A regulation method and system based on distributed photovoltaic energy storage

Abstract

The present invention discloses a control method based on distributed photovoltaic energy storage, which relates to the technical field of distributed photovoltaic power generation, including: comprehensively monitoring the photovoltaic energy storage system in real time through an intelligent monitoring network to ensure data continuity and accuracy, improve operation and maintenance efficiency and prediction accuracy, use advanced data synchronization strategies to optimize resource utilization, enhance system stability, combine LSTM and ARIMA models to predict photovoltaic power and user demand, use K-means algorithm to refine user classification, greatly improve prediction accuracy, hybrid genetic and particle swarm optimization algorithms to balance costs, user satisfaction and supply and demand, achieve dual improvements in economic benefits and user experience through intelligent energy distribution, and promote the optimal allocation of resources and a leap in the overall efficiency of the system.

Description

Technical Field

[0001] The present invention relates to the technical field of distributed photovoltaic power generation, and in particular to a control method and system based on distributed photovoltaic energy storage. Background Art

[0002] In recent years, with the urgent need for global energy structure transformation and sustainable development strategy, solar energy, as a clean and renewable energy form, has become a research hotspot in terms of improving its utilization efficiency and optimizing its energy storage system. However, how to efficiently integrate distributed photovoltaic systems and energy storage resources to achieve intelligent regulation and maximize energy utilization efficiency while meeting the dynamic needs of users is still a challenge facing the industry. Due to the lack of in-depth mining and comprehensive analysis of multi-dimensional data, as well as accurate prediction of user behavior patterns, the photovoltaic output does not match the actual needs of users. The energy storage strategy is relatively static and fails to make full use of the peak-valley electricity price difference, resulting in energy waste. There is a lag in personalized energy-saving suggestions and real-time response adjustments, which cannot effectively respond to the ever-changing needs of users and the dynamic characteristics of the power grid, reducing the economy of the overall system and user satisfaction. Summary of the invention

[0003] In view of the above existing problems, the present invention is proposed.

[0004] Therefore, the present invention provides a control method and system based on distributed photovoltaic energy storage to solve the problems of low energy efficiency caused by poor matching between energy management and user needs, static energy storage strategy and delayed response in distributed photovoltaic energy storage systems.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] In a first aspect, an embodiment of the present invention provides a control method based on distributed photovoltaic energy storage, which includes:

[0007] Deploy intelligent monitoring and data collection networks, and implement full-link multi-dimensional dynamic data synchronization strategies for data management and analysis;

[0008] Construct a photovoltaic prediction and user demand dynamic analysis model, and output photovoltaic power prediction curve and user electricity consumption behavior prediction based on the analyzed data;

[0009] The output results of the photovoltaic prediction model are integrated with the prediction of user electricity consumption behavior, and a multi-objective optimization model is established to calculate the optimal load distribution under the dynamic electricity price mechanism;

[0010] Recommend personalized energy-saving measures based on load forecast results;

[0011] Dynamically adjust the energy storage charging and discharging strategy based on real-time load changes and user response.

[0012] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, wherein: deploying an intelligent monitoring and data acquisition network includes installing smart meters, environmental sensors, energy storage system communication modules and photovoltaic power station power grid monitoring systems;

[0013] Smart meters regularly upload current, voltage, instantaneous power and accumulated energy data;

[0014] Environmental sensors acquire light intensity, temperature, humidity, and wind speed data in real time;

[0015] The energy storage system communication module monitors the battery status, charging rate, discharging rate, number of cycles, health status and fault alarm information in real time;

[0016] The photovoltaic power plant grid monitoring system regularly synchronizes frequency, voltage level and real-time load demand.

[0017] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, wherein: implementing a full-link multi-dimensional dynamic data synchronization strategy for data management and analysis includes using Apache Kafka message queue technology to build a high-throughput, low-latency data transmission channel and design a hierarchical storage architecture, hot data is stored in a high-speed memory database for instant query, cold data is regularly migrated to a distributed file system to reduce costs, the synchronization frequency is dynamically adjusted according to the importance and change frequency of the data, the Z-score method is used to identify outliers and forward and backward filling and linear interpolation methods are used to process missing data, and random noise is removed based on statistical methods and numerical data is normalized.

