A Machine Learning-Based Prediction Method for New Energy Power Consumption and Energy Storage Regulation

Abstract

The present invention discloses a prediction method for new energy power consumption and energy storage regulation based on machine learning, belonging to the field of power technology. It includes four levels: data acquisition, processing, analysis, and visualization. Grid and energy storage system data are obtained. An active power balance analysis module and a particle swarm optimization algorithm are used to calculate the power imbalance. An LSTM model is used to predict the charge and discharge curves of the energy storage system and evaluate the regulation flexibility. It solves the technical problem of reducing the active power imbalance, decreasing the curtailment amount, predicting the charge and discharge requirements of the energy storage system, and improving the operating efficiency of the energy storage system by combining the particle swarm optimization algorithm and the long short-term memory network. It reduces the active power imbalance, decreases the curtailment amount, helps the power dispatching system to reasonably dispatch energy storage resources, improves the operating efficiency of the energy storage system, reduces unnecessary electric energy waste, improves the flexibility and reliability of the power grid, optimizes the power dispatching process, and enhances the accuracy and economy of dispatching.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power, and particularly relates to a prediction method for new energy power consumption and energy storage regulation based on machine learning. Background Art

[0002] The rapid development of renewable energy sources such as wind power and solar energy provides important support for achieving the global "dual carbon" goal. However, as the proportion of these new energy sources in the power system continues to increase, their volatility and intermittency characteristics pose great challenges to the stability and economy of the power system. Traditional power system dispatching relies on stable base load power sources, such as thermal power units. However, as the proportion of new energy sources such as wind power and solar energy increases, traditional power sources are gradually replaced, and the regulation ability on the power generation side drops significantly. Therefore, more flexible and efficient regulation resources need to be introduced to ensure the stable operation of the power system.

[0003] As an important flexible regulation means, the energy storage system can balance the supply and demand relationship of the power grid, alleviate the volatility of new energy power generation, and improve the regulation ability of the power grid. However, the reasonable configuration and efficient utilization of the energy storage system are crucial for the stable operation of the power system. Currently, there are still many technical problems in the reasonable prediction of energy storage capacity, dispatching optimization, and coordinated operation with the power grid.

[0004] When applying these technologies, existing research often ignores the comprehensive consideration of multi-dimensional constraint conditions and the impact of real-time dispatching on system stability. Most existing power dispatching methods rely on static optimization models, lacking consideration of the dynamic response of new energy power generation and real-time dispatching capabilities, and it is difficult to meet the dispatching requirements of the power system under the condition of high penetration of new energy. Summary of the Invention

[0005] The purpose of the present invention is to provide a prediction method for new energy power consumption and energy storage regulation based on machine learning, which solves the technical problem of reducing the active power imbalance, reducing the curtailment power, predicting the charge and discharge requirements of the energy storage system, helping the power dispatching system to reasonably dispatch energy storage resources, and improving the operation efficiency of the energy storage system by combining the particle swarm optimization algorithm and the long short-term memory network LSTM .

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A prediction method for new energy power consumption and energy storage regulation based on machine learning includes the following steps:

[0008] Step 1: Establish a data acquisition module cluster at the data acquisition layer. Obtain power grid and energy storage data from data sources in the power grid system through the Internet, including the power grid topology and operation parameters, the predicted output curve and actual output curve of new energy power stations, the charge and discharge capacity parameters of the energy storage system, the daily load prediction curve, and historical regulation data;

[0009] Step 2: The data processing and storage layer obtains the power grid and energy storage data from the data acquisition layer, preprocesses the power grid and energy storage data through a preprocessing module to obtain raw data, and constructs a raw data database;

[0010] Step 3: The analysis and decision-making layer retrieves the raw data from the raw data database and analyzes the raw data according to the following steps:

[0011] Step 3-1: Establish an active power balance analysis module, retrieve the raw data, construct an economic dispatch objective function with the minimum cost and constraints, calculate and analyze the power imbalance, and use the particle swarm optimization algorithm PSO to find the optimal solution and output the global optimal solution and the analysis result of the power imbalance;

[0012] Step 3-2: Establish an energy storage capacity regulation prediction module, obtain the output result of Step 1 and the raw data, and adopt LSTM a long short-term memory network model to predict the charge and discharge curve of the energy storage system in the future scheduling period, evaluate the regulation flexibility and regulation time of the energy storage, obtain the regulation flexibility index, and output the charge and discharge curve of the energy storage and the regulation flexibility index;

[0013] Step 4: The user interaction and visualization layer obtains the output result of the analysis and decision-making layer and constructs a visualization module to display the global optimal solution, the analysis result of the power imbalance, the charge and discharge curve of the energy storage, and the regulation flexibility index.

