Energy storage scheduling method and system based on time sequence load prediction

Abstract

The invention belongs to the technical field of power system scheduling and resource management, and particularly relates to an energy storage scheduling method and system based on time sequence load prediction, and the method comprises the steps: obtaining the operation state information of a power system; predicting the operation load of the power system according to the acquired state information and a preset time sequence load prediction model to obtain an operation load prediction value; energy storage scheduling is carried out according to the obtained operation load prediction value, and energy storage scheduling based on time sequence load prediction is completed; wherein the preset time sequence load prediction model adopts an improved Prophet algorithm, a timestamp is determined according to a time sequence in the obtained state information, a time sequence load influence factor is determined based on the timestamp, and a time sequence load prediction model is determined according to the determined time sequence load influence factor.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system dispatching and resource management, and in particular relates to a method and system for energy storage dispatching based on time series load forecasting. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In the power system, the importance of energy storage dispatch is becoming increasingly prominent, especially in the context of the increasing proportion of new energy sources year by year. As a flexible device that can store and release energy, the energy storage system plays a key role. First, with the increasing proportion of renewable energy, the power system is facing daily and seasonal load fluctuations. The energy storage system can effectively store excess energy so that it can be released at peak demand times, thereby balancing the supply and demand on the energy supply side and the energy consumption side and improving the stability of the power system. Secondly, traditional power generation equipment is difficult to start and stop instantly, and the fast response characteristics of the energy storage system can realize charging and discharging in an instant, adapting to the needs of transient changes in the power system.

[0004] In the existing technology, the traditional dispatching method driven by rules and empirical rules is difficult to cope with the increasing complexity of the power system. Based on static rules, it cannot adapt to the dynamic and nonlinear load and energy fluctuations in the power system. And it is usually unable to effectively integrate the load forecasting and dispatching optimization on the user side, resulting in the failure to maximize the overall benefits of the system.

[0005] However, the current energy storage dispatching methods are still constrained by a series of limitations: the traditional dispatching methods driven by rules and empirical rules are difficult to cope with the growing complexity of the power system. Based on static rules, they cannot adapt to the dynamic and nonlinear load and energy fluctuations in the power system. Secondly, although the mathematical optimization model has improved the accuracy of dispatching to a certain extent, its response to the spatiotemporal dynamic changes of the power system is still insufficient, which makes the traditional dispatching methods perform poorly in dealing with new energy fluctuations, user load changes, and market electricity price uncertainties. In addition, the existing methods are usually unable to effectively integrate user-side load forecasting and dispatching optimization, resulting in the inability to maximize the overall benefits of the system. Therefore, the limitations of traditional energy storage dispatching methods urgently need to be solved through innovative methods to promote the dispatching of power systems in a more flexible, intelligent, and sustainable direction. Summary of the invention

[0006] To solve the above problems, the present invention proposes a method and system for energy storage scheduling based on time series load forecasting, which introduces load forecasting combined with time series to achieve accurate, real-time and adaptive optimization of energy storage scheduling, thereby improving the scheduling efficiency of the energy storage system and reducing the operating cost of the power system.

[0007] According to some embodiments, a first solution of the present invention provides an energy storage scheduling method based on time series load forecasting, which adopts the following technical solution: A method for energy storage scheduling based on time series load forecasting, comprising: Obtaining power system operation status information; According to the acquired state information and a preset time series load forecasting model, the operating load of the power system is forecasted to obtain an operating load forecast value; Energy storage scheduling is performed according to the obtained operating load forecast value, completing energy storage scheduling based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet The algorithm determines a timestamp according to the time sequence in the acquired state information, determines a time series load influencing factor based on the timestamp, and determines a time series load prediction model according to the determined time series load influencing factor.

[0008] As a further technical limitation, the time series load forecasting model is ;in, It is a time series load forecasting model; is a trend unit used to capture the nonlinear growth of time series load influencing factors. , , , and are all hyperparameters, is the nonlinear trend of the load, is the linear trend term used to capture the changing load demand; is a seasonal unit used to capture the periodicity of time series load influencing factors. , is the ordinal number of the seasonal component, is the total number of seasonal components, is the contribution of the ith seasonal component at time t, is the serial number of the periodic component related to the external factors, is the total number of cyclical components of external factors, is a periodic function related to external factors, is the influence weight of the periodic function; For holiday units, , is the total number of holidays, is the effect of the i-th holiday at time t; is the error term used to capture the random changes in the factors affecting the time series load, , is the error term at the previous moment, is the autoregressive parameter of the error term, is the white noise error.

