Load adjustable capability evaluation method and device based on sequence decomposition, and storage medium

Abstract

The invention relates to a load adjustable capability evaluation method and device based on sequence decomposition and a storage medium, which are applied to the technical field of power grid load adjustment. Decomposing the load sequence into a temperature sensitive load and an electricity price sensitive load; secondly, on the basis of component load sequence shape features, considering temperature sensitive load feature factors such as meteorological factors such as environment temperature and the like and electricity price sensitive load feature factors such as time-of-use electricity price and the like, adopting an XGBoost time sequence algorithm prediction model, respectively constructing component load short-term prediction models, and performing component load short-term prediction; and finally, based on the short-term prediction value of the component load, obtaining the adjustable capability range of the temperature-sensitive load at different environment temperatures, and obtaining the adjustable capability range of the electricity price-sensitive load at different moments. Therefore, the accuracy of the evaluation result is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid load regulation, and in particular to a method, device and storage medium for evaluating load adjustable capacity based on sequence decomposition. Background Art

[0002] With the continuous development of society and the decreasing fossil energy, the development of renewable energy is the key to solving resource shortage and environmental problems. However, the proportion of new energy grid connection is gradually increasing, and its inherent randomness and intermittency seriously affect the adjustable capacity of the source side, and the difficulty of source-load matching is also increasing. Therefore, in order to adapt to the development of new power systems with a high proportion of new energy as the main body, it is very important to tap the adjustment capacity of massive resources on the load side, which is directly related to the level of new energy consumption in the power system and the safe and stable operation of the power system. At present, domestic and foreign scholars have carried out extensive research on the evaluation of load adjustable capacity, which is mainly divided into two types of methods: mechanism-driven and data-driven. The mechanism-driven method usually starts from the operating mechanism and adjustment characteristics of the flexible load, establishes an adjustment capacity evaluation model, and quantifies the adjustable potential of the flexible load. The data-driven method does not rely on the detailed mechanism model of the flexible load, but mainly uses machine learning and other technical algorithms to analyze the adjustment rules in its historical data. The evaluation technology based on the hybrid drive of mechanism and data aims to improve the prediction accuracy and response efficiency of the demand-side load adjustability in the power system, so as to better adapt to the dynamic needs and market changes of power grid operation.

[0003] Most of the existing load regulation capacity assessment technologies conduct relevant research from a static perspective, but ignore the internal coupling characteristics of the load, the internal mechanism information of the load, and the impact of different external factors at different time periods on the internal operating conditions of the load. For general loads with different characteristics, the overall assessment may not accurately reflect the adjustable potential of different load characteristics, resulting in a large deviation between the assessment results and the actual situation. Summary of the invention

[0004] In view of this, the purpose of the present invention is to provide a load adjustable capacity evaluation method, device and storage medium based on sequence decomposition, so as to solve the problem in the prior art that the internal coupling characteristics of the load, the internal mechanism information of the load, and the influence of different external factors at different time periods on the internal operating conditions of the load are ignored. For general loads with different characteristics, the overall evaluation may not accurately reflect the adjustable potential of different load characteristics, resulting in a large deviation between the evaluation results and the actual situation.

[0005] According to a first aspect of an embodiment of the present invention, a method for evaluating load adjustable capacity based on sequence decomposition is provided, the method comprising:

[0006] Acquire power grid load sequence data within a period of time, and preprocess the power grid load sequence data;

[0007] Decomposing the preprocessed power grid load sequence data based on a time series decomposition algorithm to obtain the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components;

[0008] The pre-built XGBoost time series algorithm model framework is trained by using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model.

[0009] The temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model are used to obtain temperature-sensitive load forecasting values ​​and electricity price-sensitive load forecasting values ​​at different times respectively;

[0010] Obtain the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the predicted value of the temperature-sensitive load at different times and the measured temperature at the corresponding times;

[0011] A prediction data set is constructed based on the electricity price sensitive load prediction values ​​at different times, and an initial center point of the prediction data set is determined based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times;

[0012] Acquire a new location center based on an initial center point of the prediction data set and an offset vector between the electricity price-sensitive load prediction value at different times and the initial center point;

[0013] The actual value corresponding to the predicted value of the electricity price sensitive load at different times is obtained, and the adjustable capacity range of the electricity price sensitive load at different times is obtained through the predicted value of the electricity price sensitive load at different times, the corresponding actual value and the new location center.