[0014] Preferably, at the data acquisition layer, the data acquisition module cluster includes a power grid data acquisition module, a new energy power station data acquisition module, an energy storage system data acquisition module, a load prediction data acquisition module, and a historical dispatch data acquisition module;

[0015] The data sources include a power grid control system, a power grid dispatching center, a new energy power station prediction system, an energy storage management system, and a load prediction system;

[0016] The power grid data acquisition module is responsible for obtaining the power grid topology and operation parameters from the power grid control system through the Internet;

[0017] The new energy power station data acquisition module is responsible for obtaining the predicted output curve and actual output curve of the new energy power station from the new energy power station prediction system through the Internet;

[0018] The energy storage system data acquisition module is responsible for obtaining the charge and discharge capacity parameters of the energy storage system from the energy storage management system through the Internet, including the charging power limit, discharging power limit, and charge-discharge efficiency of the energy storage system;

[0019] The load forecasting data acquisition module is responsible for obtaining the daily load forecasting curve from the load forecasting system through the Internet;

[0020] The historical dispatching data acquisition module is responsible for obtaining historical regulation data from the power grid dispatching center through the Internet, including the regulation data of historical load, generator power, and energy storage power.

[0021] Preferably, both the power grid control system and the new energy power station forecasting system are SCADA system.

[0022] Preferably, when performing step 3-1, the specific steps are as follows:

[0023] Step 3-1-1: Construct an economic dispatching objective function with the minimum cost and constraint conditions. The specific formula is as follows:

[0024] ;

[0025] Among them, is the cost of purchasing electricity from the power plant; is the amount of curtailed electricity; is the operating cost of the energy storage system; is the amount of electricity exported; is the amount of load loss; ρ is the curtailment penalty coefficient, with the unit of yuan / kWh , ε is the revenue coefficient of the exported electricity, with the unit of yuan / kWh , γ is the load loss penalty coefficient, with the unit of yuan / kWh ;

[0026] ;

[0027] Among them, is the unit power generation cost of the i th generator, with the unit of yuan / kWh ; is the output power of the i th generator at time t , with the unit of MW ; N gen is the total number of generators;

[0028] ;

[0029] Among them, is thei The predicted output of a new energy power station at time t , with the unit of MW . is the i th new energy power station's actual output at time t , with the unit of MW ; N renewable is the total number of new energy power stations; is the time step;

[0030] ;

[0031] Among them, C charge is the unit cost of energy storage charging, with the unit of yuan / kWh . C discharge is the unit cost of energy storage discharging, with the unit of yuan / kWh . is the total charging power of the energy storage system at t moment, with the unit of MW . is the total discharging power of the energy storage system at t moment, with the unit of MW ;

[0032] ;

[0033] Among them, is the power transmitted to the grid, with the unit of MW ; T is the number of time steps;

[0034] ;

[0035] Among them, is t moment's load demand, with the unit of MW ; t takes values from 1 to T ; is the i th energy storage device's output power at t moment;

[0036] Constraint conditions:

[0037] Power balance constraint:

[0038] ;

[0039] is the total power output of the energy storage system at time t ;

[0040] is the time t of the abandoned power:

[0041] ;

[0042] The constraint is: ;

[0043] Energy storage power limit:

[0044] ;

[0045] Among them, represents the maximum power limit of the energy storage system during discharge, represents the maximum power limit of the energy storage system during charging; is the power output of the energy storage system at time t ;

[0046] Generator power limit:

[0047] ;

[0048] Among them, represents the minimum power output limit of the i th generator set, represents the maximum power output limit of the i th generator set;

[0049] Step 3-1-2: Use the particle swarm optimization algorithm PSO to find the optimal solution, which specifically includes the following steps:

[0050] Step 3-1-2-1: Initialize the number of particles in the particle swarm and randomly generate a position in the solution space for each particle. The position of each particle represents a scheduling solution, and the position of the particle contains multiple variables, specifically including and ;

[0051] Step 3-1-2-2: For each particle, calculate the value of the objective function, and the objective function is expressed as:

[0052] ;

[0053] Step 3-1-2-3: According to the PSO algorithm, each particle updates its position and velocity based on its own historical best solution and the historical best solution of all particles:

[0054] ;

[0055] Among them, is the particle i at the t iteration speed; is the particle i at the t iteration position; w is the inertia weight; c Both 1 and c 2 are acceleration constants, controlling the speed at which the particle approaches the individual optimal solution pbest and the global optimal solution gbest ; r Both 1 and r 2 are random numbers, with a value range of [0, 1]; pbest i is the individual optimal solution of the particle i ; gbest is the global optimal solution of all particles;

[0056] Step 3-1-2-4: Constraint handling. Specifically, if the solution of the particle violates the constraint conditions, a penalty term is added to the solution to increase the objective function value and reduce the priority during the search process;

[0057] Step 3-1-2-5: Stopping condition. Specifically, reaching the maximum number of iterations or convergence of the global optimal solution;

[0058] Step 3-1-3: Output results, including outputting the global optimal solution and the analysis results of the power imbalance;

[0059] The global optimal solution is the optimal scheme for power grid scheduling, including the output plans of each motor and the charge and discharge plans of each energy storage system;

[0060] The analysis results of the power imbalance include if , it indicates insufficient peak shaving capacity or imbalance between supply and demand. Otherwise, there is no imbalance.