[0018] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, wherein: a photovoltaic prediction and user demand dynamic analysis model is constructed, and a photovoltaic power prediction curve and a user electricity consumption behavior prediction are output based on the analyzed data, including using an LSTM model to receive the standardized historical light intensity sequence and weather forecast data at the input layer, and a hidden layer using multi-layer STM units to capture the time series pattern, and an output layer predicting the future photovoltaic power curve, applying a K-means clustering algorithm, grouping based on the user's historical electricity consumption pattern, applying an ARIMA model in each user group to capture the electricity consumption pattern of users in the group, training an ARIMA model for each group, inputting historical electricity consumption data, and outputting a predicted future electricity demand curve.

[0019] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, wherein: the output results of the photovoltaic prediction model and the user's electricity consumption behavior prediction are integrated to establish a multi-objective optimization model, and the optimal load distribution under the dynamic electricity price mechanism is calculated, including integrating the photovoltaic prediction model output with the user's electricity consumption behavior prediction data and the current system state as the input data set of the optimization model, setting a hybrid strategy of the genetic algorithm and the particle swarm optimization algorithm, and combining the advantages of both to perform multiple rounds of iterations. In each iteration, the genetic algorithm and the particle swarm optimization algorithm respectively generate new solution sets, evaluate the quality of the solution through the fitness function, select high-quality solutions for crossover and mutation operations, and obtain the Pareto solution set after convergence. These solutions achieve the best compromise between cost, user satisfaction and supply and demand balance. From the Pareto solution set, a solution that meets the specific target priority is selected, and the dynamic electricity price mechanism is output to guide the energy storage charging and discharging strategy and photovoltaic output adjustment to ensure the optimization of load distribution.

[0020] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, in which: combined with the load forecast results, personalized energy-saving measures are recommended, including the load forecast curve, cost and user satisfaction index obtained by analyzing the solution of the multi-objective optimization model, dividing the time period into peak, flat peak and valley according to the load forecast, setting the corresponding electricity price strategy, setting the peak period to 150% of the benchmark electricity price, the flat peak period to 100% of the benchmark electricity price, and the valley period to 50% of the benchmark electricity price, encouraging users to shift peaks and fill valleys, integrating the dynamic electricity price calculation logic into the existing billing system, automatically adjusting the electricity price according to time, designing and developing a user interactive APP and SMS push platform, integrating a real-time electricity price information interface, automatically pushing real-time electricity price information according to the electricity price period in the user's area, developing a personalized energy-saving suggestion algorithm, generating energy-saving measures based on the user's historical electricity usage habits and equipment information, and regularly analyzing user feedback and electricity usage data, evaluating energy-saving effects and user satisfaction, optimizing the dynamic electricity price strategy according to the feedback and analysis results, adjusting the energy-saving suggestion algorithm, and improving user participation and satisfaction.

[0021] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, wherein: according to real-time load changes and user responses, dynamically adjusting the energy storage charging and discharging strategy includes real-time collection of grid load data, actual user power consumption, energy storage system status, and real-time power of photovoltaic power stations, uploading to the central monitoring system every 5 minutes through the Internet of Things platform, real-time comparison of the predicted load curve and the actual load data, calculating the deviation value, if the deviation exceeds a predetermined threshold, triggering an adaptive adjustment mechanism, designing a fuzzy logic controller for the adaptive adjustment strategy, defining a maximum load deviation of 5%, inputting load deviation, energy storage SOC, and photovoltaic power change rate, outputting energy storage charging and discharging instructions, photovoltaic power station adjustment instructions, and constructing a fuzzy rule base, if the load deviation is large and the SOC is high, then increasing the discharge amount, if the photovoltaic power change rate is lower than expected and the energy storage SOC is low, then reducing the energy storage discharge amount, and retaining the power for use during the evening peak.

[0022] As a preferred solution of the control method based on distributed photovoltaic energy storage described in the present invention, the SOC safety threshold is set to 20% to 80%. If it is close to the upper limit of 80%, charging is stopped and discharge is prepared. If it is lower than 20%, only charging is allowed except in emergency situations. Once the threshold is reached, the automatic locking operation is started, and it can only be unlocked if manual intervention is obtained with advanced authority and the system automatically determines that an emergency situation has occurred.

[0023] The beneficial effects of the present invention are: comprehensive and real-time monitoring of the photovoltaic energy storage system through an intelligent monitoring network to ensure data continuity and accuracy, improve operation and maintenance efficiency and prediction accuracy, use advanced data synchronization strategies to optimize resource utilization, enhance system stability, combine LSTM and ARIMA models to predict photovoltaic power and user demand, and use K-means algorithm to refine user classification, greatly improving prediction accuracy. Hybrid genetic and particle swarm optimization algorithms balance costs, user satisfaction and supply and demand, and achieve dual improvements in economic benefits and user experience through intelligent energy distribution, which promotes the optimal allocation of resources and a leap in the overall efficiency of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0025] Figure 1 This is a flow chart of the control method based on distributed photovoltaic energy storage in Example 1.