[0009] Furthermore, the parameters of the time series load forecasting model are optimized by maximizing the log-likelihood estimation. In the process of parameter optimization, the parameters of the trend item, seasonal unit and holiday unit are adjusted by gradient descent, and the iteration is repeated until the prediction error of the time series load forecasting model is minimized.

[0010] As a further technical limitation, in the process of energy storage scheduling according to the obtained operating load forecast value, an improved particle swarm algorithm is used to simulate the group behavior of birds in a flock, and the particle position is continuously adjusted to find the optimal solution, thereby obtaining the optimal operating load forecast value, and completing the optimal scheduling of energy storage according to the obtained optimal operating load forecast value.

[0011] Furthermore, in the process of energy storage optimization scheduling, chaotic mapping is used to generate the speed and position of the initial particle swarm, the inertia weight is adjusted based on the adaptive weight and dynamic parameter adjustment mechanism, the speed and position of the particles are iteratively updated, and continuous iterations are carried out to prevent falling into the local optimum through fuzzy logic control, and the optimal energy storage scheduling searched by the particle swarm is obtained, that is, the global optimal energy storage scheduling strategy, which determines the charging and discharging operations of the energy storage system in each time period.

[0012] As a further technical limitation, the acquired power system operation status information includes at least historical power consumption data, market electricity price information, meteorological data, holiday power consumption data, and major event power consumption data.

[0013] According to some embodiments, the second solution of the present invention provides an energy storage scheduling system based on time series load forecasting, which adopts the following technical solutions: An energy storage dispatching system based on time series load forecasting, comprising: An acquisition module, configured to acquire power system operation status information; A prediction module is configured to predict the operating load of the power system according to the acquired state information and a preset time series load prediction model to obtain an operating load prediction value; A dispatching module, which is configured to perform energy storage dispatch according to the obtained operating load forecast value, and complete energy storage dispatch based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet algorithm, determines the timestamp according to the time sequence in the acquired status information, determines the time series load influencing factors based on the timestamp, and determines the time series load forecasting model according to the determined time series load influencing factors.

[0014] According to some embodiments, a third solution of the present invention provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a program thereon, which, when executed by a processor, implements the steps of a method for energy storage scheduling based on time series load forecasting as described in the first solution of the present invention.

[0015] According to some embodiments, a fourth solution of the present invention provides an electronic device, which adopts the following technical solution: An electronic device comprises a memory, a processor and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps in a method for energy storage scheduling based on time series load forecasting as described in the first solution of the present invention are implemented.

[0016] According to some embodiments, a fifth solution of the present invention provides a computer program product, which adopts the following technical solution: A computer program product includes software codes, wherein the program in the software codes executes the steps in the energy storage scheduling method based on time series load forecasting as described in the first solution of the present invention.

[0017] Compared with the prior art, the present invention has the following beneficial effects: The present invention introduces a time series load forecasting method to achieve accurate, real-time and adaptive optimization of energy storage scheduling, accurately predict the user-side load and then use it as the basis for scheduling on the energy supply side, to achieve accurate scheduling of the energy storage system, to meet the power demand to the maximum extent, and to improve the system operation efficiency; it supports the formulation of optimal functional strategies under different market conditions, making energy utilization more economical and sustainable, and reducing the cost of operation. Through real-time monitoring and adjustment of scheduling, it can flexibly respond to fluctuations in electricity market prices; it has obvious advantages in improving the efficiency of the power system, reducing operating costs and promoting intelligence. By comprehensively considering users and the energy supply side, it introduces advanced technical means to provide innovative solutions for the future development of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings in the specification that constitute a part of this embodiment are used to provide a further understanding of this embodiment. The schematic embodiments of this embodiment and their descriptions are used to explain this embodiment and do not constitute improper limitations on this embodiment.