[0061] Preferably, when performing Step 3-2, it specifically includes the following steps:

[0062] Step 3-2-1: The input data includes the output results obtained from Step 3-1 and the original data;

[0063] Step 3-2-2: Construct a LSTM model based on time series prediction. The input layer includes the predicted output of new energy, load prediction, historical regulation data, charge and discharge capacity parameters of the energy storage system, the output plans of each generator, and the charge and discharge plans of the energy storage system; the hidden layer includes extracting time series features through multiple LSTM units; the output layer includes predicting the charge and discharge power of the energy storage system at future moments;

[0064] Step 3-2-3: Use historical regulation data forLSTM Train the model to optimize the model parameters;

[0065] Step 3-3-4: Output the prediction results, including the charge-discharge curve of the energy storage and the regulation flexibility index;

[0066] The charge-discharge curve of the energy storage is the charge-discharge power curve used to predict the energy storage system at future moments;

[0067] The regulation flexibility index is used to evaluate the regulation ability of the energy storage system in different time periods, including the maximum response time of charging / discharging and the flexibility index.

[0068] A new energy power consumption and energy storage regulation prediction method based on machine learning according to the present invention solves the problem of combining the particle swarm optimization algorithm and the long short-term memory network LSTM , reducing the active power imbalance, reducing the curtailment power, predicting the charge-discharge demand of the energy storage system, helping the power dispatching system to reasonably dispatch the energy storage resources, and improving the operation efficiency of the energy storage system. The present invention optimizes the active power balance of the power grid through the particle swarm optimization ( PSO ) algorithm, can quickly respond to the fluctuations of new energy and load changes, reduce the active power imbalance, reduce the curtailment power, and improve the stability and economy of the power grid. Through LSTM The network can accurately predict the charge-discharge demand of the energy storage system by learning historical data, help the power dispatching system to reasonably dispatch the energy storage resources, improve the operation efficiency of the energy storage system, reduce unnecessary electric energy waste, and perform real-time scheduling through the particle swarm optimization algorithm, can timely respond to the dynamic changes in the operation of the power grid, quickly adjust the power grid load and the charge-discharge plan of the energy storage system, improve the flexibility and reliability of the power grid, comprehensively consider multi-dimensional constraint conditions such as the power grid topology structure, generator output, energy storage power limit, curtailment penalty cost, etc., optimize the power dispatching process, and improve the accuracy and economy of the dispatching. Description of the Drawings

[0069] Figure 1 is the main flow chart of the present invention;

[0070] Figure 2 is the flow chart of Step 3-1 of the present invention;

[0071] Figure 3 is the flow chart of Step 3-2 of the present invention;

[0072] Figure 4 is the system architecture diagram of the present invention. Detailed Embodiments

[0073] Consisting of Figures 1 - 4 A new energy power consumption and energy storage regulation prediction method based on machine learning shown in the following steps:

[0074] Step 1: Establish a data acquisition module cluster at the data acquisition layer. Obtain power grid and energy storage data from data sources in the power grid system through the Internet, including the power grid topology and operation parameters, the predicted output curve and actual output curve of new energy power stations, the charge and discharge capacity parameters of the energy storage system, the daily load prediction curve, and historical regulation data;

[0075] At the data acquisition layer, the data acquisition module cluster includes a power grid data acquisition module, a new energy power station data acquisition module, an energy storage system data acquisition module, a load prediction data acquisition module, and a historical dispatching data acquisition module;

[0076] The data sources include a power grid control system, a power grid dispatching center, a new energy power station prediction system, an energy storage management system, and a load prediction system; the power grid control system, the power grid dispatching center, the new energy power station prediction system, the energy storage management system, and the load prediction system are all existing technologies, and in this embodiment, data in these systems is obtained through the Internet.

[0077] The power grid data acquisition module is responsible for obtaining the power grid topology and operation parameters from the power grid control system through the Internet;

[0078] The power grid control system ( Grid Control System , GCS ) is one of the core components of the power system, responsible for the real-time monitoring, dispatching, and fault handling of the power grid. It ensures the stable operation of the power grid through the control of power grid equipment (such as substations, switches, generators, etc.).

[0079] In this embodiment, the power grid control system is usually based on SCADA ( Supervisory Control and Data Acquisition ) system, which is a widely used remote monitoring and control technology that can obtain the operating status of the power grid (such as parameters like current, voltage, frequency, power, etc.) in real time and perform necessary control operations.

[0080] The power grid topology and operation parameters include the configuration and operation parameters of power grid nodes, lines, power sources, loads, etc. The input data participating in the calculation in Step 3 includes:

[0081] The output power of each generator at time t ;

[0082] The minimum power output limit of each generator set;

[0083] The maximum power output limit of each generating unit;

[0084] The total output power of the energy storage system at time t ;

[0085] At t the load demand at the moment;

[0086] The unit power generation cost of each generator;

[0087] The power transmitted out of the power grid.