[0026] Figure 2 This is a diagram for determining the risk level of the user behavior pattern in Example 1. DETAILED DESCRIPTION

[0027] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0028] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0030] Example 1

[0031] Reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a control method based on distributed photovoltaic energy storage, comprising the following steps:

[0032] S1. Deploy an intelligent monitoring and data collection network, implement a full-link multi-dimensional dynamic data synchronization strategy for data management and analysis;

[0033] Construct a photovoltaic prediction and user demand dynamic analysis model, and output photovoltaic power prediction curve and user electricity consumption behavior prediction based on the analyzed data;

[0034] The output results of the photovoltaic prediction model are integrated with the prediction of user electricity consumption behavior, and a multi-objective optimization model is established to calculate the optimal load distribution under the dynamic electricity price mechanism;

[0035] Recommend personalized energy-saving measures based on load forecast results;

[0036] Dynamically adjust the energy storage charging and discharging strategy based on real-time load changes and user response.

[0037] S2, deploying intelligent monitoring and data collection network including installing smart meters, environmental sensors, energy storage system communication modules and photovoltaic power station grid monitoring system;

[0038] Smart meters regularly upload current, voltage, instantaneous power and accumulated energy data;

[0039] Environmental sensors acquire light intensity, temperature, humidity, and wind speed data in real time;

[0040] The energy storage system communication module monitors the battery status, charging rate, discharging rate, number of cycles, health status and fault alarm information in real time;

[0041] The photovoltaic power plant grid monitoring system regularly synchronizes frequency, voltage level and real-time load demand.

[0042] S3. Implement a full-link, multi-dimensional dynamic data synchronization strategy for data management and analysis, including using Apache Kafka message queue technology to build a high-throughput, low-latency data transmission channel and design a layered storage architecture. Hot data is stored in a high-speed memory database for instant query, and cold data is regularly migrated to a distributed file system to reduce costs. The synchronization frequency is dynamically adjusted according to the importance and change frequency of the data. The Z-score method is used to identify outliers and forward and backward filling and linear interpolation methods are used to process missing data. Random noise is removed based on statistical methods and numerical data is normalized.

[0043] Furthermore, the Z-score method is applied to identify outliers. Data points beyond the 3σ range are marked as outliers and processed. The expression is:

[0044]

[0045] Among them, X is the observation value, the specific value to check whether it is an outlier, μ is the arithmetic mean of all observations, representing the center position or "typical" value of the data set, σ is the standard deviation, which is an indicator of the degree of dispersion of data points relative to the mean. The larger the standard deviation, the greater the difference between the data, and the smaller the standard deviation, the more concentrated the data points. Z is the score after the observation value is standardized, indicating how many standard deviations the original value X is from the mean value μ. If the value of Z is large, it means that the observation is extreme or unusual relative to other observations.

[0046] Furthermore, we perform Min-Max normalization on all types of data and map all values ​​to the interval [0, 1] to facilitate model training and comparison. The expression is:

[0047]

[0048] Among them, X norm is a normalized value, which will be mapped to the interval [0, 1]. min is the minimum value of the feature in the original data set, X max is the maximum value of the feature in the original data set.

[0049] S4. Construct a photovoltaic prediction and user demand dynamic analysis model, and output the photovoltaic power prediction curve and user electricity consumption behavior prediction based on the analyzed data, including using the LSTM model to receive the standardized historical light intensity series and weather forecast data at the input layer, and using multi-layer STM units in the hidden layer to capture the time series pattern. The output layer predicts the future photovoltaic power curve, applies the K-means clustering algorithm, groups users based on their historical electricity consumption patterns, applies the ARIMA model in each user group to capture the electricity consumption patterns of users in the group, trains the ARIMA model for each group, inputs historical electricity consumption data, and outputs the predicted future electricity load demand curve.

