Cooperative operation control method of alternative energy storage system

Abstract

The invention relates to a collaborative operation control method for an alternative energy storage system, and the method comprises the steps: determining a current application scene according to the operation data, obtained in real time, of the alternative energy storage system, and determining the current application scene from the application scenes of a pre-constructed space-time correlation model; calling a model according to the current scene, wherein the model input information comprises a current timestamp; the model determines a judgment standard based on the current timestamp and outputs a judgment result; the judgment result comprises a state judgment result, and when the state judgment result is abnormal, outputting a regulation and control target of an abnormal state at the same time; configuring a regulation strategy according to a regulation target output by the space-time correlation model; through the specific regulation and control strategies, the energy storage system can fully play a role in a peak load shifting application scene, the operation cost of a power system is optimized, and the reliability and stability of power supply are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of alternative energy storage, and in particular relates to a coordinated operation control method of an alternative energy storage system. Background Art

[0002] Alternative energy storage technology aims to solve the environmental problems and unstable energy supply caused by traditional fossil energy dependence. It is an important part of modern energy transformation, and the operation and control of alternative energy storage systems is also a top priority. Existing operation control methods include energy management system (EMS), automatic control, predictive control, etc. EMS is the general term for modern power grid dispatching automation system (including hardware and software), including data acquisition and monitoring system (SCADA system), automatic generation control (AGC) and economic dispatch control (EDC), power system state estimation (State Estimator), security analysis (Security Analysis), and dispatcher simulation training system (DTS).

[0003] Alternative energy storage operation control methods based on EMS include the following two aspects: (1) Real-time monitoring: Real-time monitoring of the state of the energy storage system, including voltage, current, temperature and other parameters, through sensors and data acquisition systems. (2) Optimal scheduling: Based on real-time data and forecast information (such as load forecast, renewable energy output forecast, etc.), optimize the charging and discharging strategy of the energy storage system to achieve efficient energy utilization; Alternative energy storage operation control methods based on automatic control include the following two aspects: (1) Closed-loop control: Automatically adjust the operating parameters of the energy storage system using feedback loops to ensure that it operates within the set target range. (2) PID control: Proportional-integral-differential controllers are used to adjust the output of the energy storage system to ensure its stability and response speed; Alternative energy storage operation control methods based on predictive control include the following two aspects: (1) Model predictive control (MPC): Based on the mathematical model of the energy storage system and its environment, predict the system behavior in the future and formulate the optimal control strategy accordingly. (2) Machine learning: Use machine learning algorithms to analyze historical data, predict future load and renewable energy output, and provide forward-looking control instructions for the energy storage system.

[0004] Problems with existing technologies: In the process of formulating optimal control strategies based on mathematical models, the real-time variables or variables within a certain period are often input into a unified model to output prediction results based on the variables. However, in this process, it is not considered that the differences in time and space may cause nonlinear characteristics in the variables obtained in real time, which will lead to large deviations in the prediction results obtained based on such nonlinear characteristics with obvious fluctuations, thus affecting the accuracy of subsequent control strategy formulation. Therefore, how to improve the accuracy of the output results of the mathematical model in view of the differences in time and space is a problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of this invention is that alternative energy storage has not yet been built on a large scale, and related research is also in the qualitative demonstration stage. Therefore, the applicable scenarios for evaluating the alternative effect of grid energy storage are usually limited to the construction of theoretical models. A mathematical model is designed that can improve the accuracy and comprehensiveness of the output results based on temporal and spatial differences.

[0006] The technical solution adopted by the present invention is: a coordinated operation control method of an alternative energy storage system, comprising the following steps:

[0007] S1: Determine the current application scenario based on the real-time acquired operation data of the alternative energy storage system, and determine the current application scenario from the application scenarios of the pre-built spatiotemporal correlation model;

[0008] S2: Call the model according to the current scenario, and the model input information includes the current timestamp;

[0009] S3: The model determines the judgment criteria based on the current timestamp and outputs the judgment result; the judgment result includes the state judgment result. When the state judgment result is abnormal, the control target of the abnormal state is output at the same time;

[0010] S4: Configure the control strategy according to the control objectives output by the spatiotemporal correlation model.