[0014] Preferably,

[0015] The preprocessing of the power grid load sequence data comprises:

[0016] Missing value filling, outlier correction and normalization processing;

[0017] The missing value filling includes: filling the missing values ​​using linear interpolation;

[0018] The outlier correction includes: a horizontal processing method and a vertical processing method;

[0019] The horizontal processing method includes: if the difference between the load value at any moment and the load values ​​at the adjacent moments before and after exceeds the preset horizontal threshold, the load value at the moment is judged to be an abnormal value, and the average value of the load values ​​at the adjacent moments before and after the abnormal value is used as the corrected load value;

[0020] The vertical processing method includes: if the difference between the load value at any moment and the load value at the same moment exceeds a preset vertical threshold; then the load value at the moment is judged to be an abnormal value, and the average of the load values ​​at the same moment is used as the corrected load value;

[0021] The normalization process includes: performing normalization process on the power grid load sequence data after missing value filling and abnormal value correction by using a linear normalization method.

[0022] Preferably,

[0023] The pre-built XGBoost time series algorithm model framework is trained by respectively using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model, including:

[0024] Construct an XGBoost time series algorithm model framework, set the objective function and loss function of the XGBoost time series algorithm model framework, add a regularization term to the loss function, iteratively train the XGBoost time series algorithm model framework using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component respectively, accumulate the model after each iterative update with the model before the iterative update, update the predicted value of the model until the preset number of iterations is reached, and obtain a temperature-sensitive load forecasting model and an electricity price-sensitive load forecasting model.

[0025] Preferably,

[0026] The step of obtaining the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the predicted value of the temperature-sensitive load at different times and the measured temperature at the corresponding times includes:

[0027] Obtain the measured temperature corresponding to different moments, and obtain the basic electric load at the measured temperature corresponding to different moments; obtain the temperature-sensitive load power consumption at the measured temperature corresponding to different moments based on the temperature-sensitive load prediction value at different moments, the corresponding measured temperature, and the basic electric load at the measured temperature;

[0028] The upward adjustment value and the downward adjustment value of the temperature-sensitive load are obtained through the power consumption of the temperature-sensitive load at different measurement temperatures, and the adjustable capacity range of the temperature-sensitive load at different temperatures is obtained based on the upward adjustment value and the downward adjustment value of the temperature-sensitive load.

[0029] Preferably,

[0030] The acquiring of a new location center based on the initial center point of the prediction data set and the offset vector between the electricity price-sensitive load prediction value at different times and the initial center point comprises:

[0031] Obtaining offset vectors of electricity price-sensitive load forecast values ​​at different times and the initial center point respectively; obtaining weights of electricity price-sensitive load forecast values ​​at different times using a Gaussian kernel function according to the distance of the offset vectors;

[0032] The weights of the electricity price-sensitive load forecast values ​​at different times are weighted averaged to obtain a new location center.

[0033] Preferably,

[0034] The method of obtaining the adjustable capacity range of the electricity price sensitive load at different times by using the electricity price sensitive load forecast value at different times, the corresponding true value and the new location center includes:

[0035] Obtain the load fluctuation value at different times through the electricity price sensitive load forecast value at different times and the corresponding true value;

[0036] The deviation value between the predicted load curve at different times and the average power consumption curve of the power consumption mode is obtained through the power price sensitive load prediction value at different times and the new location center;

[0037] The adjustable capacity range of the electricity price sensitive load at different times is obtained based on the load fluctuation value at different times and the deviation value between the predicted load curve and the average power consumption curve of the power consumption mode.