[0050] Furthermore, a photovoltaic prediction model is constructed to output the photovoltaic power prediction curve based on the analyzed data. The expression is:

[0051] P PV (t+1)=LSTM([L(t), W(t), T(t), H(t)]; θ LSTM )

[0052] Among them, P PV (t+1) is the predicted photovoltaic power value at time point t+1, L(t) is the observed value of light intensity, W(t) is the observed value of wind speed, T(t) is the observed value of temperature, H(t) is the observed value of humidity, θ LSTM It is a set of parameters of the LSTM model, including weights and bias terms.

[0053] Furthermore, the historical electricity consumption data U of user i at time t is i (t) Apply K-means algorithm for clustering and get n groups, expressed as:

[0054] C=KMeans(U1(t),U2(t),…,U m (t); n)

[0055] Where C is the clustering result output by the K-means algorithm, which contains the distribution of n clusters, m is the number of users, U1(t), U2(t), ..., U m (t) is the electricity consumption data point of m users at time t, indicating that cluster analysis is being performed on all electricity consumption data of all users at the same time point, and n is the number of clusters to be formed.

[0056] Furthermore, for each cluster group j, the ARIMA model is trained based on its historical electricity consumption data to predict the total electricity load demand in the next period. The expression is:

[0057]

[0058] in is the total predicted electricity load demand of cluster j in the next period, U j (t) is the observed power consumption of cluster j at time t, is the set of ARIMA model parameters specific to cluster j.

[0059] S5. Integrate the output results of the photovoltaic prediction model with the user's electricity consumption behavior prediction, establish a multi-objective optimization model, and calculate the optimal load distribution under the dynamic electricity price mechanism. This includes integrating the photovoltaic prediction model output with the user's electricity consumption behavior prediction data and the current system status as the input data set of the optimization model, setting a hybrid strategy of the genetic algorithm and the particle swarm optimization algorithm, combining the advantages of both, and performing multiple rounds of iterations. In each iteration, the genetic algorithm and the particle swarm optimization algorithm respectively generate new solution sets, evaluate the quality of the solution through the fitness function, select high-quality solutions for crossover and mutation operations, and obtain the Pareto solution set after convergence. These solutions achieve the best compromise between cost, user satisfaction and supply and demand balance. From the Pareto solution set, select a solution that meets the specific target priority, output the dynamic electricity price mechanism, guide the energy storage charging and discharging strategy and photovoltaic output adjustment, and ensure the optimization of load distribution.

[0060] Furthermore, the optimization objectives are clarified and a multi-objective optimization model is established, as follows:

[0061] Define the objective function:

[0062] Cost function: f1(x) = C g +C s

[0063] Among them, C g It is the cost of purchasing electricity from the power grid. In order to meet user needs, in addition to the electricity generated by itself, it is also necessary to purchase electricity from the external power grid. This cost varies with the peak and valley electricity prices and the purchase volume. s The energy storage operation cost refers to the cost of using energy storage equipment to store and release electrical energy in the system;

[0064] User satisfaction function:

[0065] in, is the actual satisfaction of the ith user, The expected satisfaction of the i-th user, N is the total number of users;

[0066] Supply and demand balance indicator: f3(x)=|∑(Load pred (t)-P PV (t)-P s (t))|

[0067] Among them, Load pred(t) is the total power demand of the user's predicted load at time t, P PV (t) is the predicted PV power at time t, P s (t) is the electric energy provided by the energy storage system to the grid and users at time t;

[0068] The comprehensive cost-effectiveness index is constructed through the objective function, and the expression is:

[0069] I(x,t)=w1·f1(x)+w2·f2(x)+w3·f3(x)

[0070] Among them, w1 is the weight coefficient of the cost function, w2 is the weight coefficient of the user satisfaction function, and w3 is the weight coefficient of the supply and demand balance indicator;

[0071] Combining the basic electricity price, current load forecast and time period characteristics, the dynamic electricity price for each time period is calculated, and the expression is:

[0072] P price (t) = γ base ×(1+δ(t)×f load (t)+h(I(x,t)))

[0073] Among them, P price (t) is the dynamic electricity price at time t, γ base It is the basic electricity price without time, load and market adjustment factors. δ(t) is the time adjustment coefficient, which is used to reflect the difference in electricity cost in different time periods of the day. During peak hours, δ(t)>0, and during off-peak hours, δ(t)<0. load (t) is the load peak-valley coefficient, which reflects the current position relative to the load level throughout the day. load (t)>1, low period f load (t)<1, the average period is close to 1, and h(I(x, t))) is a monotonic function used to reflect the reasonable impact of the comprehensive cost-effectiveness index I on electricity prices.