[0011] Specifically, in step S1, it is necessary to first construct a spatiotemporal correlation model in an application scenario, and one application scenario corresponds to one spatiotemporal correlation model; the specific steps are as follows:

[0012] Step S1.1: Divide the application scenarios according to the functions of alternative energy storage;

[0013] Step S1.2: Construct a spatiotemporal correlation model for each application scenario based on the historical operation data of the energy storage system and the power system for each application scenario.

[0014] Furthermore, in step S1.1, application scenarios are divided according to alternative energy storage functions, including peak shaving, ancillary services, renewable energy smoothing, microgrid and island operation, and load management.

[0015] Furthermore, in step S1, a discriminant model may be trained in advance to determine the current application scenario based on the input operation data.

[0016] Furthermore, the operating data includes electricity price data (electricity prices in different time periods), load data (electricity demand in different time periods), renewable energy data (solar or wind power generation forecasts), and environmental data (weather forecast information).

[0017] In the step S2, after constructing the model of each application scenario, it can be uniformly stored and managed, and the storage methods include blockchain, Internet of Things system or cloud platform; after determining the current application scenario, the model is called from the chain through the pre-assigned API interface, or the model is called from the management end of the Internet of Things system or cloud platform; when the corresponding model is called, the acquired operation data is input as the key variable into the called spatiotemporal correlation model, and the corresponding and corresponding timestamps of the operation data are also input.

[0018] In step S3, the spatiotemporal correlation model determines the judgment standard of the current time period according to the timestamp, and the discriminator in the spatiotemporal correlation model analyzes the input operation data according to the current judgment standard, thereby outputting a judgment result based on the time dimension.

[0019] In step S4, when the output result of the spatiotemporal correlation model indicates that the current system is operating normally, it means that the current parameters do not need to be adjusted and the current operating state is maintained; when the output result of the spatiotemporal correlation model indicates that the current system is operating abnormally, it means that the current parameters need to be adjusted, and a corresponding control strategy is formulated according to the control target synchronously output by the spatiotemporal correlation model.

[0020] The beneficial effects of the present invention are as follows: (1) different spatiotemporal correlation models are established in multiple application scenarios, so that the spatiotemporal correlation models can analyze and judge the operating status in combination with spatial features;

[0021] (2) Collect the operation data of the power system and alternative energy storage system in the whole life cycle under different seasonal backgrounds as training samples, and establish the correlation relationship between spatial features and associated features according to the time axis;

[0022] (3) The timestamp is used to call the judgment criteria of the discriminator, so that the analysis of the current spatiotemporal features can be matched according to the running data of the input model, and accurate judgment results can be output. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flow chart of the steps of the present invention; DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention, and are specifically described in combination with the embodiments below.

[0025] like Figure 1 As shown, the present invention comprises the following steps:

[0026] S1: According to the real-time acquired operation data of the alternative energy storage system, the current application scenario is determined from the application scenarios of the pre-built spatiotemporal correlation model. The specific process is as follows:

[0027] First of all, the application scenarios need to be divided according to the functions of alternative energy storage.

[0028] The classification based on functions is based on the functions that alternative energy storage realizes in the distribution network. In this way, the control strategy of alternative energy storage system is relatively stable, which is convenient for unified management and operation stability. According to the functions, the application scenarios are divided into etc.

[0029] Delay the expansion and upgrade of power transmission and distribution: Build energy storage at the receiving end of the grid section and upstream of the grid congested lines to increase the transmission capacity of the system section and alleviate transmission and distribution congestion.

[0030] Replacement of basic power supply in remote areas: Ensure stable power supply for agricultural and fishery production as well as the lives of rural and island residents in remote rural areas and isolated islands. Their renewable energy resources exceed the electricity load. Through new energy storage, support the development and utilization of renewable energy such as solar energy and wind energy, meet local energy needs, and reduce the comprehensive construction cost of power grid infrastructure.