[0019] Figure 1 This is a flow chart of the energy storage scheduling method based on time series load forecasting in Embodiment 1 of the present invention; Figure 2 This is an architecture diagram of the energy storage scheduling method based on time series load forecasting in the first embodiment of the present invention; Figure 3This is a structural block diagram of an energy storage scheduling system based on time series load forecasting in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0020] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0021] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.

[0022] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0023] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0024] In the present invention, terms such as "fixed connection", "connected", "connection", etc. should be understood in a broad sense, indicating that it can be fixedly connected, integrally connected or detachably connected; it can be directly connected or indirectly connected through an intermediate medium. For relevant scientific research or technical personnel in this field, the specific meanings of the above terms in the present invention can be determined according to specific circumstances, and they cannot be understood as limitations on the present invention.

[0025] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0026] Embodiment 1 Embodiment 1 of the present invention introduces an energy storage scheduling method based on time series load forecasting.

[0027] like Figure 1 A method for energy storage scheduling based on time series load forecasting is shown, comprising: Obtaining power system operation status information; According to the acquired state information and a preset time series load forecasting model, the operating load of the power system is forecasted to obtain an operating load forecast value; Energy storage scheduling is performed according to the obtained operating load forecast value, completing energy storage scheduling based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet The algorithm determines a timestamp according to the time sequence in the acquired state information, determines a time series load influencing factor based on the timestamp, and determines a time series load prediction model according to the determined time series load influencing factor.

[0028] This embodiment obtains power system operation status information through a monitoring system, a sensor or historical records, including: (1) Historical electricity consumption data: Obtain the user's hourly or minute-by-minute electricity consumption data through monitoring systems or sensors; (2) Market electricity price information: Considering the impact of peak and valley periods of market electricity prices on electricity load, the market electricity price is used as a predictable covariate input to the model; (3) Meteorological data: Obtain meteorological data that has a significant correlation with power load, including temperature, humidity, etc.; (4) Holiday and special event data: Considering the impact of holidays and major events on electricity consumption, date data is obtained as a static covariate for model input. Time series data is represented as ,in, Indicates the user's electricity consumption data. represents the other covariates proposed above.

[0029] After obtaining the power system operating status information, this embodiment needs to preprocess the data, mainly including data cleaning and resampling, and use interpolation to handle missing values ​​and abnormal values. If there are irregular time intervals, resampling is required to ensure that the time series has the same time interval.

[0030] The Prophet algorithm is a prediction algorithm based on time series. It has good adaptability to the periodic changes and trends of data. The algorithm can effectively integrate holiday information that has a significant impact on electricity consumption, greatly improving the accuracy of the prediction.

[0031] In this embodiment, the improved Prophet algorithm is used to capture the main patterns of load changes by fitting the trends, seasonality, holidays and other characteristics in the historical time series; the time series data is embodied as a series of electricity consumption data points arranged in chronological order, each data point includes a timestamp, the corresponding electricity consumption and other related covariates. Through these timestamps, features such as diurnal and seasonal changes, and fluctuations in electricity consumption on specific holidays can be extracted. The correlation between time series and load forecasting lies in that by observing and modeling patterns in historical data, the improved Prophet algorithm can extrapolate and predict future loads based on the historical characteristics of the time series, so that the model can capture the dynamic changes in electricity consumption over time. The trend part is used to simulate long-term changes, the seasonal part is used to capture periodic fluctuations, and the holiday part is used to deal with the impact of special time points.