[0074] S6. Combined with the load forecast results, personalized energy-saving measures are recommended, including the load forecast curve, cost and user satisfaction indicators obtained by analyzing the solution of the multi-objective optimization model. According to the load forecast, the time periods are divided into peak, off-peak and off-peak periods, and the corresponding electricity price strategy is set. The peak period is set at 150% of the benchmark electricity price, the off-peak period is 100% of the benchmark electricity price, and the off-peak period is 50% of the benchmark electricity price. Users are encouraged to shift peak loads and fill valley loads. The dynamic electricity price calculation logic is integrated into the existing billing system, and the electricity price is automatically adjusted according to time. A user interactive APP and SMS push platform are designed and developed, and a real-time electricity price information interface is integrated. According to the electricity price period in the user's area, real-time electricity price information is automatically pushed. A personalized energy-saving suggestion algorithm is developed. Based on the user's historical electricity usage habits and equipment information, energy-saving measures are generated and user feedback and electricity usage data are regularly analyzed to evaluate the energy-saving effect and user satisfaction. According to the feedback and analysis results, the dynamic electricity price strategy is optimized, and the energy-saving suggestion algorithm is adjusted to improve user participation and satisfaction.

[0075] Furthermore, combining the load forecast results and the real-time electricity price in each period, the energy-saving recommendation algorithm designed for user a is expressed as:

[0076]

[0077] Among them, Ss rec (a) is the energy saving amount recommended for user a, based on the user's historical data and behavior pattern, D u (a, t) predicts the power consumption of user a at time t, E eff (a) is the energy efficiency coefficient of user a, which reflects the economic efficiency of energy use.

[0078] S7. According to the real-time load changes and user response, the energy storage charging and discharging strategy is dynamically adjusted, including real-time collection of grid load data, actual user power consumption, energy storage system status, and real-time power of photovoltaic power stations. The data are uploaded to the central monitoring system every 5 minutes through the Internet of Things platform. The predicted load curve is compared with the actual load data in real time, and the deviation value is calculated. If the deviation exceeds the predetermined threshold, the adaptive adjustment mechanism is triggered. A fuzzy logic controller is designed for the adaptive adjustment strategy. The maximum load deviation is defined as 5%. The load deviation, energy storage SOC, and photovoltaic power change rate are input. The energy storage charging and discharging instructions and photovoltaic power station adjustment instructions are output and a fuzzy rule base is constructed. If the load deviation is large and the SOC is high, the discharge amount is increased. If the photovoltaic power change rate is lower than expected and the energy storage SOC is low, the energy storage discharge amount is reduced to reserve power for use during the evening peak.

[0079] Furthermore, based on the historical load data, the load demand in the future period is predicted, and the deviation between the predicted load and the actual load is calculated. The expression is:

[0080] ΔLoad=|Loadpred (t)-Load act (t)|

[0081] Among them, Load act (t) is the actual total load demand of users at time t, Load pred (t) is the total power demand of the user's predicted load at time t, ΔLoad is the deviation between the predicted load and the actual load;

[0082] The maximum load deviation is defined as 5%. When the calculated deviation value ΔLoad exceeds the preset maximum load deviation by 5%, the dynamic adjustment mechanism of the energy storage system is triggered.

[0083] S8. Set the SOC safety threshold to 20% to 80%. If it is close to the upper limit of 80%, stop charging and prepare for discharging. If it is below 20%, only charging is allowed except in emergencies. Once the threshold is reached, the automatic locking operation is initiated. The vehicle can only be unlocked if manual intervention is obtained with advanced authority or the system automatically determines an emergency situation.

[0084] This embodiment also provides a computer device, which is suitable for the case of a control method based on distributed photovoltaic energy storage, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the control method based on distributed photovoltaic energy storage proposed in the above embodiment.

[0085] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0086] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the control method based on distributed photovoltaic energy storage proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0087] In summary, the present invention comprehensively monitors the photovoltaic energy storage system in real time through an intelligent monitoring network to ensure data continuity and accuracy, improve operation and maintenance efficiency and prediction accuracy, use advanced data synchronization strategies to optimize resource utilization, enhance system stability, and combine LSTM and ARIMA models to predict photovoltaic power and user demand. The K-means algorithm refines user classification and greatly improves prediction accuracy. The hybrid genetic and particle swarm optimization algorithms balance costs, user satisfaction, and supply and demand. Through intelligent energy distribution, both economic benefits and user experience are improved, which promotes the optimal allocation of resources and a leap in the overall efficiency of the system.