[0031] Alternative power supply guarantee construction: energy storage is installed in important military facilities, government agencies, important conference venues, hospitals, precision production lines, etc.; during the summer load peak period, mobile energy storage is installed at the load center to ensure stable power supply.

[0032] Grid power supply improvement construction: At the end of the grid, in areas with weak distribution networks, and in areas with high penetration of new energy, energy storage is deployed to improve the voltage support capacity and grid power supply capacity.

[0033] Improve the construction of distributed new energy carrying capacity: In order to address the problem of power backflow or even reverse overload in areas with high new energy penetration, energy storage can be configured to assist in adjusting the supply and demand balance and improve the level of local consumption.

[0034] The application scenarios divided according to functions are generally determined when the alternative energy storage plan is formulated, that is, defined in advance; however, an alternative energy storage system can have multiple functions, and its functions can be flexibly adjusted according to changes in time and specific electricity parameters (electricity prices in different time periods, load data in different time periods, renewable resource power generation forecasts, meteorological data, etc.), that is, its application scenarios can change.

[0035] After completing the division of application scenarios, train the spatiotemporal correlation model for each application scenario.

[0036] The training process is as follows:

[0037] Step 1: Sample data collection

[0038] Collect historical operating data of the energy storage system (such as charge and discharge volume, efficiency, etc.), as well as related power system operating data (such as grid load, electricity price, etc.) and weather data (such as temperature, light intensity, etc.), that is, the power system operating data of the area covered by alternative energy storage; during the training process, the training objectives of the model may include state judgment results and abnormal state control objectives; the specific content of the training objectives is formulated according to its application scenario and can be configured by manual labeling.

[0039] For example:

[0040] In the peak shaving and valley filling application scenario, the training goal of the spatiotemporal correlation model is mainly to improve the energy storage system's ability to regulate the grid load, that is, to release stored energy when the grid load is peak and store energy when the load is low, so as to balance the grid load and improve the grid stability and economy.

[0041] Marked training target content:

[0042] Status judgment results: Based on the input information, the future change trend of the grid load is accurately predicted, including peak and valley periods. Based on the prediction results, the optimal charging and discharging state of the energy storage system is determined, such as when to charge, when to discharge, and the power of charging and discharging, so as to determine whether the current parameters are operating normally.

[0043] Control target for abnormal status: When the status judgment result shows that the current parameters are operating abnormally, identify and report abnormal conditions in the energy storage system or power grid, such as energy storage system failure and abnormal grid load fluctuation, and determine the specific parameters that need to be adjusted in the charging and discharging strategy based on the specific content of the abnormal situation, that is, the control target.

[0044] Step 2: Data preprocessing

[0045] First, the sample data is cleaned, standardized, and converted to ensure data quality and unify the format.

[0046] Data cleaning: remove missing values, outliers, etc.

[0047] Standardization: Standardize numerical features.

[0048] Then, time alignment and space matching are performed to ensure the consistency of time series data and spatial data in time and space dimensions.

[0049] Step 3: Feature extraction

[0050] Select features useful for model prediction from preprocessed data. Set the extracted features based on historical experience, or select the extracted features based on statistics or automatic feature selection algorithms. The extracted features need to include three aspects:

[0051] Time series features: Extract time series features from the historical operation data of the energy storage system, such as trend features, seasonal features, and periodic features.

[0052] Spatial characteristics: Based on geographic location information, spatial characteristics are extracted, such as the spatial distribution characteristics of regional electricity demand and supply, as well as natural geographical characteristics such as topography and landforms.

[0053] Correlation characteristics: Construct correlation characteristics between the energy storage system and the power system and weather system, such as the correlation coefficient between the charging and discharging of the energy storage system and the grid load, and the relationship between the energy storage efficiency and temperature.