[0038] According to a second aspect of an embodiment of the present invention, a load adjustable capacity evaluation device based on sequence decomposition is provided, the device comprising:

[0039] Data acquisition module: used to acquire power grid load sequence data within a period of time and pre-process the power grid load sequence data;

[0040] Decomposition module: used for decomposing the preprocessed power grid load sequence data based on the time series decomposition algorithm, obtaining the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components;

[0041] Model training acquisition module: used to train the pre-built XGBoost time series algorithm model framework through the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component respectively, to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model;

[0042] Prediction module: used to obtain temperature-sensitive load prediction values ​​and electricity price-sensitive load prediction values ​​at different times respectively through the temperature-sensitive load prediction model and the electricity price-sensitive load prediction model;

[0043] Temperature load adjustable capacity acquisition module: used to obtain the adjustable capacity range of temperature-sensitive load at different temperatures according to the relationship between the temperature-sensitive load prediction value at different times and the measured temperature at the corresponding times;

[0044] Initial center point acquisition module: used to construct a prediction data set based on the electricity price sensitive load prediction values ​​at different times, and determine the initial center point of the prediction data set based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times;

[0045] A location center acquisition module: used to acquire a new location center based on an initial center point of the prediction data set and an offset vector between the electricity price-sensitive load prediction value at different times and the initial center point;

[0046] Electricity price load adjustable capacity acquisition module: used to obtain the actual value corresponding to the electricity price sensitive load forecast value at different times, and obtain the electricity price sensitive load adjustable capacity range at different times through the electricity price sensitive load forecast value at different times, the corresponding actual value and the new location center.

[0047] According to a third aspect of an embodiment of the present invention, there is provided a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented.

[0048] The technical solution provided by the embodiments of the present invention may have the following beneficial effects:

[0049] This application uses a time series decomposition algorithm to decompose the load sequence into temperature-sensitive loads and electricity price-sensitive loads based on the load standardization data set obtained through data preprocessing and the different load variation characteristics; secondly, based on the shape characteristics of the component load sequence, temperature-sensitive load characteristic factors such as meteorological factors such as ambient temperature and electricity price-sensitive load characteristic factors such as time-of-use electricity prices are considered, and the XGBoost time series algorithm prediction model is adopted to respectively construct component load short-term prediction models to perform component load short-term predictions; finally, based on the short-term prediction values ​​of the component loads, the adjustable capacity range of the temperature-sensitive loads under different ambient temperatures is obtained, as well as the adjustable capacity range of the electricity price-sensitive loads at different times; this solves the problem in the prior art of ignoring the internal coupling characteristics of the load, the internal mechanism information of the load, and the influence of different external factors at different time periods on the internal operating conditions of the load, resulting in a large deviation between the evaluation results and the actual situation for general loads with different characteristics.

[0050] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0052] Figure 1 is a flow chart of a method for evaluating load adjustable capacity based on sequence decomposition according to an exemplary embodiment;

[0053] Figure 2 is a system schematic diagram of a load adjustable capacity evaluation device based on sequence decomposition according to another exemplary embodiment;

[0054] In the attached figure: 1-data acquisition module, 2-decomposition module, 3-model training acquisition module, 4-prediction module, 5-temperature load adjustable capacity acquisition module, 6-initial center point acquisition module, 7-location center acquisition module, 8-electricity price load adjustable capacity acquisition module. DETAILED DESCRIPTION

[0055] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of devices and methods consistent with some aspects of the present invention as detailed in the appended claims.

[0056] Embodiment 1

[0057] Figure 1 is a flow chart of a method for evaluating load adjustable capacity based on sequence decomposition according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0058] S1, obtaining power grid load sequence data within a period of time, and preprocessing the power grid load sequence data;

[0059] S2, decomposing the preprocessed power grid load series data based on a time series decomposition algorithm to obtain the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components;

[0060] S3, respectively training the pre-built XGBoost time series algorithm model framework through the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component to obtain a temperature-sensitive load forecasting model and an electricity price-sensitive load forecasting model;

[0061] S4, obtaining temperature-sensitive load forecast values ​​and electricity price-sensitive load forecast values ​​at different times respectively through the temperature-sensitive load forecast model and the electricity price-sensitive load forecast model;

[0062] S5, obtaining the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the predicted value of the temperature-sensitive load at different times and the measured temperature at the corresponding times;

[0063] S6, constructing a prediction data set based on the electricity price sensitive load prediction values ​​at different times, and determining an initial center point of the prediction data set based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times;

[0064] S7, obtaining a new location center based on the initial center point of the prediction data set and the offset vectors of the electricity price-sensitive load prediction values ​​at different times and the initial center point;

[0065] S8, obtaining the actual value corresponding to the predicted value of the electricity price sensitive load at different times, and obtaining the adjustable capacity range of the electricity price sensitive load at different times through the predicted value of the electricity price sensitive load at different times, the corresponding actual value and the new location center;