[0032] The time series load forecasting model used in this embodiment is: Specific: (1) Trend unit is a trend unit used to capture the nonlinear growth of time series load influencing factors, that is, ; in, , , and are all hyperparameters, is the nonlinear trend of the load, is the linear trend term used to capture the changing load demand; (2) Seasonal unit is a seasonal unit used to capture the periodicity of the factors affecting the time series load, that is, ; in, is the serial number of the seasonal component, which is used to traverse all periodic components; represents the total number of seasonal components. This embodiment takes into account the periodicity of year, month, and week. =3; represents the contribution of the ith seasonal component at time t. The annual cycle affects the electricity load in winter and summer, and the weekly cycle is used to reflect the difference in electricity consumption between weekdays and weekends. Indicates the serial number of the periodic component related to the external factor, which is used to traverse the periodic components of all external factors; represents the total number of periodic components of external factors. This embodiment takes into account the periodic changes in temperature, weather and electricity demand. is a periodic function related to external factors, is the influence weight of the periodic function; (3) Holiday Unit is the holiday unit, that is ; in, is the total number of holidays, including the Spring Festival, National Day, etc. Each holiday has a significant impact on electricity demand; represents the effect of the i-th holiday at time t, which is used to describe the impact of a specific holiday on the electricity load; (4) Error term is the error term used to capture the random changes of the factors affecting the time series load, that is, ; in, is the error term of the previous moment, that is, the autocorrelation part in the autoregressive model, and is introduced The purpose is to capture the short-term fluctuation characteristics of time series data so that the model can more accurately predict short-term load changes; is the autoregressive parameter of the error term, is the white noise error.

[0033] In the process of training the time series load forecasting model, this embodiment optimizes the parameters of the model by maximizing the log-likelihood estimate to ensure a good fit to the time series data. During the optimization process, the parameters of trend, seasonality and holiday effects are adjusted by gradient descent, and the loss function is defined as the sum of squares of the forecast errors ,in, is the actual load value, is the model prediction value, are all model parameters. For the loss function Derivatives are taken to obtain the gradient of each parameter, which is used to represent the rate of change of the loss function relative to each parameter: , according to the learning rate Update each parameter and adjust the formula as follows: , the above process is iterated continuously until the prediction error is minimized. For example, for the parameter , , the same goes for other parameters.

[0034] This example uses a cross-validation method to evaluate the model and verify its performance on unseen data. 80% of the data is used for training and the remaining 20% ​​of the data is used for validation to ensure that the model has good generalization ability.

[0035] In the prediction stage, this embodiment inputs the pre-processed future influencing factors (such as meteorological data, holidays, etc.) into the trained model, and the improved Prophet algorithm combines these influencing factors to predict the load value of the future period by extrapolating the trend and seasonality. In the prediction process, the characteristics of the time series are reflected by capturing the time dependency of historical data and the correlation of load changes. The improved Prophet algorithm uses the timestamps in the time series to identify the change patterns of different time scales such as day, week, and month, so as to infer future load changes. For example, by learning the load patterns of different seasons and times in past electricity consumption data, the model can extrapolate similar patterns in the prediction, thereby improving the prediction accuracy. Time dependency ensures that the model can predict the future based on the past load change rules, and the correlation between load and time enables the model to capture the dynamic change characteristics of electricity consumption over time.

[0036] It should be noted that the load forecast value for the future period obtained in this embodiment includes the load value of each future time step and provides a confidence interval to help decision makers evaluate the reliability of the forecast as a basis for subsequent formulation of scheduling strategies.

[0037] In the process of optimizing the scheduling, the optimization problem is defined as minimizing the system operation cost while satisfying the power market demand and energy inventory constraints. The optimization problem can be expressed as:

[0038]

[0039]

[0040]

[0041]

[0042]

[0043]

[0044] in, Purchase power for the grid, is the charging and discharging efficiency of the energy storage system, is the load demand, and The unit cost of purchasing electricity for the grid and charging and discharging the energy storage system, and The maximum limit for power purchase from the grid and the power of the energy storage system, is the SOC of the energy storage system, is the charging efficiency of the energy storage system, is the time step, and are the maximum and minimum limits of the energy storage system SOC.

[0045] This embodiment uses an improved particle swarm algorithm as the scheduling optimization algorithm. The particle swarm algorithm simulates the group behavior of birds in a flock and finds the optimal solution by continuously adjusting the position of particles (i.e., potential solutions). It can more effectively adapt to the complex scenarios of energy storage scheduling optimization.

[0046] The improved algorithm is as follows: (1) Chaotic map initialization In the initialization stage of the particle swarm, Logistic mapping is used to generate the initial position and speed. The ergodic and random characteristics of Logistic mapping can effectively enhance the diversity of the particle swarm and prevent it from falling into the local optimum; that is, ;in, To control the parameters, after multiple rounds of experimental verification, the hyperparameter value was 3.82, and the chaos of the system was optimal.