[0054] Step 4: Model construction

[0055] Use deep learning models (such as LSTM, GRU and other recurrent neural networks) or machine learning models (such as support vector machines, random forests, etc.) to build a spatiotemporal correlation model in combination with spatiotemporal features. When designing the model structure, consider using convolutional neural networks (CNN) to process spatial features, using recurrent neural networks (RNN) to process time series features, and using fully connected layers (FC) for feature fusion and prediction.

[0056] During the model building process, feature fusion is required, that is, the time series features, spatial features and correlation features are fused to construct a comprehensive feature vector; among them, the correlation relationship between spatial features and correlation features is constructed according to the time series features, so that the spatial features and correlation features are arranged and fused according to the time axis, so that the judgment criteria for different time nodes in the same application scenario can be defined later.

[0057] Step 5: Model training

[0058] For example, for the peak shaving and valley filling application scenario, the training process is as follows:

[0059] The objective function should reflect the operating cost of the energy storage system and take into account the physical limitations of the energy storage system and the needs of the power grid. The objective function of a simple peak shaving and valley filling application scenario can be expressed as:

[0060] J(P charge ,P discharge )=∑ t (p(t)·P discharge (t)-p(t)·P charge (t))

[0061] Among them, J(P charge ,P discharge ) is the system operation cost within a certain period; p(t) is the electricity price at time t; P charge (t) is the charging power at time t; P discharge (t) is the discharge power at time t.

[0062] Constraints:

[0063] Condition 1: Energy conservation, the calculation formula is:

[0064]

[0065] Among them, S(t+1) and S(t) are the energy storage battery power at time t+1 and time t respectively; η charge and η discharge are the charging efficiency and discharging efficiency respectively, and Δt is the time interval.

[0066] Condition 2: State of charge limit, the calculation formula is:

[0067] S min ≤S(t)≤S max

[0068] Among them, S min and S max They are the minimum state of charge and maximum state of charge of the energy storage system respectively.

[0069] Condition 3: Power limitation, the calculation formula is:

[0070] 0≤P charge(t) ≤P max,charge

[0071] 0≤P discharge(t) ≤P max,discharge

[0072] Among them, P max,charge and P max,discharge They are the maximum charging power and maximum discharging power of the energy storage system respectively.

[0073] Status judgment result:

[0074] State judgment: The model needs to determine whether the current time period is a peak period or a low period based on the current time and electricity price.

[0075] Abnormal conditions: The model needs to be able to identify abnormal fluctuations in electricity demand and take action when necessary, such as adjusting charging and discharging strategies or triggering alarms.

[0076] Training process:

[0077] Data preparation: Collect historical electricity price data, electricity demand data, and energy storage system operation data.

[0078] Feature extraction: Extract useful features from historical data, such as electricity price trends, electricity demand fluctuations, etc.

[0079] Model training: Use machine learning algorithms (such as reinforcement learning, deep learning, etc.) to train the model so that it can make optimal charging and discharging decisions based on the current status (such as electricity price, state of charge, etc.).

[0080] Step 6: Model Validation

[0081] Use the validation data set to evaluate the performance of the model. Common evaluation indicators include mean square error (MSE), root mean square error (RMSE), R 2 Scores, etc.

[0082] Step 7: Model optimization and deployment

[0083] The model is tuned according to the verification results, such as adjusting model parameters, changing algorithms, etc. Finally, the optimized model is deployed in actual applications.

[0084] After the spatiotemporal correlation models of different application scenarios are established, unified storage management can be carried out. The storage locations can include blockchain, Internet of Things, cloud platforms, etc.

[0085] Obtain the operation data of the current cycle (one week or one month) of the alternative energy storage system, including but not limited to: electricity price data (electricity prices in different time periods), load data (electricity demand in different time periods), renewable energy data (solar or wind power generation forecast), environmental data (weather forecast information), and determine which application scenario the current scenario belongs to in the preset scenario (that is, the spatiotemporal correlation model is established) based on the operation data. Here, a discriminant model can be trained to determine the current application scenario. The training process refers to the training process of the above-mentioned spatiotemporal correlation model.

