data analysis according to claim 1 is characterized in that: The energy balance constraint: Among them, E i,t represents the energy storage state of device i at time t, and Δt is the time step; The energy storage device capacity constraints: in, and are the minimum and maximum energy storage capacity of device i, respectively; The charge and discharge power constraints: Among them, E i,t represents the current energy storage state of distributed energy storage device i at time t, E i,t+1 represents the current energy storage state of distributed energy storage device i at time t+1, Δt represents the time step, η dis Represents the discharge efficiency of the energy storage device, η ch represents the charging efficiency of the energy storage device, and They represent the minimum and maximum energy storage capacity of device i, Indicates the maximum charge and discharge power of device i.

5. The method for optimizing energy supply in a zero-carbon park based on big data analysis according to claim 4 is characterized in that: The S4 comprises the following steps: S41. Initialize the parameters of the zero-carbon park firefly optimization algorithm, set the firefly population size M, the maximum number of iterations G max , initial random perturbation step α max , light intensity attraction coefficient β0 and light intensity attenuation factor γ, and define the firefly position vector according to the zero-carbon park distributed energy storage control optimization model: Among them, X m,g represents the distributed energy storage control strategy corresponding to the mth firefly in the gth iteration, represents the charge and discharge control variable corresponding to device i at time t, N is the total number of distributed energy storage devices in the park, and T is the length of the control optimization time window; S42. Calculate the brightness value I of the distributed energy storage control strategy corresponding to firefly m m,g The brightness value is based on the objective function J(X m,g ) Characterization: Among them, I m,g represents the brightness value of firefly m at iteration number g; S43. Dynamically adjust the random perturbation step size based on the real-time energy operation data and prediction error of the zero-carbon park, and define the dynamic adjustment strategy of the perturbation step size as follows: Among them, α m,g is the random perturbation step size of firefly m at the gth iteration, λ1 is the step size adjustment coefficient, Δ m,g is the prediction error between the energy supply and demand forecast of the park by Firefly m at the gth iteration and the real-time data, Δ max is the maximum value of the prediction error within the current iteration population; S44. Based on the energy regulation needs of zero-carbon parks, an improved position update method is constructed by integrating variable-step random walk and firefly brightness adaptation: Among them, X m,g+1 is the position vector of firefly m after the g+1th iteration update, X k,g is the position vector of firefly k whose brightness value is better than that of firefly m in the current iteration, r m,k represents the distance between firefly m and firefly k, reflecting the degree of difference between distributed energy storage control strategies, α m,g is the dynamic perturbation step length, I max,g with I min,g are the maximum and minimum brightness values of fireflies in the population in the g-th iteration, L μ,ν For random variables that satisfy the Lévy distribution, μ and ν control the scale and stability of the Lévy distribution respectively, and γ1 is the weight coefficient; S45. Through the firefly position adaptive update process of steps S41 to S44, a plurality of candidate solution sets of distributed energy storage control strategies adapted to the energy supply and demand of the park are generated: X candidate ={X m,Gmax ∣m=1,2,…,M}。 6. A zero-carbon park energy supply optimization system based on big data analysis, used to execute a zero-carbon park energy supply optimization method based on big data analysis according to any one of claims 1 to 5, characterized in that: Includes the following modules: The data acquisition module is used to collect energy operation data, pre-process the collected energy operation data, and build a unified campus energy operation data set; A data processing and analysis module, configured to pre-process the data collected by the data collection module and construct a park energy characteristic matrix; Energy demand forecasting module, which is used to forecast energy demand based on the park energy characteristic matrix and adopt the long short-term memory network model to generate energy demand forecast results; The energy storage control optimization module is used to build a distributed energy storage control optimization model for a zero-carbon campus based on energy demand forecast results and real-time campus energy operation data sets; An optimization solution module, configured to solve the energy storage control optimization model based on an improved hybrid optimization algorithm, and generate multiple candidate solutions for distributed energy storage control strategies through iterative optimization; The strategy execution and feedback module is used to evaluate multiple candidate solutions for distributed energy storage control strategies and select the optimal control strategy based on preset fitness indicators.

Images