[0047] The fluctuations in electricity demand and market prices are uncertain, so the diversity of the initial solutions of the particle swarm directly affects the ability to search for the global optimal energy storage scheduling strategy. Generating the initial particle positions through chaotic mapping can more effectively cover various possible scheduling strategies, thereby increasing the probability of finding the global optimal solution.

[0048] For example, in the energy storage dispatching system, the initial position represents the charging and discharging strategy of the energy storage equipment in different time periods in the future. The diverse initial schemes generated by the chaotic mapping ensure that the system can explore a variety of different dispatching strategies in the early stages, thereby adapting to the fluctuations in power demand and the uncertainty of market prices, and increasing the effectiveness and reliability of the dispatching strategy.

[0049] (2) Adding adaptive weights and dynamic parameter adjustment In the application scenario of energy storage dispatching system, the volatility of power load and the dynamic changes of market environment determine that the dispatching strategy needs to have flexible response capabilities. and learning factor , which is usually fixed and difficult to adapt to the complexity of the scheduling environment. The improved algorithm uses adaptive weights and dynamic parameter adjustment mechanisms to enable the algorithm to effectively perform adaptive adjustments in different search stages, thereby improving the efficiency and flexibility of scheduling optimization.

[0050] Inertia Weight Dynamic adjustments are made at different iteration stages to balance the weight between global search and local search: ; Use a larger inertia weight at the beginning of the search , to enhance the global search capability; gradually reduce the weight in the later stages of the search , to enhance the local development capability and help the particles converge to the optimal solution more finely.

[0051] Learning Factor It controls the degree to which the particle follows its own historical optimal position and the global optimal position. In the improved algorithm, the learning factor is adaptively adjusted according to the current state of the search to enhance the adaptability of the algorithm at different stages.

[0052] Initial Exploration Stage: Increasing Individual Learning Factors This can enhance the particle's pursuit of its own historical optimal position. This can encourage particles to independently explore their own optimal solutions, increase the coverage of the search space, and prevent the particle swarm from falling into the local optimum at an early stage.

[0053] Adjustment formula of individual learning factor: ; in, and are the maximum and minimum values ​​of the individual learning factor respectively.

[0054] Late convergence stage: increasing the group learning factor This can enhance the particles' pursuit of the global optimal position, help the particles converge quickly to the global optimal solution, and improve the convergence speed of the algorithm.

[0055] Adjustment formula of group learning factor: ; in, and are the minimum and maximum values ​​of the group learning factor, respectively.

[0056] Through such adaptive adjustments, the behavior of particles in different search stages can be more flexible, achieving a balance between exploration and development, thereby ensuring that in the energy storage scheduling scenario, strategies that maximize economic benefits can be formulated while improving the stability and accuracy of the search.

[0057] (3) Introducing fuzzy logic controller The purpose of introducing a fuzzy logic controller is to supplement the above-mentioned adaptive weights and learning factors. Compared with the above-mentioned adaptive weights and dynamic learning factors, which mainly depend on the number of iterations and particle states, the introduced fuzzy logic controller adjusts parameters by monitoring the dynamic characteristics of the particle swarm (mainly including the aggregation degree and convergence speed of the particle swarm) based on the global trend of the algorithm operation. This adjustment is based on real-time state feedback. For example, when it is detected that the particle swarm is over-concentrated, the inertia weight is increased or the learning factor is reduced to enhance the exploration ability and prevent it from falling into the local optimum.

[0058] When the particle cluster concentration is high, increase the inertia weight To enhance exploration; if the particle swarm converges slowly, increase the global learning factor , ensuring that the particle swarm converges to the global optimum as quickly as possible; in each iteration, the particle velocity and position updates can be expressed as: ; ; in, is the inertia weight, which is used to balance the influence of current speed and historical speed. and are learning factors, which control the degree to which the particle follows its own historical optimal position and the global optimal position. , , These are all values ​​that are further corrected by the fuzzy logic controller based on adaptive adjustment. is the particle speed, reflecting the change range of the scheduling strategy. For particles In time The position of represents the charging and discharging strategy of the energy storage system, that is, the power and time period of charging and discharging. and is a random number between 0 and 1. Its purpose is to introduce randomness and increase the particle swarm's ability to explore unknown space, that is, to increase the range of strategy choices. is the historical optimal position, which represents the best scheduling solution found by the particle. is the global optimal position, representing the optimal energy storage scheduling strategy found by the entire particle swarm.