[0066] It is understandable that obtaining the grid load series data within a period of time and preprocessing the obtained time series data include:

[0067] Missing value filling:

[0068] For missing data, methods such as mean filling and linear interpolation can be used. The load data is sorted in time series and the intervals are not large, so the linear interpolation method is used. The formula of the linear interpolation method is as follows:

[0069]

[0070] In the formula, y m is the load value at the mth moment; y m+1 is the load value at the m+1th moment; y m+k is the missing load value at time m+k;

[0071] Outlier correction:

[0072] (1) Horizontal processing method

[0073] If the difference between the current load value and the adjacent load value in the load data set exceeds the preset value, it can be determined that the data is abnormal and the horizontal processing method can be used to handle the abnormality. The processing method is as follows:

[0074]

[0075] In the formula, is the threshold set by the horizontal processing method, y(t) is the load value at the tth moment, and y(t) is the corrected load value.

[0076] (2) Vertical processing method

[0077] Since load data has a strong periodicity, if the data at a certain moment differs too much from the data at the same moment and exceeds the set threshold, the value will be marked as abnormal data. The abnormal value will be corrected according to the mean value of the load data at the same moment. The processing method is as follows:

[0078]

[0079]

[0080] Where β is the threshold set by the vertical processing method, y(t) is the load value at the tth moment, is the average load value of the same type of time in the load data, and y(t) is the corrected load value.

[0081] Normalization:

[0082] In order to ensure that different variables such as load and weather factors do not cause gradient deviation due to different magnitudes, it is necessary to normalize the historical data to obtain the same weight tendency and map different variables to the value range [0,1]. Therefore, this embodiment uses linear normalization to normalize the input variables. The normalization formula is as follows:

[0083]

[0084] In the formula, is the data after load normalization; x is the initial data of load; is the average value of the initial load data; x max is the maximum value of the initial load data; x min is the minimum value of the initial load data;

[0085] Load breakdown:

[0086] Time series decomposition is the process of decomposing a time series into several components, which usually include trend, seasonality and random components. The core idea is to decompose the time series into components at multiple scales, each scale corresponding to a specific frequency range in the time series, so as to allow the algorithm to capture the changing characteristics at different time scales. This embodiment is based on the time series decomposition algorithm to decompose the load series into trend, multiple periodic components and residual components, where the multiple periodic components include temperature-sensitive loads and electricity price-sensitive loads. The specific decomposition formula is as follows:

[0087] Y t =Y t trend +Y t cycle +Y t remainder

[0088] Where Y t is the general load data at time t; Y t trend is the trend load component at time t, Y t cycle is the periodic component at time t, Y t remainder is the residual component at time t;

[0089] Y t cycle =Y t temp +Y t price

[0090] Where Y t temp is the temperature sensitive periodic load component at time t, Y t price is the electricity price sensitive periodic load component at time t;

[0091] Model construction:

[0092] Based on the load standardized data set and load sequence shape characteristics of data preprocessing, temperature-sensitive load influencing factors such as ambient temperature and thermodynamic equivalent model parameters, as well as price-sensitive load influencing factors such as time-of-use electricity price, electricity consumption and peak-to-valley difference rate are considered. The date types of weekdays and holidays are distinguished, and the XGBoost time series algorithm prediction model is used to set 15 minutes as a prediction period to construct short-term prediction models for component loads respectively.

[0093] First, the integration method of serial result tree models is used to build the basic decision tree model. The formula is as follows:

[0094]

[0095] In the formula, x i is the input value of the i-th sample, y i (0) is the initial value of the model, y i (1) is the prediction result value of the i-th sample of the first model, y i (2) is the prediction result value of the second model for the i-th sample, f1(x i ) is the new decision tree of the first model iteration, f2(xi ) is the new decision tree for the second model iteration.