[0088] The data acquisition module of the new energy power station is responsible for obtaining the predicted output curve and actual output curve of the new energy power station from the new energy power station prediction system through the Internet;

[0089] In this embodiment, the predicted output curve and the actual output curve are the power output data of new energy power stations such as wind power and photovoltaic power, including predicted data and actual operation data. The input data participating in the calculation in step 3 includes:

[0090] The predicted output of each new energy power station at time t ;

[0091] The actual output of each new energy power station at time t ;

[0092] The amount of abandoned electricity in the calculation of the abandoned electricity cost;

[0093] The new energy power station prediction system is mainly used to predict the power generation output of new energy (such as wind power and photovoltaic power) power stations. The power generation of new energy is affected by factors such as weather and environment, and has volatility and uncertainty. Therefore, the prediction system can provide more accurate new energy power generation predictions for the power grid dispatching center.

[0094] The new energy prediction system can be modeled based on meteorological data, historical output data, and machine learning algorithms (such as neural networks, LSTM etc.) to predict the output of new energy power stations such as wind power and photovoltaic power.

[0095] The data acquisition module of the energy storage system is responsible for obtaining the charge and discharge capacity parameters of the energy storage system from the energy storage management system through the Internet, including the charging power limit, discharging power limit, and charge and discharge efficiency of the energy storage system. The input data participating in the calculation in step 3 includes:

[0096] The energy storage system at tTotal charging power at a moment;

[0097] The energy storage system at t Total discharging power at a moment;

[0098] C charge Unit cost of energy storage charging;

[0099] C discharge Unit cost of energy storage discharging;

[0100] Maximum power limit for the energy storage system to discharge;

[0101] Maximum power limit for the energy storage system to charge.

[0102] The energy storage management system ( Energy Storage Management System , ESMS ) is a technical system used to manage the operation and scheduling of large-scale energy storage facilities. It can maximize its efficiency and economic value by intelligently controlling the charging and discharging processes of the energy storage system.

[0103] Modern energy storage management systems integrate advanced monitoring, control, and optimized scheduling functions. Energy storage systems include various forms such as battery energy storage and flywheel energy storage. ESMS It can monitor the charging and discharging status of the battery, manage the health status of the battery pack, and perform scheduling according to factors such as grid demand and market price.

[0104] The load forecasting data acquisition module is responsible for obtaining the daily load forecasting curve from the load forecasting system through the Internet;

[0105] The daily load forecasting curve is the forecasting data of future load demand. The input data participating in the calculation in step 3 includes:

[0106] Load shedding amount.

[0107] In this embodiment, the power grid dispatching center is equipped with a power dispatching automation system ( EMS , Energy Management System ), which integrates functions such as load forecasting, power grid optimized scheduling, and real-time monitoring, and is widely used in the operation and scheduling of the power grid. EMS It is not only responsible for dispatching but also can analyze the load fluctuations of the power grid in real time and adjust the output of the generating units.

[0108] The power grid dispatching center is the core unit in the power system responsible for coordinating various power resources and dispatching power generation, energy storage, and load to meet the load demand of the power grid. It conducts real-time load forecasting, power generation dispatching, and optimal dispatching through the operation of the dispatching system.

[0109] Currently, the load forecasting system can build models by combining historical load data, weather factors, economic activities, etc. Common algorithms include time series analysis (such as ARIMA ), machine learning methods (such as XGBoost , LSTM ), etc. Its purpose is to predict the load demand of the power grid in advance to avoid problems of power shortage or excess.

[0110] The load forecasting system ( Load Forecasting System ) is used to predict the change of the power grid load in a future period of time, helping the power grid dispatching center reasonably arrange power generation and dispatching plans.

[0111] The historical dispatching data acquisition module is responsible for obtaining historical regulation data from the power grid dispatching center through the Internet, including historical load, generator power, and energy storage power regulation data. The input data participating in the calculation in step 3 includes:

[0112] In the historical data:

[0113] The output power of each generator at time t ;

[0114] The total charging power of the energy storage system at t time;

[0115] The total discharging power of the energy storage system at t time;

[0116] The output power of each energy storage device at t time.

[0117] Both the power grid control system and the new energy power station forecasting system are SCADA systems.

[0118] The present invention adopts a multi-layer architecture, including a data acquisition layer, a data processing layer, an analysis and decision-making layer, and a user interaction and visualization layer. This architecture supports efficient data flow and module collaboration.

[0119] Step 2: The data processing and storage layer obtains the power grid and energy storage data of the data acquisition layer, preprocesses the power grid and energy storage data through a preprocessing module to obtain raw data, and constructs a raw data database;

[0120] The preprocessing includes data cleaning, data normalization, data interpolation, etc.

[0121] Step 3: The analysis and decision-making layer retrieves the original data from the original data database and analyzes the original data according to the following steps:

[0122] Step 3-1: Establish an active power balance analysis module, retrieve the original data, construct an economic dispatch objective function with the minimum cost and constraints, calculate and analyze the power imbalance, and use the particle swarm optimization algorithm PSO to find the optimal solution, and output the global optimal solution and the analysis result of the power imbalance;

[0123] When performing Step 3-1, the specific steps are as follows:

[0124] Step 3-1-1: Construct an economic dispatch objective function with the minimum cost and constraints. The specific formulas are as follows:

[0125] ;

[0126] Among them, is the cost of purchasing electricity from the power plant; is the amount of curtailed electricity; is the operating cost of the energy storage system; is the amount of electricity transmitted out; is the amount of load loss; ρ is the curtailment penalty coefficient, with the unit of yuan / kWh , ε is the revenue coefficient of the electricity transmitted out, with the unit of yuan / kWh , γ is the load loss penalty coefficient, with the unit of yuan / kWh ;

[0127] In this embodiment, ρ represents the economic cost or penalty caused by curtailed electricity. The output of new energy (such as wind power and photovoltaic power) may not be fully consumed in some cases (i.e., "curtailed"). In order to reflect the economic loss of this curtailed electricity, this embodiment sets a curtailment penalty coefficient ρ to calculate the additional cost brought by curtailed electricity. In the objective function, represents the penalty cost caused by curtailed electricity.