[0086] S2: Call the model according to the current scenario. The model input information includes the current timestamp. The specific process is as follows:

[0087] After determining the current application scenario, the model is called from the chain through the pre-assigned API interface, or from the management end of the IoT system or cloud platform.

[0088] When the corresponding model is called, the operating data obtained in step 2 is input as a key variable into the called spatiotemporal correlation model, along with the corresponding timestamps of the operating data.

[0089] S3: The model determines the judgment standard based on the current timestamp and outputs the judgment result; the judgment result includes the state judgment result. When the state judgment result is abnormal, the control target of the abnormal state is output at the same time. The specific process is: the spatiotemporal correlation model determines the judgment standard of the current time period according to the timestamp, and the discriminator in the spatiotemporal correlation model analyzes the input operation data according to the current judgment standard, thereby outputting the judgment result based on the time dimension.

[0090] The judgment results output by the spatiotemporal correlation model include a state judgment result and a control target of an abnormal state when the judgment result is abnormal.

[0091] S4: Configure the control strategy according to the control target output by the spatiotemporal correlation model. The specific process is as follows:

[0092] When the output result of the spatiotemporal correlation model indicates that the current system is operating normally, it means that the current parameters do not need to be adjusted and the current operating state is maintained;

[0093] When the output result of the spatiotemporal correlation model is that the current system is operating abnormally, it means that the current parameters need to be adjusted. According to the control objectives synchronously output by the spatiotemporal correlation model, the corresponding control strategy is formulated.

[0094] For example:

[0095] In the peak shaving and valley filling scenario, the control objectives output by the spatiotemporal correlation model include the following aspects:

[0096] Objective 1. Schedule charging and discharging, including:

[0097] Off-peak charging: Based on electricity price data and electricity demand data, the model identifies time periods with lower electricity prices (usually at night) and marks them as suitable time periods for charging the energy storage system. To achieve this goal, the specific strategy is as follows: During the night when electricity prices are lower (such as 22:00 to 06:00), the model recommends that the energy storage system be charged at maximum charging power until the set state of charge (SOC) upper limit is reached.

[0098] Discharge during peak hours: The model identifies time periods with higher electricity prices (usually during the day or evening) and marks them as recommended time periods for energy storage system discharge. For this goal, the specific strategy is: during the daytime when electricity prices are higher (such as 10:00 to 18:00), the model recommends that the energy storage system discharge at maximum discharge power until the set SOC lower limit is reached.

[0099] Objective 2. Dynamic SOC threshold control, including:

[0100] SOC threshold setting: The model will recommend dynamic adjustment of the upper and lower SOC thresholds based on the current electricity price, electricity demand and the state of the energy storage system. To achieve this goal, the specific strategy is as follows: when the electricity price is low, the model will set the SOC upper limit higher to store as much electricity as possible; when the electricity price is high, the model will set the SOC lower limit lower to release as much electricity as possible.

[0101] Objective 3. Response to abnormal situations, including:

[0102] Anomaly detection: The model monitors changes in electricity demand and electricity prices in real time, and outputs whether there are abnormal fluctuations. To achieve this goal, the specific strategy is: if electricity demand suddenly increases significantly during a certain period of time, the model will immediately start the energy storage system to discharge to supplement the power supply capacity of the power grid.

[0103] Objective 4. Forecasting and scheduling, including:

[0104] Load forecasting: The model combines historical data with current data to predict future electricity demand and electricity price changes, and outputs energy storage recommendations. To achieve this goal, if it is predicted that there will be a peak electricity price period tomorrow, the model will charge in advance tonight for use tomorrow.

[0105] Dynamic scheduling: Based on the prediction results, the model output dynamically adjusts the charging and discharging strategy to cope with future changes. To achieve this goal, the specific strategy is as follows: If it is predicted that tomorrow's electricity demand will increase significantly, the model will charge in advance tonight to ensure sufficient electricity supply tomorrow.