[0096] Secondly, based on the gradient boosting decision tree, a new decision tree objective function is constructed by adding a regularization term to the loss function:

[0097]

[0098] In the formula, F obj is the objective function of the decision tree; η(ξ) is the loss function of the strong learner; Ω(f(x k )) is the complexity of the i corresponding weak learners in the strong learner; λ and γ are hyperparameters; T is the total number of leaf nodes in the set decision tree; ω is the feature vector composed of the predicted values ​​output by all leaf nodes through the decision tree model; finally, with the optimal loss function as the goal, multiple prediction iterations are performed, the newly constructed decision tree is accumulated with the previous model, and the predicted value of the model is updated until the preset number of iterations is reached, and the temperature-sensitive load and electricity price sensitive load forecast

[0099]

[0100] In the formula, y i (j) is the predicted value of the i-th sample after j model iterations, y i (j-1) is the predicted value of the i-th sample after j-1 model iterations, f j (x i ) is the new decision tree of the jth model iteration;

[0101] According to the temperature-sensitive load and production-characteristic-sensitive load decomposed according to the load change characteristics, the temperature-sensitive load generally presents a single-peak characteristic, and the electricity price-sensitive load presents a step-like characteristic. The specific evaluation methods for the regulation capacity of loads with different characteristics are as follows:

[0102] Temperature sensitive load regulation capability range assessment:

[0103] Considering the characteristic relationship between temperature-sensitive load and temperature, the hourly load-temperature sensitivity curve is fitted, the time-sharing basic load is calculated, and the temperature-sensitive load component in the general load curve is identified. Among them, the temperature-sensitive load is stripped, and the curve between load and temperature is fitted in different time periods. The fitting method is as follows:

[0104]

[0105] In the formula, is the predicted value of the electric load data at time t, which is obtained through the above model prediction; T tis the temperature measured at time t; a, b, c are fitting coefficients, representing the coefficients of the quadratic term, the linear term, and the constant term, respectively; L' t (T) is the basic electric load at the temperature T℃ at time t, and the load value when the temperature derivative of the district electric load is 0 is taken; It is the temperature sensitive load at the temperature T℃ at time t;

[0106] The above formula is used to obtain L at temperatures T, T-1 and T+1 respectively;

[0107] Next, the mapping curve relationship between different outdoor temperatures and temperature-sensitive loads is established. Then, the mapping relationship between the user's acceptable indoor temperature adjustment range and the air conditioning load change is identified through mechanism modeling, and the adjustable value of the temperature-sensitive load within the appropriate temperature range is calculated. The formula is:

[0108] ΔL up =L(T+1)-L(T)

[0109] ΔL down =L(T)-L(T-1)

[0110] Where, ΔL up , ΔL down are the upward and downward adjustment values ​​of the temperature-sensitive load, L(T-1), L(T), and L(T+1) are the power consumption of the temperature-sensitive load corresponding to the temperatures of T-1, T, and T+1, respectively. That is to say, the adjustment capacity range of the temperature-sensitive load is (ΔL up , ΔL down );

[0111] Evaluation of the scope of regulation capability of electricity price sensitive loads:

[0112] According to the forecast value of price-sensitive load, considering the scale and behavior of price-sensitive load, the mean shift clustering analysis algorithm is used to classify the power consumption mode of price-sensitive load. The initial center point of the forecast data set is determined based on the distribution characteristics of the forecast values ​​at different times, and the offset vector between the data in the data set and is calculated:

[0113]

[0114] In the formula, Δy i is the offset vector of the i-th electricity price sensitive load, is the predicted value of the i-th electricity price sensitive load at time t, is the initial center point in the prediction data set;

[0115] According to the distance of the offset vector, the weight of each data point is calculated, and the Gaussian kernel function K(x) is used to measure the distance. The closer the point is to the center of the window, the greater the weight:

[0116]

[0117] In the formula, K(Δy i ) is the weight of the i-th electricity price sensitive load, d is the data dimension, and h represents the bandwidth parameter value;

[0118] Calculate the new center position based on the weighted average of the data points:

[0119]

[0120] In the formula, is the new center point of the price-sensitive load at time t, K(Δy i ) is the weight of the i-th electricity price sensitive load, is the predicted value of the i-th electricity price-sensitive load at time t, and m is the number of data points in the window;

[0121] According to the electricity consumption pattern of price-sensitive load obtained by cluster analysis, the load fluctuation value Δd1 based on the historical load curve in different time periods is obtained, the deviation value Δd2 between the predicted load curve in different time periods and the average electricity consumption curve of the electricity consumption pattern is obtained, and the adjustable range (Δd1, Δd2) of the price-sensitive load in different time periods is obtained.