[0128] ε is the revenue coefficient of the electricity transmitted out, represents the revenue from the electricity output to the external power grid. If the power grid can output more electricity to other power grids or regions, then will increase, thus bringing more revenue. Since it is in the form of a negative sign, it means that an increase in the amount of electricity transmitted out will reduce the total cost.

[0129] γ is the load shedding penalty coefficient, which represents the penalty cost of load shedding. If the power supply of the power grid cannot meet the predicted load demand, it will increase, resulting in an increase in system cost. Load shedding usually means unstable grid frequency or power outage events, and such situations need to be avoided as much as possible in this embodiment.

[0130] In this embodiment, the objective function aims to optimize the economic dispatch of the power grid, minimize the operating cost of the power grid, and at the same time balance various factors such as curtailed electricity, power transmitted out, and load shedding. By adjusting these penalty coefficients, the optimization algorithm can find the optimal power supply plan under various dispatch conditions to ensure the economy and stability of the power grid.

[0131] Curtailed electricity penalty coefficient ρ encourages reducing curtailed electricity and maximizing the consumption of new energy power; the revenue coefficient of the power transmitted out ε encourages the power transmitted out and increases the economic revenue of the power grid. The load shedding penalty coefficient γ penalizes the situation where the power grid cannot meet the load demand to ensure the reliability of the power grid power supply.

[0132] ;

[0133] Among them, is the unit power generation cost of the i th generator, with the unit of yuan / kWh ; is the output power of the i th generator at time t , with the unit of MW ; N gen is the total number of generators;

[0134] ;

[0135] Among them, is the predicted output of the i th new energy power station at time t , with the unit of MW , is the actual output of the i th new energy power station at time t , with the unit of MW ; N renewable is the total number of new energy power stations; is the time step;

[0136] ;

[0137] Among them, Ccharge is the unit cost of energy storage charging, in yuan / kWh , C discharge is the unit cost of energy storage discharging, in yuan / kWh , is the total charging power of the energy storage system at t moment, in MW , is the total discharging power of the energy storage system at t moment, in MW ;

[0138] ;

[0139] Among them, is the power transmitted by the power grid, in MW ; T is the number of time steps;

[0140] ;

[0141] Among them, is t the load demand at MW moment, in t takes values from 1 to T ; is the output power of the i th energy storage device at t moment;

[0142] Constraint conditions:

[0143] Power balance constraint:

[0144] ;

[0145] is the total power output of the energy storage system at time t ;

[0146] is the curtailed power at time t :

[0147] ;

[0148] The constraint of is:

[0149] Energy storage power limit:

[0150] ;

[0151] Among them, Indicates the maximum power limit for the energy storage system during discharge, Indicates the maximum power limit for the energy storage system during charging; is the power output of the energy storage system at time t ;

[0152] Generator power limit:

[0153] ;

[0154] Among them, represents the minimum power output limit of the i th generator set, represents the maximum power output limit of the i th generator set;

[0155] Step 3-1-2: Use the particle swarm optimization algorithm PSO to find the optimal solution, which specifically includes the following steps:

[0156] Step 3-1-2-1: Initialize the number of particles in the particle swarm and randomly generate a position for each particle in the solution space. The position of each particle represents a scheduling solution, and the position of the particle contains multiple variables, specifically including and ;

[0157] Step 3-1-2-2: For each particle, calculate the value of the objective function, which is expressed as:

[0158] ;

[0159] Step 3-1-2-3: According to the PSO algorithm, each particle updates its position and velocity based on its own historical best solution and the historical best solution of all particles:

[0160] ;

[0161] Among them, is the velocity of particle i at the t th iteration; is the position of particle i at the t th iteration; w is the inertia weight; c 1 and c 2 are both acceleration constants, controlling the speed at which the particle approaches the individual optimal solution pbest and the global optimal solution gbest ; r 1 and r 2 are both random numbers, with a value range of [0, 1]; pbesti is the individual optimal solution of the particle i ; gbest is the global optimal solution of all particles;

[0162] Step 3-1-2-4: Constraint handling. Specifically, if the solution of the particle violates the constraint conditions, a penalty term is added to the solution to increase the objective function value and reduce the priority during the search process;

[0163] Step 3-1-2-5: Stopping condition. Specifically, it is when the maximum number of iterations is reached or the global optimal solution converges;

[0164] Step 3-1-3: Output results, including outputting the global optimal solution and the analysis result of the power imbalance;

[0165] The global optimal solution is the optimal scheme for power grid scheduling, including the output plans of each generator and the charge and discharge plans of each energy storage system;

[0166] In this embodiment, the output plan of the generator includes the power output plan of each generator in each time period;

[0167] The charge and discharge plan of the energy storage system includes the charge and discharge plans of each energy storage system in each time period.