[0106] Through these specific regulation strategies, the energy storage system can fully play its role in peak shaving and valley filling application scenarios, optimize the operating costs of the power system, and improve the reliability and stability of power supply.

[0107] The present invention determines the current application scenario based on the real-time acquired operation data of the alternative energy storage system, and determines the current application scenario from the application scenarios of the pre-constructed spatiotemporal correlation model; calls the model according to the current scenario, and the model input information includes the current timestamp; the model determines the judgment standard based on the current timestamp and outputs the judgment result; the judgment result includes the state judgment result, and when the state judgment result is abnormal, the regulation target of the abnormal state is output at the same time; the regulation strategy is configured according to the regulation target output by the spatiotemporal correlation model. The present invention establishes different spatiotemporal correlation models under multiple application scenarios, so that the spatiotemporal correlation model can analyze and judge the operation status in combination with spatial features, thereby improving the accuracy of the model output results; collects the operation data of the power system and the alternative energy storage system in the whole life cycle under different seasonal backgrounds as training samples, and establishes the correlation relationship between spatial features and correlation features according to the time axis, so that the spatiotemporal correlation model can analyze and judge the operation status in combination with time features, further improving the accuracy and reliability of the output results; uses the timestamp to call the judgment standard of the discriminator, so that the analysis of the current spatiotemporal features can be matched according to the operation data of the input model, and an accurate discrimination result can be output.

[0108] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations within the meaning and scope of the elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A coordinated operation control method for an alternative energy storage system, characterized in that: The steps include: S1: Determine the current application scenario based on the real-time acquired operation data of the alternative energy storage system, and determine the current application scenario from the application scenarios of the pre-built spatiotemporal correlation model; S2: Call the model according to the current scenario, and the model input information includes the current timestamp; S3: The model determines the judgment criteria based on the current timestamp and outputs the judgment result; the judgment result includes the state judgment result. When the state judgment result is abnormal, the control target of the abnormal state is output at the same time; S4: Configure the control strategy according to the control objectives output by the spatiotemporal correlation model.

2. The coordinated operation control method of an alternative energy storage system according to claim 1, characterized in that: In step S1, it is necessary to first construct a spatiotemporal correlation model in the application scenario, and one application scenario corresponds to one spatiotemporal correlation model; the specific steps are as follows: Step S1.1: Divide the application scenarios according to the functions of alternative energy storage; Step S1.2: Construct a spatiotemporal correlation model for each application scenario based on the historical operation data of the energy storage system and the power system for each application scenario.

3. The coordinated operation control method of an alternative energy storage system according to claim 1, characterized in that: In the step S2, after constructing the model of each application scenario, it can be uniformly stored and managed, and the storage methods include blockchain, Internet of Things system or cloud platform; after determining the current application scenario, the model is called from the chain through the pre-assigned API interface, or the model is called from the management end of the Internet of Things system or cloud platform; when the corresponding model is called, the acquired operation data is input as the key variable into the called spatiotemporal correlation model, and the corresponding and corresponding timestamps of the operation data are also input.

4. The coordinated operation control method of an alternative energy storage system according to claim 1, characterized in that: In step S3, the spatiotemporal correlation model determines the judgment standard of the current time period according to the timestamp, and the discriminator in the spatiotemporal correlation model analyzes the input operation data according to the current judgment standard, thereby outputting a judgment result based on the time dimension.

5. The coordinated operation control method of an alternative energy storage system according to claim 1, characterized in that: In step S4, when the spatiotemporal correlation model outputs a result that the current system is operating normally, it means that the current parameters do not need to be adjusted and the current operating state is maintained; When the output result of the spatiotemporal correlation model is that the current system is operating abnormally, it means that the current parameters need to be adjusted. According to the control objectives synchronously output by the spatiotemporal correlation model, the corresponding control strategy is formulated.

Images