[0059] This embodiment uses chaotic mapping to generate the speed and position of the initial particle swarm, sets the chaotic parameter r=3.82, and sets the initial position Randomly generate in the range of [0,1]; generate the position and speed of the initial particle through the Logistic mapping formula, indicating the possible scheduling scheme, including the time period and power of charging and discharging. Set the maximum and minimum values ​​of the inertia weight to , ; Iterate and update the speed and position of particles. Through continuous iteration, the particle swarm searches for the best energy storage scheduling strategy to minimize the system operation cost and maximize the economic benefits; output the global optimal position , that is, the globally optimal energy storage scheduling strategy, which clarifies the charging and discharging operations of the energy storage system in each time period.

[0060] This embodiment introduces a time series load forecasting method to achieve accurate, real-time and adaptive optimization of energy storage scheduling, accurately predict the user-side load and then use it as the basis for scheduling on the energy supply side, achieve accurate scheduling of the energy storage system, maximize power demand, and improve system operating efficiency; support the formulation of optimal functional strategies under different market conditions, so that energy utilization is more economical and sustainable, and the cost of operation is reduced. Through real-time monitoring and adjustment of scheduling, it can flexibly respond to fluctuations in electricity market prices; it has obvious advantages in improving the efficiency of the power system, reducing operating costs and promoting intelligence. By comprehensively considering users and the energy supply side, it introduces advanced technical means to provide innovative solutions for the future development of the power system.

[0061] Embodiment 2 Embodiment 2 of the present invention introduces an energy storage scheduling system based on time series load forecasting.

[0062] like Figure 3 The energy storage dispatching system shown in the figure based on time series load forecasting includes: An acquisition module, configured to acquire power system operation status information; A prediction module is configured to predict the operating load of the power system according to the acquired state information and a preset time series load prediction model to obtain an operating load prediction value; A dispatching module, which is configured to perform energy storage dispatch according to the obtained operating load forecast value, and complete energy storage dispatch based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet algorithm, determines the timestamp according to the time sequence in the acquired status information, determines the time series load influencing factors based on the timestamp, and determines the time series load forecasting model according to the determined time series load influencing factors.

[0063] The detailed steps are the same as those of the energy storage scheduling method based on time series load forecasting provided in Example 1, and will not be repeated here.

[0064] Embodiment 3 Embodiment 3 of the present invention provides a computer-readable storage medium.

[0065] A computer-readable storage medium stores a program thereon, which, when executed by a processor, implements the steps of a method for energy storage scheduling based on time series load forecasting as described in Embodiment 1 of the present invention.

[0066] The detailed steps are the same as those of the energy storage scheduling method based on time series load forecasting provided in Example 1, and will not be repeated here.

[0067] Embodiment 4 A fourth embodiment of the present invention provides an electronic device.

[0068] An electronic device comprises a memory, a processor and a program stored in the memory and running on the processor, wherein when the processor executes the program, the steps in a method for energy storage scheduling based on time series load forecasting as described in Embodiment 1 of the present invention are implemented.

[0069] The detailed steps are the same as those of the energy storage scheduling method based on time series load forecasting provided in Example 1, and will not be repeated here.

[0070] Embodiment 5 Embodiment 5 of the present invention provides a computer program product.

[0071] A computer program product includes software codes, wherein the program in the software codes executes the steps in the energy storage scheduling method based on time series load forecasting as described in the first embodiment of the present invention.

[0072] The detailed steps are the same as those of the energy storage scheduling method based on time series load forecasting provided in Example 1, and will not be repeated here.

[0073] The above description is only a preferred embodiment of the present embodiment and is not intended to limit the present embodiment. For those skilled in the art, the present embodiment may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present embodiment shall be included in the protection scope of the present embodiment.