[0122]

[0123] In the formula, y t is the true value of the electricity price sensitive load at time t.

[0124] Embodiment 2

[0125] Figure 2 is a system schematic diagram of a load adjustable capacity evaluation device based on sequence decomposition according to another exemplary embodiment, the device comprising:

[0126] Data acquisition module 1: used to acquire power grid load sequence data within a period of time and pre-process the power grid load sequence data;

[0127] Decomposition module 2: used for decomposing the preprocessed power grid load sequence data based on the time series decomposition algorithm, obtaining the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components;

[0128] Model training acquisition module 3: used to train the pre-built XGBoost time series algorithm model framework through the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component respectively, to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model;

[0129] Prediction module 4: used to obtain temperature-sensitive load prediction values ​​and electricity price-sensitive load prediction values ​​at different times respectively through the temperature-sensitive load prediction model and the electricity price-sensitive load prediction model;

[0130] Temperature load adjustable capacity acquisition module 5: used to obtain the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the temperature-sensitive load prediction value at different times and the measured temperature at the corresponding times;

[0131] Initial center point acquisition module 6: used to construct a prediction data set based on the electricity price sensitive load prediction values ​​at different times, and determine the initial center point of the prediction data set based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times;

[0132] Position center acquisition module 7: used to acquire a new position center based on the initial center point of the prediction data set and the offset vector between the electricity price sensitive load prediction value at different times and the initial center point;

[0133] The electricity price load adjustable capacity acquisition module 8 is used to obtain the actual value corresponding to the electricity price sensitive load forecast value at different times, and obtain the electricity price sensitive load adjustable capacity range at different times through the electricity price sensitive load forecast value at different times, the corresponding actual value and the new location center.

[0134] Embodiment three:

[0135] This embodiment provides a storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a main controller, each step in the above method is implemented;

[0136] It is understandable that the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0137] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0138] It should be noted that, in the description of the present invention, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0139] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0140] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0141] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0142] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0143] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0144] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0145] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A load adjustable capacity evaluation method based on sequence decomposition, characterized in that: The method comprises: Acquire power grid load sequence data within a period of time, and preprocess the power grid load sequence data; Decomposing the preprocessed power grid load sequence data based on a time series decomposition algorithm to obtain the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components; The pre-built XGBoost time series algorithm model framework is trained by using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model. The temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model are used to obtain temperature-sensitive load forecasting values ​​and electricity price-sensitive load forecasting values ​​at different times respectively; Obtain the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the predicted value of the temperature-sensitive load at different times and the measured temperature at the corresponding times; A prediction data set is constructed based on the electricity price sensitive load prediction values ​​at different times, and an initial center point of the prediction data set is determined based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times; Acquire a new location center based on an initial center point of the prediction data set and an offset vector between the electricity price-sensitive load prediction value at different times and the initial center point; The actual value corresponding to the predicted value of the electricity price sensitive load at different times is obtained, and the adjustable capacity range of the electricity price sensitive load at different times is obtained through the predicted value of the electricity price sensitive load at different times, the corresponding actual value and the new location center.

2. The method according to claim 1, characterized in that The preprocessing of the power grid load sequence data comprises: Missing value filling, outlier correction and normalization processing; The missing value filling includes: filling the missing values ​​using linear interpolation; The outlier correction includes: a horizontal processing method and a vertical processing method; The horizontal processing method includes: if the difference between the load value at any moment and the load values ​​at the adjacent moments before and after exceeds the preset horizontal threshold, the load value at the moment is judged to be an abnormal value, and the average value of the load values ​​at the adjacent moments before and after the abnormal value is used as the corrected load value; The vertical processing method includes: if the difference between the load value at any moment and the load value at the same moment exceeds a preset vertical threshold; then the load value at the moment is judged to be an abnormal value, and the average of the load values ​​at the same moment is used as the corrected load value; The normalization process includes: performing normalization process on the power grid load sequence data after missing value filling and abnormal value correction by using a linear normalization method.