[0168] The analysis result of the power imbalance is , including if , it indicates insufficient peak shaving capacity or imbalance between supply and demand. On the contrary, there is no imbalance.

[0169] Step 3-2: Establish an energy storage capacity regulation prediction module, obtain the output result of Step 1 and the original data, and use LSTM a long short-term memory network model to predict the charge and discharge curve of the energy storage system in the future scheduling period, evaluate the regulation flexibility and regulation time of the energy storage, obtain the regulation flexibility index, and output the charge and discharge curve of the energy storage and the regulation flexibility index;

[0170] When performing Step 3-2, it specifically includes the following steps:

[0171] Step 3-2-1: The input data includes the output result obtained from Step 3-1 and the original data;

[0172] Step 3-2-2: Construct a LSTM model based on time series prediction. The input layer includes the predicted output of new energy, load prediction, historical regulation data, charge and discharge capacity parameters of the energy storage system, the output plans of each generator, and the charge and discharge plans of the energy storage system; the hidden layer includes extracting time series features through multiple LSTM units; the output layer includes predicting the charge and discharge power of the energy storage system at future moments;

[0173] In this embodiment, LSTM The time series features input by the input layer of the model include:

[0174] The output plan of each generator;

[0175] The charge and discharge plan of the energy storage system;

[0176] The analysis result of power imbalance ;

[0177] The i predicted output of the t th new energy power station at time

[0178] The load demand at t moment;

[0179] In the historical data the output power of generator i at time t and the i th energy storage device at t time;

[0180] The maximum power limit for the energy storage system to discharge;

[0181] The maximum power limit for the energy storage system to charge.

[0182] The hidden layer extracts time series features through multiple LSTM units, capturing the time dependencies between each moment and different inputs;

[0183] The output layer outputs the charge and discharge power of the energy storage system within the future scheduling period, specifically including the energy storage charge and discharge curve , indicating the predicted charge and discharge power of each energy storage system i at time t ; The regulation flexibility index, including the maximum response time and the flexibility index, is used to evaluate the regulation ability of the energy storage system, that is, when facing different load or new energy output fluctuations, the charging / discharging ability of the energy storage system and its duration.

[0184] Step 3-2-3: Use historical regulation data to LSTM train the model and optimize the model parameters;

[0185] In this embodiment, use historical regulation data to LSTM train the model and optimize the model parameters (for example, LSTMThe weights and biases). The goal of training is to enable the model to predict how the energy storage system responds and the charging and discharging strategies of the energy storage under different grid loads and new energy outputs.

[0186] The objective function during the training process is:

[0187] ;

[0188] Among them, is the actual charging and discharging power in the historical data, is LSTM the charging and discharging power predicted by the model.

[0189] Step 3-3-4: Output the prediction results, including the energy storage charging and discharging curve and the regulation flexibility index;

[0190] The energy storage charging and discharging curve is the curve for predicting the charging and discharging power of the energy storage system at future moments;

[0191] The regulation flexibility index is used to evaluate the regulation ability of the energy storage system in different time periods, including the maximum response time of charging / discharging and the flexibility index.

[0192] Step 4: The user interaction and visualization layer obtains the output results of the analysis and decision-making layer, constructs a visualization module to display the power supply-demand balance state of the grid, the energy storage regulation ability and the energy storage charging and discharging curve, the generator output and the dispatching plan, and the power imbalance analysis.

[0193] In this embodiment, the data content displayed by the user interaction and visualization layer is all presented in the form of charts, specifically including:

[0194] The power supply-demand balance state of the grid: including showing the grid load at the current moment and the gap between the grid power supply capacity (i.e., the power provided by all generators and energy storage systems);

[0195] Showing the curtailed power .

[0196] The energy storage regulation ability and the energy storage charging and discharging curve: including showing the energy storage charging and discharging curve during the prediction period ;

[0197] Showing the regulation flexibility index, such as the maximum response time of charging / discharging, the flexibility index, etc.;

[0198] Showing the maximum power limit of the energy storage system during discharging and the maximum power limit of the energy storage system during charging;

[0199] Showing the output power of the energy storage system at this moment;

[0200] Show the charge and discharge plan of the overall energy storage system.

[0201] Generator output and scheduling plan: including showing the output plan of the overall generator.

[0202] Power imbalance analysis: including showing the power imbalance amount (load shedding amount ) of the power grid.

[0203] In this embodiment, the data acquisition layer, the data processing and storage layer, the analysis and decision-making layer, and the user interaction and visualization layer are all deployed in the cloud server and communicate with each other through the Internet.