[0088] Example 2

[0089] Referring to Table 1, which is the second embodiment of the present invention, in order to further verify the advancement of the present invention, experimental simulation data of a control method based on distributed photovoltaic energy storage is provided.

[0090] We conducted a one-month experiment covering a variety of climate conditions and user behavior patterns. In the experimental design, we deployed a complete intelligent monitoring and data acquisition network, which includes 100 smart meters, 50 environmental sensors, 20 energy storage system communication modules, and a photovoltaic power station grid monitoring system. All equipment is configured as required and connected to a unified monitoring and data processing platform. Smart meters upload current, voltage, instantaneous power and accumulated electric energy data every 15 minutes, environmental sensors continuously monitor light intensity, temperature, humidity and wind speed, and transmit data in real time. The energy storage system communication module tracks the battery status in real time to ensure the accuracy of the data. The photovoltaic power station grid monitoring system is responsible for synchronizing the grid frequency, voltage level and real-time load demand, and is updated every hour.

[0091] The data synchronization strategy uses Apache Kafka to achieve efficient data flow. Hot data is stored in the in-memory database to ensure instant query needs, while cold data is transferred to the distributed file system on a weekly basis to effectively control costs. The data processing process includes outlier identification and processing, random noise filtering, and data normalization to ensure data quality.

[0092] Based on the LSTM and ARIMA models, we built a photovoltaic forecasting and user demand analysis model. The LSTM model used the light intensity series and weather forecast of the past week to predict the photovoltaic power curve for the next 24 hours. K-means clustering divided users into three categories. Each user group used an independent ARIMA model to predict electricity consumption behavior, which improved the personalization and accuracy of the forecast.

[0093] Through the hybrid strategy of genetic algorithm and particle swarm optimization algorithm, we established a multi-objective optimization model. After 200 rounds of iterations, we obtained the optimal solution for cost, user satisfaction and supply-demand balance. The dynamic electricity price mechanism is set according to the model output. The peak, off-peak and off-peak electricity prices are 150%, 100% and 50% of the benchmark price respectively, and users are notified in real time through the APP and SMS platform.

[0094] The energy storage charging and discharging strategy is adaptively adjusted by the fuzzy logic controller, and is dynamically adjusted according to the load deviation, energy storage SOC and photovoltaic power change rate. When the SOC approaches the safety threshold, it automatically performs locking or unlocking operations to ensure stable operation of the system.

[0095] The details are shown in the following table:

[0096] Table 1 Experimental record table

[0097]

[0098] Through the four typical weather conditions shown in the table above, we can see that under sunny conditions, the photovoltaic predicted power reaches 120kW, which is less than the actual load of 105kW, indicating that the LSTM model predicts accurately, and the dynamic electricity price mechanism effectively guides user behavior, so that the actual load and the predicted curve maintain good consistency. At the same time, the user satisfaction survey results show that the peak shifting and valley filling response during peak hours is positive, indicating that the price strategy is effective. Compared with the conventional energy storage control method that does not adopt the present invention, the photovoltaic power prediction accuracy of this embodiment on cloudy and rainy days is improved by 15% and 20% respectively, and the user's electricity demand prediction error is reduced by 10%, which is directly reflected in the dynamic electricity price mechanism. It can be adjusted more flexibly, balance the supply and demand relationship, and avoid unnecessary power waste. Especially in the rapidly changing weather from sunny to cloudy, the system can quickly adjust the energy storage charging and discharging strategy, reduce load deviation, ensure the stable operation of the power grid, and demonstrate its adaptability and control ability to complex environments.

