Joint demand forecasting method and device for multiple types of energy storage equipment

Abstract

The embodiment of the present invention relates to a method and device for joint demand forecasting of multiple types of energy storage devices, the method comprising the steps of: obtaining historical data related to multiple types of energy storage devices, new energy generation and power grid operation within the scope of the power system to be forecasted; classifying the demand of the multiple types of energy storage devices; performing time series decomposition according to the results of the demand classification and the role of each energy storage device in the power grid operation; establishing prediction models based on the historical data and the results of the time series decomposition, and using the prediction models to perform predictions to obtain prediction results. The technical solution provided by the embodiment of the present invention classifies the energy storage devices in combination with the differences in the demands of multiple types of energy storage for the operation of the power system, and then establishes prediction models respectively, which overcomes the inaccurate prediction of single-type energy storage demand and improves the accuracy of demand forecasting for multiple types of energy storage devices.

Description

Technical Field

[0001] Embodiments of the present invention relate to the technical field of power system energy storage prediction, and in particular to a method and device for joint demand prediction of multiple types of energy storage devices. Background Art

[0002] With the rapid development of renewable energy and the growth of electricity demand, the power system is facing more and more challenges. The volatility and uncontrollability of renewable energy have brought certain difficulties to the stable operation of the power system. In order to solve this problem, energy storage technology has become an important means to improve the flexibility of the power system and resist power fluctuations. Different types of energy storage technologies play different roles in the power system, including power balance, peak load and backup capacity, primary and secondary frequency regulation, and safe and stable control. Therefore, studying the growth demand for multi-type energy storage capacity in the power system is of great significance to promoting the sustainable development and reliable operation of the power system.

[0003] At present, various types of energy storage technologies have been applied in power systems, including lithium-ion batteries, pumped storage, compressed air energy storage and supercapacitors. These energy storage technologies play different roles in different scenarios to meet the diverse needs of power systems. The factors affecting the growth of demand for multi-type energy storage capacity in power systems mainly include the development of renewable energy, the growth of electricity demand, energy policies and economic environment. The development of renewable energy will drive the growth of demand for energy storage technology, especially the application of energy storage technology in regulating the volatility of renewable energy. The growth of electricity demand will put higher requirements on the energy storage capacity demand of energy storage technology. Changes in energy policies and economic environment will also have an impact on the development and application of energy storage technology.

[0004] In order to predict the growth demand for various types of energy storage capacity in power systems, researchers have adopted a variety of methods and models. Common methods include trend analysis based on historical data, demand analysis based on load forecasting, and economic analysis based on the power market. Among them, the trend analysis method predicts the growth direction of future energy storage capacity demand by analyzing historical data and development trends. The demand analysis method based on load forecasting predicts the future demand for energy storage capacity by predicting load demand and combining the application scenarios of energy storage technology. In addition, the economic analysis method based on the power market takes into account the economics and market demand of energy storage technology to predict the scale of energy storage capacity demand. With the advancement of data acquisition and processing technology, future research will tend to data-driven prediction methods. Through big data analysis and artificial intelligence technology, a large amount of data in the power system can be mined and analyzed to improve the prediction precision and accuracy.

[0005] However, current research on energy storage demand forecasting problems mostly focuses on a single type of energy storage technology, such as pumped storage or electrochemical energy storage such as lithium batteries, and there is little research on the combined demand forecasting of multiple types of energy storage. Summary of the invention

[0006] Based on the above situation of the prior art, the purpose of the embodiments of the present invention is to provide a method and device for joint demand prediction of multiple types of energy storage equipment, combining the differences in power and energy characteristics of multiple types of energy storage equipment, and combining the needs for power system stability control, frequency regulation, peak load standby and comprehensive energy balance optimization, to obtain the total demand prediction results of multiple types of energy storage, thereby improving the accuracy of demand prediction of multiple types of energy storage equipment.