[0204] A new energy power consumption and energy storage regulation prediction method based on machine learning according to the present invention solves the problem of combining the particle swarm optimization algorithm and the long short-term memory network LSTM , reducing the active power imbalance amount, reducing the curtailment amount, predicting the charge and discharge demand of the energy storage system, helping the power dispatching system to reasonably dispatch the energy storage resources, and improving the operation efficiency of the energy storage system. The present invention optimizes the active power balance of the power grid through the particle swarm optimization ( PSO ) algorithm, can quickly respond to new energy fluctuations and load changes, reduce the active power imbalance amount, reduce the curtailment amount, and improve the stability and economy of the power grid. Through LSTM The network can accurately predict the charge and discharge demand of the energy storage system by learning historical data, help the power dispatching system to reasonably dispatch the energy storage resources, improve the operation efficiency of the energy storage system, and reduce unnecessary electric energy waste. Through real-time scheduling by the particle swarm optimization algorithm, it can timely respond to the dynamic changes in the operation of the power grid, quickly adjust the power grid load and the charge and discharge plan of the energy storage system, improve the flexibility and reliability of the power grid, comprehensively consider multi-dimensional constraint conditions such as the power grid topology structure, generator output, energy storage power limit, and curtailment penalty cost, optimize the power dispatching process, and improve the accuracy and economy of dispatching.

Claims

1. A prediction method for new energy power consumption and energy storage regulation based on machine learning, characterized in that: It includes the following steps: Step 1: Establish a data acquisition module cluster at the data acquisition layer. Obtain power grid and energy storage data from data sources in the power grid system through the Internet, including the power grid topology and operation parameters, the predicted output curve and actual output curve of new energy power stations, the charge and discharge capacity parameters of the energy storage system, the daily load prediction curve, and historical regulation data; Step 2: The data processing and storage layer obtains the power grid and energy storage data from the data acquisition layer, preprocesses the power grid and energy storage data through a preprocessing module to obtain raw data, and constructs a raw data database; Step 3: The analysis and decision-making layer retrieves the raw data from the raw data database and analyzes the raw data according to the following steps: Step 3-1: Establish an active power balance analysis module, retrieve the original data, construct an economic dispatch objective function with the minimum cost and constraints, calculate and analyze the power imbalance, and use the particle swarm optimization algorithm PSO to find the optimal solution, and output the global optimal solution and the analysis results of the power imbalance; Construct an economic dispatch objective function with the minimum cost and constraint conditions. The specific formulas are as follows: ； Among them, is the cost of purchasing electricity from the power plant; is the amount of curtailed electricity; is the operating cost of the energy storage system; is the amount of electricity transmitted out; is the amount of load loss; ρ is the curtailment penalty coefficient, with the unit of yuan / kWh , ε is the revenue coefficient of the electricity transmitted out, with the unit of yuan / kWh , γ is the load loss penalty coefficient, with the unit of yuan / kWh ; ； Among them, is the unit power generation cost of the i th generator, with the unit of yuan / kWh ; is the output power of the i th generator at time t , with the unit of MW ; N gen is the total number of generators; ； Among them, is the i th predicted output of the new energy power station at time t , with the unit of MW ; is the i th actual output of the new energy power station at time t , with the unit of MW ; N renewable is the total number of new energy power stations; is the time step; ； Among them, C charge is the unit cost of energy storage charging, with the unit of yuan / kWh , C discharge is the unit cost of energy storage discharging, with the unit of yuan / kWh , is the total charging power of the energy storage system at t moment, with the unit of MW , is the total discharging power of the energy storage system at t moment, with the unit of MW ; ； Among them, is the power transmitted out of the power grid, with the unit of MW ; T is the time step; ； Among them, is t the load demand at the moment, with the unit of MW ; t takes values from 1 to T ; is the output power of the i th energy storage device at the moment of t ; Step 3-2: Establish an energy storage capacity regulation prediction module, obtain the output result of Step 1 and the original data, and use LSTM a long short-term memory network model to predict the energy storage charge and discharge curve of the energy storage system within the future scheduling period, evaluate the regulation flexibility and regulation time of the energy storage, obtain the regulation flexibility index, and output the energy storage charge and discharge curve and the regulation flexibility index; Step 4: The user interaction and visualization layer obtains the output results of the analysis and decision-making layer, and constructs a visualization module to display the global optimal solution, the analysis results of power imbalance, the charge and discharge curve of the energy storage, and the regulation flexibility index.

2. A prediction method for new energy power consumption and energy storage regulation based on machine learning according to claim 1, characterized in that: At the data acquisition layer, the data acquisition module cluster includes a power grid data acquisition module, a new energy power station data acquisition module, an energy storage system data acquisition module, a load prediction data acquisition module, and a historical dispatch data acquisition module; The data sources include a power grid control system, a power grid dispatch center, a new energy power station prediction system, an energy storage management system, and a load prediction system; The power grid data acquisition module is responsible for obtaining the power grid topology and operation parameters from the power grid control system through the Internet; The new energy power station data acquisition module is responsible for obtaining the predicted output curve and actual output curve of the new energy power station from the new energy power station prediction system through the Internet; The energy storage system data acquisition module is responsible for obtaining the charge and discharge capacity parameters of the energy storage system from the energy storage management system through the Internet, including the charging power limit, discharging power limit, and charge and discharge efficiency of the energy storage system; The load prediction data acquisition module is responsible for obtaining the daily load prediction curve from the load prediction system through the Internet; The historical dispatch data acquisition module is responsible for obtaining historical regulation data from the power grid dispatch center through the Internet, including historical load, generator power, and energy storage power regulation data.