3. The method according to claim 2, characterized in that The pre-built XGBoost time series algorithm model framework is trained by respectively using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model, including: Construct an XGBoost time series algorithm model framework, set the objective function and loss function of the XGBoost time series algorithm model framework, add a regularization term to the loss function, iteratively train the XGBoost time series algorithm model framework using the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component respectively, accumulate the model after each iterative update with the model before the iterative update, update the predicted value of the model until the preset number of iterations is reached, and obtain a temperature-sensitive load forecasting model and an electricity price-sensitive load forecasting model.

4. The method according to claim 3, characterized in that The step of obtaining the adjustable capacity range of the temperature-sensitive load at different temperatures according to the relationship between the predicted value of the temperature-sensitive load at different times and the measured temperature at the corresponding times includes: Obtain the measured temperature corresponding to different moments, and obtain the basic electric load at the measured temperature corresponding to different moments; obtain the temperature-sensitive load power consumption at the measured temperature corresponding to different moments based on the temperature-sensitive load prediction value at different moments, the corresponding measured temperature, and the basic electric load at the measured temperature; The upward adjustment value and the downward adjustment value of the temperature-sensitive load are obtained through the power consumption of the temperature-sensitive load at different measurement temperatures, and the adjustable capacity range of the temperature-sensitive load at different temperatures is obtained based on the upward adjustment value and the downward adjustment value of the temperature-sensitive load.

5. The method according to claim 4, characterized in that The acquiring of a new location center based on the initial center point of the prediction data set and the offset vector between the electricity price-sensitive load prediction value at different times and the initial center point comprises: Obtaining offset vectors of electricity price-sensitive load forecast values ​​at different times and the initial center point respectively; obtaining weights of electricity price-sensitive load forecast values ​​at different times using a Gaussian kernel function according to the distance of the offset vectors; The weights of the electricity price-sensitive load forecast values ​​at different times are weighted averaged to obtain a new location center.

6. The method according to claim 5, characterized in that The method of obtaining the adjustable capacity range of the electricity price sensitive load at different times by using the electricity price sensitive load forecast value at different times, the corresponding true value and the new location center includes: Obtain the load fluctuation value at different times through the electricity price sensitive load forecast value at different times and the corresponding true value; The deviation value between the predicted load curve at different times and the average power consumption curve of the power consumption mode is obtained through the power price sensitive load prediction value at different times and the new location center; The adjustable capacity range of the electricity price sensitive load at different times is obtained based on the load fluctuation value at different times and the deviation value between the predicted load curve and the average power consumption curve of the power consumption mode.

7. A load adjustable capacity evaluation device based on sequence decomposition, characterized in that: The device comprises: Data acquisition module: used to acquire power grid load sequence data within a period of time and pre-process the power grid load sequence data; Decomposition module: used for decomposing the preprocessed power grid load sequence data based on the time series decomposition algorithm, obtaining the decomposed periodic components, and decomposing the periodic components into temperature-sensitive periodic load components and electricity price-sensitive periodic load components; Model training acquisition module: used to train the pre-built XGBoost time series algorithm model framework through the temperature-sensitive periodic load component and the electricity price-sensitive periodic load component respectively, to obtain the temperature-sensitive load forecasting model and the electricity price-sensitive load forecasting model; Prediction module: used to obtain temperature-sensitive load prediction values ​​and electricity price-sensitive load prediction values ​​at different times respectively through the temperature-sensitive load prediction model and the electricity price-sensitive load prediction model; Temperature load adjustable capacity acquisition module: used to obtain the adjustable capacity range of temperature-sensitive load at different temperatures according to the relationship between the temperature-sensitive load prediction value at different times and the measured temperature at the corresponding times; Initial center point acquisition module: used to construct a prediction data set based on the electricity price sensitive load prediction values ​​at different times, and determine the initial center point of the prediction data set based on the distribution characteristics of the electricity price sensitive load prediction values ​​at different times; A location center acquisition module: used to acquire a new location center based on an initial center point of the prediction data set and an offset vector between the electricity price-sensitive load prediction value at different times and the initial center point; Electricity price load adjustable capacity acquisition module: used to obtain the actual value corresponding to the electricity price sensitive load forecast value at different times, and obtain the electricity price sensitive load adjustable capacity range at different times through the electricity price sensitive load forecast value at different times, the corresponding actual value and the new location center.

8. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the main controller, each step of the load adjustable capacity evaluation method based on sequence decomposition as described in any one of claims 1 to 6 is implemented.

Images