[0007] To achieve the above object, according to a first aspect of the present invention, a method for joint demand forecasting of multiple types of energy storage devices is provided, comprising the steps of:

[0008] S1. Acquire historical data related to various types of energy storage equipment, renewable energy power generation and grid operation within the power system to be predicted, wherein the renewable energy power generation includes at least wind power generation and photovoltaic power generation;

[0009] S2. Classifying the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices during energy storage charging and discharging;

[0010] S3, performing time series decomposition according to the results of the demand classification and the role of each energy storage device in the operation of the power grid;

[0011] S4, establishing prediction models based on the historical data and the results of time series decomposition, and using the prediction models to perform predictions to obtain prediction results;

[0012] S5, correcting the prediction result of the prediction model to obtain a corrected result;

[0013] S6. Perform a second correction based on the first correction result to obtain the final prediction result.

[0014] Furthermore, the multiple types of energy storage devices include multiple types of supercapacitors, flywheel energy storage devices, lithium-ion batteries, fuel cell hydrogen storage devices and pumped storage devices.

[0015] Furthermore, the historical data includes: the annual cumulative installed capacity of the various types of energy storage equipment, the annual cumulative installed capacity of wind power generation, the annual cumulative installed capacity of photovoltaic power generation, and the proportion of new energy power generation in the total installed capacity of power generation.

[0016] Furthermore, the step S2 includes classifying the requirements of the multiple types of energy storage devices according to the following formula:

[0017]

[0018] Among them, Class represents the demand classification of energy storage equipment. The value of 1 is the first type, indicating the stable control type; the value of 2 is the second type, indicating the primary and secondary frequency regulation type; the value of 3 is the third type, indicating the peak load standby type; the value of 4 is the fourth type, indicating the comprehensive energy balance optimization type; t response Indicates the response speed of the energy storage device; ΔT disc harge Indicates the maximum time that the energy storage device can continuously discharge at rated power; t1, t2, t3 and ΔT min They represent the first speed threshold, the second speed threshold, the third speed threshold and the time threshold respectively.

[0019] Furthermore, in step S3, time series decomposition is performed according to the result of the demand classification and the role of each energy storage device in the operation of the power grid, including: performing time series decomposition according to the following formula:

[0020]

[0021] Y t =T t +S t +I t

[0022] Where y represents the installed capacity of the corresponding energy storage device; the subscript t represents the current time; Y t is the time series of the original total energy storage capacity; T t represents the peak load demand component of the power grid, S t represents the power grid security and stability demand component, I t represents the comprehensive energy balance optimization component, with and Y t Same dimension; T t Contains the third type of energy storage device; S t comprising a first type and a second type of energy storage device; I t Includes requirements for a fourth type of energy storage device.

[0023] Furthermore, prediction models are established based on the results of the time series decomposition, including:

[0024] T t =f1(T t-1 ,T t-2 ,L,T t-k ,Wind t-1 ,Wind t-2 ,L,Wind t-k ,Solar t-1 ,Solar t-2,L,Solar t-k ) (3)

[0025] S t =f2(S t-1 ,S t-2 ,L,S t-k ,Proportion t-1 ) (4)

[0026] I t =f3(I t-1 ,I t-2 ,L,I t-k ) (5)

[0027] The subscript k indicates the length of the historical data used; Wind t-1 Represents the cumulative installed capacity of wind power generation at time t-1 in the past; Solar t-1 Represents the cumulative installed capacity of photovoltaic power generation at time t-1 in the past; Proportion t―1 Indicates the proportion of renewable energy power generation in total installed power generation capacity.

[0028] Furthermore, the method further comprises:

[0029] The feature fusion of historical data is performed according to the following formula:

[0030]

[0031] X S =[S t-1 S t-2 LS t-k |Proportion t-1 ]

[0032] X I =[I t-1 I t-2 LI t-k ]

[0033] Among them, X T represents the input vector after the characteristics of the power grid peak demand component prediction model are fused; X S represents the input vector after the characteristics of the power grid security and stability demand component prediction model are fused; X I It represents the input vector after the features of the comprehensive energy balance optimization component prediction model are fused;

[0034] The input vector is used as an input of a prediction model for prediction.