3. A method for predicting new energy power consumption and energy storage regulation based on machine learning according to claim 2, characterized in that: The grid control system and the new energy power station prediction system are both SCADA systems.

4. A method for predicting new energy power consumption and energy storage regulation based on machine learning according to claim 1, characterized in that: When performing Step 3-1, the specific steps are as follows: Step 3-1-1: Construct an economic dispatch objective function with the minimum cost and constraint conditions. The specific formulas are as follows: ； Among them, is the cost of purchasing electricity from the power plant; is the amount of curtailed electricity; is the operating cost of the energy storage system; is the amount of electricity transmitted out; is the amount of load loss; ρ is the curtailment penalty coefficient, with the unit of yuan / kWh , ε is the revenue coefficient of the electricity transmitted out, with the unit of yuan / kWh , γ is the load loss penalty coefficient, with the unit of yuan / kWh ; ； Among them, is the unit power generation cost of the i th generator, with the unit of yuan / kWh ; is the output power of the i th generator at time t , with the unit of MW ; N gen is the total number of generators; ； Among them, is the i th predicted output of the new energy power station at time t , with the unit of MW ; is the i th actual output of the new energy power station at time t , with the unit of MW ; N renewable is the total number of new energy power stations; is the time step; ； Among them, C charge is the unit cost of energy storage charging, in yuan / kWh , C discharge is the unit cost of energy storage discharging, in yuan / kWh , is the total charging power of the energy storage system at t moment, in MW , is the total discharging power of the energy storage system at t moment, in MW ; ； Among them, is the power transmitted out of the power grid, with the unit of MW ; T is the time step; ； Among them, is t the load demand at the moment, with the unit of MW ; t takes values from 1 to T ; is the output power of the i th energy storage device at the moment of t ; Constraint conditions: Power balance constraint: ； is the total power output of the energy storage system at time t ; is the time t of the curtailed electricity volume: ； The constraints are as follows: ; Energy storage power limit: ； Among them, represents the maximum power limit for the energy storage system to discharge, represents the maximum power limit for the energy storage system to charge; is the power output of the energy storage system at time t ; Generator power limit: ； Among them, represents the minimum power output limit of the i th generator set, represents the maximum power output limit of the i th generator set; Step 3-1-2: Use the particle swarm optimization algorithm PSO to find the optimal solution, which specifically includes the following steps: Step 3-1-2-1: Initialize the number of particles in the particle swarm, and randomly generate a position in the solution space for each particle. The position of each particle represents a scheduling solution, and the position of the particle contains multiple variables, specifically including and ; Step 3-1-2-2: For each particle, calculate the value of the objective function, and the objective function is expressed as: ； Step 3-1-2-3: According to PSO the algorithm, each particle updates its position and velocity based on its own historical best solution and the historical best solution of all particles: ； wherein, is the velocity of the particle i at the t -th iteration; is the position of the particle i at the t -th iteration; w is the inertia weight; c 1 and c 2 are both acceleration constants, controlling the velocity of the particle approaching the individual optimal solution pbest and the global optimal solution gbest ; r 1 and r 2 are both random numbers, with values in the range [0, 1]; pbest i is the individual optimal solution of the particle i ; gbest is the global optimal solution of all particles; Step 3-1-2-4: Constraint handling. Specifically, if the solution of the particle violates the constraint conditions, a penalty term is added to the solution to increase the objective function value and reduce the priority in the search process; Step 3-1-2-5: Stop condition, specifically reaching the maximum number of iterations or convergence of the global optimal solution; Step 3-1-3: Output results, including outputting the global optimal solution and the analysis results of power imbalance; The global optimal solution is the optimal plan for power grid dispatching, including the output plans of each motor and the charge and discharge plans of each energy storage system; The analysis result of power imbalance includes that if , it indicates insufficient peak regulation capacity or imbalance between supply and demand. On the contrary, there is no imbalance.

5. A method for predicting new energy power consumption and energy storage regulation based on machine learning according to claim 4, characterized in that: When performing Step 3-2, it specifically includes the following steps: Step 3-2-1: The input data includes the output results obtained from Step 3-1 and the original data; Step 3-2-2: Construct a LSTM model. The input layer includes the predicted output of new energy, load prediction, historical regulation data, charge-discharge capacity parameters of the energy storage system, output plans of each generator, and charge-discharge plans of the energy storage system; the hidden layer includes LSTM units for extracting time series features; the output layer includes the predicted charge-discharge power of the energy storage system at future moments. Step 3-2-3: Use historical adjustment data to LSTM train the model and optimize the model parameters; Step 3-3-4: Prediction result output, including the charge and discharge curve of the energy storage and the regulation flexibility index; The charge and discharge curve of the energy storage is the curve for predicting the charge and discharge power of the energy storage system at future moments; The regulation flexibility index is used to evaluate the regulation ability of the energy storage system in different time periods, including the maximum response time of charging or discharging and the flexibility index.

Images