Claims

1. A control method based on distributed photovoltaic energy storage, characterized in that: include, Deploy intelligent monitoring and data collection networks, and implement full-link multi-dimensional dynamic data synchronization strategies for data management and analysis; Construct a photovoltaic prediction and user demand dynamic analysis model, and output photovoltaic power prediction curve and user electricity consumption behavior prediction based on the analyzed data; The output results of the photovoltaic prediction model are integrated with the prediction of user electricity consumption behavior, and a multi-objective optimization model is established to calculate the optimal load distribution under the dynamic electricity price mechanism; Recommend personalized energy-saving measures based on load forecast results; Dynamically adjust energy storage charging and discharging strategies based on real-time load changes and user responses; A photovoltaic prediction model is constructed, and the photovoltaic power prediction curve is output based on the analyzed data. The expression is: P PV (t+1)=LSTM([L(t),W(t),T(t),H(t)];θ LSTM ) Among them, P PV (t+1) is the predicted photovoltaic power value at time point t+1, L(t) is the observed value of light intensity, W(t) is the observed value of wind speed, T(t) is the observed value of temperature, H(t) is the observed value of humidity, θ LSTM is the parameter set of the LSTM model; The historical electricity consumption data U of user u at time t i (t) Apply K-means algorithm for clustering and get n groups, expressed as: C=KMeans(U1(t),U2(t),...,U m (t);n) Among them, c is the clustering result output by the K-means algorithm, m is the number of users, U1(t),U2(t),...,U m (t) is the electricity consumption data point of m users at time t, and n is the number of clusters to be formed; For each cluster group j, the ARIMA model is trained based on its historical electricity consumption data to predict the total electricity load demand in the next period. The expression is: in is the total predicted electricity load demand of cluster j in the next period, U j (t) is the observed power consumption of cluster j at time t, is the set of ARIMA model parameters specific to cluster group j; Clarify the optimization objectives and establish a multi-objective optimization model, as follows: Define the objective function: Cost function: f1(x) = C g +C s Among them, C g It is the cost of purchasing electricity from the power grid. In order to meet user needs, in addition to the electricity generated by itself, it is also necessary to purchase electricity from the external power grid. This cost varies with the peak and valley electricity prices and the purchase volume. s The energy storage operation cost refers to the cost of using energy storage equipment to store and release electrical energy in the system; User satisfaction function: in, is the actual satisfaction of the ith user, The expected satisfaction of the i-th user, N is the total number of users; Supply and demand balance indicator: f3(x)=|∑(Load pred (t)-P PV (t)-P s (t))| Among them, Load pred (t) is the total power demand of the user's predicted load at time t, P PV (t) is the predicted PV power at time t, P s (t) is the electric energy provided by the energy storage system to the grid and users at time t; Combining the basic electricity price, current load forecast and time period characteristics, the dynamic electricity price for each period is calculated, and the expression is: P price (t)=γ base ×(1+δ(t)×f load (t)+h(I(x,t))) Among them, P price (t) is the dynamic electricity price at time t, γ base It is the basic electricity price without time, load and market adjustment factors. δ(t) is the time adjustment coefficient, which is used to reflect the difference in electricity cost in different time periods of the day. During peak hours, δ(t)>0, and during off-peak hours, δ(t)<0. load (t) is the load peak-valley coefficient, which reflects the current position relative to the load level throughout the day. load (t)>1, low period f load (t)<1, the average period is close to 1, and h(I(x,t))) reflects the reasonable impact of the comprehensive cost-effectiveness index I on the electricity price; Combining the load forecast results and the real-time electricity price in each period, the energy-saving recommendation algorithm designed for user a is expressed as: Among them, S srec (a) is the energy saving amount recommended for user a, D u (a,t) predicts the power consumption of user a at time t, E eff (a) is the energy efficiency coefficient of user a.

2. The control method based on distributed photovoltaic energy storage according to claim 1, characterized in that: The deployment of the intelligent monitoring and data acquisition network includes installing smart meters, environmental sensors, energy storage system communication modules and photovoltaic power station grid monitoring systems; Smart meters regularly upload current, voltage, instantaneous power and accumulated energy data; Environmental sensors acquire light intensity, temperature, humidity, and wind speed data in real time; The energy storage system communication module monitors the battery status, charging rate, discharging rate, number of cycles, health status and fault alarm information in real time; The photovoltaic power plant grid monitoring system regularly synchronizes frequency, voltage level and real-time load demand.

3. The control method based on distributed photovoltaic energy storage according to claim 1, characterized in that: The implementation of a full-link multi-dimensional dynamic data synchronization strategy for data management and analysis includes using Apache Kafka message queue technology to build a high-throughput, low-latency data transmission channel and design a hierarchical storage architecture; Hot data is stored in a high-speed memory database for instant query, and cold data is regularly migrated to a distributed file system to reduce costs. The synchronization frequency is dynamically adjusted based on the importance and change frequency of the data. The Z-score method is used to identify outliers, and forward and backward filling and linear interpolation methods are used to process missing data. Random noise is removed based on statistical methods and numerical data is normalized.