Claims

1. A method for energy storage scheduling based on time series load forecasting, characterized in that: include: Obtaining power system operation status information; According to the acquired state information and a preset time series load forecasting model, the operating load of the power system is forecasted to obtain an operating load forecast value; Energy storage scheduling is performed according to the obtained operating load forecast value, completing energy storage scheduling based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet The algorithm determines a timestamp according to the time sequence in the acquired state information, determines a time series load influencing factor based on the timestamp, and determines a time series load prediction model according to the determined time series load influencing factor.

2. A method for energy storage scheduling based on time series load forecasting as claimed in claim 1, characterized in that: The time series load forecasting model is ;in, It is a time series load forecasting model; is a trend unit used to capture the nonlinear growth of time series load influencing factors. , , , and are all hyperparameters, is the nonlinear trend of the load, is the linear trend term used to capture the changing load demand; is a seasonal unit used to capture the periodicity of time series load influencing factors. , is the ordinal number of the seasonal component, is the total number of seasonal components, is the contribution of the ith seasonal component at time t, is the serial number of the periodic component related to the external factors, is the total number of cyclical components of external factors, is a periodic function related to external factors, is the influence weight of the periodic function; For holiday units, , is the total number of holidays, is the effect of the i-th holiday at time t; is the error term used to capture the random changes in the factors affecting the time series load, , is the error term at the previous moment, is the autoregressive parameter of the error term, is the white noise error.

3. A method for energy storage scheduling based on time series load forecasting as claimed in claim 2, characterized in that: The parameters of the time series load forecasting model are optimized by maximizing the log-likelihood estimation. In the process of parameter optimization, the parameters of the trend item, the seasonal unit and the holiday unit are adjusted by gradient descent, and the iteration is repeated until the prediction error of the time series load forecasting model is minimized.

4. The energy storage scheduling method based on time series load forecasting as claimed in claim 1, characterized in that: In the process of energy storage scheduling according to the obtained operating load forecast value, an improved particle swarm algorithm is used to simulate the group behavior of birds in a flock, and the particle position is continuously adjusted to find the optimal solution, thereby obtaining the optimal operating load forecast value, and completing the optimal scheduling of energy storage according to the obtained optimal operating load forecast value.

5. A method for energy storage scheduling based on time series load forecasting as claimed in claim 4, characterized in that: In the process of energy storage optimization scheduling, chaotic mapping is used to generate the speed and position of the initial particle swarm, the inertia weight is adjusted based on the adaptive weight and dynamic parameter adjustment mechanism, the speed and position of the particles are iteratively updated, and continuous iteration is carried out. Fuzzy logic control is used to prevent falling into the local optimum, and the optimal energy storage scheduling searched by the particle swarm is obtained, that is, the global optimal energy storage scheduling strategy, which determines the charging and discharging operations of the energy storage system in each time period.

6. The energy storage scheduling method based on time series load forecasting as claimed in claim 1, characterized in that: The acquired power system operation status information includes at least historical power consumption data, market power price information, meteorological data, holiday power consumption data, and major event power consumption data.

7. An energy storage dispatching system based on time series load forecasting, characterized in that: include: An acquisition module, configured to acquire power system operation status information; A prediction module is configured to predict the operating load of the power system according to the acquired state information and a preset time series load prediction model to obtain an operating load prediction value; A dispatching module, which is configured to perform energy storage dispatch according to the obtained operating load forecast value, and complete energy storage dispatch based on time series load forecast; Among them, the preset time series load forecasting model adopts the improved Prophet algorithm, determines the timestamp according to the time sequence in the acquired status information, determines the time series load influencing factors based on the timestamp, and determines the time series load forecasting model according to the determined time series load influencing factors.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of a method for energy storage scheduling based on time series load forecasting as described in any one of claims 1 to 6 are implemented.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the program, the steps of a method for energy storage scheduling based on time series load forecasting as described in any one of claims 1-6 are implemented.

10. A computer program product comprising software code, characterized in that The program in the software code executes the steps of a method for energy storage scheduling based on time series load forecasting as described in any one of claims 1-6.

Images