4. The control method based on distributed photovoltaic energy storage according to claim 3, characterized in that: The photovoltaic prediction and user demand dynamic analysis model is constructed, and the photovoltaic power prediction curve and user power consumption behavior prediction are output based on the analyzed data, including using an LSTM model to receive the standardized historical light intensity sequence and weather forecast data at the input layer, using a multi-layer STM unit in the hidden layer to capture the time series pattern, and predicting the future photovoltaic power curve at the output layer; The K-means clustering algorithm is applied to group users based on their historical electricity consumption patterns. The ARIMA model is applied in each user group to capture the electricity consumption patterns of users in the group. The ARIMA model is trained for each group, and the historical electricity consumption data is input to output the predicted future electricity demand curve.

5. The control method based on distributed photovoltaic energy storage according to claim 4, characterized in that: The output of the comprehensive photovoltaic prediction model and the user's electricity consumption behavior prediction are used to establish a multi-objective optimization model to calculate the optimal load distribution under the dynamic electricity price mechanism, including integrating the photovoltaic prediction model output with the user's electricity consumption behavior prediction data and the current system status as an input data set for the optimization model; A hybrid strategy of genetic algorithm and particle swarm optimization algorithm is set, and the advantages of both are combined to perform multiple rounds of iterations. In each iteration, the genetic algorithm and particle swarm optimization algorithm respectively generate new solution sets, evaluate the quality of the solution through the fitness function, select high-quality solutions for crossover and mutation operations, and obtain the Pareto solution set after convergence. These solutions achieve the best compromise between cost, user satisfaction and supply and demand balance. From the Pareto solution set, a solution that meets the specific target priority is selected, and a dynamic electricity price mechanism is output to guide the energy storage charging and discharging strategy and photovoltaic output adjustment to ensure the optimization of load distribution.

6. The control method based on distributed photovoltaic energy storage according to claim 5, characterized in that: The personalized energy-saving measures recommended in combination with the load forecast results include load forecast curves, costs and user satisfaction indicators obtained by analysis based on the solution of the multi-objective optimization model, dividing the time periods into peak, off-peak and off-peak periods according to the load forecast, setting corresponding electricity price strategies, setting the peak period at 150% of the benchmark electricity price, the off-peak period at 100% of the benchmark electricity price, and the off-peak period at 50% of the benchmark electricity price, encouraging users to shift peak loads to fill off-peak loads, integrating the dynamic electricity price calculation logic into the existing billing system, and automatically adjusting the electricity price according to time; Design and develop a user interactive APP and SMS push platform, integrate a real-time electricity price information interface, automatically push real-time electricity price information according to the electricity price period in the user's area, develop a personalized energy-saving suggestion algorithm, generate energy-saving measures based on the user's historical electricity usage habits and equipment information, and regularly analyze user feedback and electricity usage data to evaluate energy-saving effects and user satisfaction. Based on feedback and analysis results, optimize the dynamic electricity price strategy, adjust the energy-saving suggestion algorithm, and improve user participation and satisfaction.

7. The control method based on distributed photovoltaic energy storage according to claim 6, characterized in that: The dynamic adjustment of energy storage charging and discharging strategies according to real-time load changes and user responses includes real-time collection of grid load data, actual power consumption of users, energy storage system status, and real-time power of photovoltaic power stations, which are uploaded to the central monitoring system through the Internet of Things platform every 5 minutes, and real-time comparison of the predicted load curve with the actual load data to calculate the deviation value. If the deviation exceeds a predetermined threshold, the adaptive adjustment mechanism is triggered; A fuzzy logic controller is designed for adaptive adjustment strategy. The maximum load deviation is defined as 5%. The load deviation, energy storage SOC, and photovoltaic power change rate are input. The energy storage charging and discharging instructions and photovoltaic power station adjustment instructions are output and a fuzzy rule base is constructed. If the load deviation is large and the SOC is high, the discharge amount is increased. If the photovoltaic power change rate is lower than expected and the energy storage SOC is low, the energy storage discharge amount is reduced to reserve power for use during the evening peak.

8. The control method based on distributed photovoltaic energy storage according to claim 7, characterized in that: The SOC safety threshold is set between 20% and 80%. If it is close to the upper limit of 80%, charging is stopped and discharge is prepared. If it is below 20%, only charging is allowed except in emergencies. Once the threshold is reached, the automatic locking operation is initiated and the vehicle can only be unlocked if manual intervention is obtained with advanced permissions or the system automatically determines an emergency situation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the control method based on distributed photovoltaic energy storage according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the control method based on distributed photovoltaic energy storage according to any one of claims 1 to 8 are implemented.

Images