[0035] Further, the prediction result of the prediction model is corrected once to obtain a corrected result, including correcting once using the following formula:

[0036]

[0037] The result of the first correction is:

[0038]

[0039] Where, ΔT t-1 It represents the increment of installed capacity of pumped storage equipment at time t-1 compared with the previous time; ΔS t-1 Represents the increment of installed capacity of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment at time t-1 compared with the previous time; η t-1,Ｔ represents the proportion of pumped storage equipment in the total energy storage at time t-1; η t-1,S It indicates the proportion of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage devices in the total energy storage at time t-1.

[0040] Furthermore, a secondary correction is performed based on the primary correction result to obtain the final prediction result, including the secondary correction using the following formula:

[0041]

[0042] The final prediction result is:

[0043] Y t =T t ″+S t ″+I t ″

[0044] Among them, λ t,S Represents the first correction coefficient for the predicted values ​​of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment; λ t,I Represents the second correction coefficient for the predicted values ​​of other thermal storage and cold storage equipment.

[0045] According to another aspect of the present invention, a joint demand forecasting device for multiple types of energy storage devices is provided, comprising:

[0046] A data acquisition module, used to acquire historical data related to various types of energy storage devices, renewable energy power generation and power grid operation within the scope of the power system to be predicted, wherein the renewable energy power generation includes at least wind power generation and photovoltaic power generation;

[0047] A classification module, used to classify the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices to be predicted during the energy storage charging and discharging process;

[0048] A time series decomposition module, used to establish prediction models based on the historical data and the results of the time series decomposition, and use the prediction models to perform predictions to obtain prediction results;

[0049] A model building module, used to build prediction models based on the original data and the results of time series decomposition;

[0050] A primary correction module, used to perform a primary correction on the prediction result of the prediction model to obtain a primary correction result;

[0051] The secondary correction module is used to perform secondary correction based on the primary correction result to obtain the final prediction result.

[0052] In summary, an embodiment of the present invention provides a method and device for joint demand forecasting of multiple types of energy storage devices, the method comprising the steps of: obtaining historical data related to multiple types of energy storage devices, new energy generation and power grid operation within the scope of the power system to be forecasted; classifying the demand of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices during energy storage charging and discharging; performing time series decomposition according to the results of the demand classification and the role of each energy storage device in the power grid operation; establishing prediction models based on the historical data and the results of the time series decomposition, and using the prediction models to predict and obtain prediction results; making a correction to the prediction results of the prediction model to obtain a primary correction result; making a secondary correction based on the primary correction result to obtain a final prediction result. The technical solution provided by the embodiment of the present invention classifies the energy storage devices in combination with the differences in the demands of multiple types of energy storage in the operation of the power system, and then establishes prediction models respectively, which overcomes the inaccurate prediction of the demand for single-type energy storage and improves the accuracy of the demand prediction of multiple types of energy storage devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A flowchart of a method for joint demand forecasting of multiple types of energy storage provided in an embodiment of the present invention.

[0054] Figure 2 It is a schematic diagram of model prediction results of different components using the component prediction method provided by an embodiment of the present invention.

[0055] Figure 3 To compare the results of energy storage equipment demand forecasting using different methods. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments and with reference to the accompanying drawings. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concept of the present invention.

[0057] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in one or more embodiments of the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in one or more embodiments of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.

[0058] The technical solution of the present invention is described in detail below with reference to the accompanying drawings. An embodiment of the present invention provides a method for forecasting joint demand of multiple types of energy storage devices. Figure 1 A flow chart of the method is shown in Figure 1 As shown, the method comprises the following steps:

[0059] S1. Obtain historical data related to various types of energy storage equipment, new energy power generation and power grid operation within the scope of the power system to be predicted. The power system to be predicted includes multiple types of energy storage equipment, new energy power generation equipment and other traditional types of power generation equipment including thermal power generation equipment. Among them, the various types of energy storage equipment include, for example, supercapacitors, flywheel energy storage equipment, lithium-ion batteries, fuel cell hydrogen storage equipment and pumped storage equipment; the new energy power generation includes at least wind power generation and photovoltaic power generation, and may also include other types of new energy power generation. Historical data refers to historical data related to new energy power generation, load, energy storage equipment and power grid operation requirements, including but not limited to the cumulative installed capacity of various types of energy storage equipment year by year, the cumulative installed capacity of wind power generation year by year, the cumulative installed capacity of photovoltaic power generation year by year and the proportion of new energy power generation in the total installed power generation capacity. In this step, historical data related to various types of energy storage equipment, new energy generation and power grid operation within the scope of the power system to be predicted are obtained. For example, the energy storage installed capacity data of China from 2014 to 2022 is selected as the data of this embodiment, and the year-on-year change data of its total installed capacity and the proportion of installed capacity of different types of energy storage equipment in each year are collected. In addition, wind power installed capacity, photovoltaic installed capacity and new energy installed capacity can also be selected as auxiliary data for model establishment. The data is acquired through the file reading interface, and the data set is divided, the data from 2014 to 2020 is used as training data, and the data from 2021 and 2022 is used as prediction verification data.

[0060] S2. Classify the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices during the energy storage charging and discharging process. In this step, the requirements of the multiple types of energy storage devices can be classified according to the following formula:

[0061]

[0062] Among them, Class represents the demand classification of energy storage equipment. The value of 1 is the first type, indicating the stable control type; the value of 2 is the second type, indicating the primary and secondary frequency regulation type; the value of 3 is the third type, indicating the peak load standby type; the value of 4 is the fourth type, indicating the comprehensive energy balance optimization type; t response Indicates the response speed of the energy storage device; ΔT disc harg e Indicates the maximum time that the energy storage device can continuously discharge at rated power; t1, t2, t3 and ΔT min They represent the first speed threshold, the second speed threshold, the third speed threshold and the time threshold respectively.

[0063] In this embodiment, according to the response speed t of different types of energy storage devices response and continuous energy storage and release time ΔT dischargeEnergy storage elements are divided into four categories. For example, when energy storage devices include supercapacitor energy storage devices, flywheel energy storage devices, pumped storage and compressed air energy storage devices, electrochemical energy storage devices such as lithium-ion batteries and lead-acid batteries, and molten salt thermal energy storage devices, the classification results are: supercapacitor and flywheel energy storage devices are the first type; pumped storage and compressed air energy storage devices are the second type; electrochemical energy storage devices such as lithium-ion batteries and lead-acid batteries are the third type; molten salt thermal energy storage devices are the fourth type.

[0064] S3, according to the result of the demand classification and the role of each energy storage device in the operation of the power grid, the time series is decomposed. In this step, the time series is decomposed according to the following formula:

[0065]

[0066] Y t =T t +S t +I t

[0067] Where y represents the installed capacity of the corresponding energy storage device; the subscript t represents the current time; Y t is the time series of the original total energy storage capacity; T t represents the peak load demand component of the power grid, S t represents the power grid security and stability demand component, I t represents the comprehensive energy balance optimization component, with and Y t Same dimension; T t Contains the third type of energy storage device; S t comprising a first type and a second type of energy storage device; I t Includes requirements for a fourth type of energy storage device.

[0068] According to the role of energy storage equipment in the power system, the time series of total energy storage installed capacity is decomposed into three different components: T, S, and I. Among them, the power grid peak demand component T includes the second type of energy storage equipment, the power grid safety and stability demand component S includes the first type and geothermal type of energy storage equipment, and the comprehensive energy balance optimization component I includes the fourth type of energy storage equipment. The division of energy storage equipment types in this embodiment of the present invention takes into account the time scale of safe and stable control during the operation of the power grid as well as frequency modulation and peak regulation. Considering that the time scale of emergency control of the second line of defense in the three lines of defense for safe and stable operation of the power grid is 300ms, when dividing, the first type of energy storage equipment can meet this requirement in terms of response speed. The time scale of frequency modulation is at the second and minute levels, and the second type of energy storage equipment can meet the demand. These two are divided into the power grid safety and stability demand component S. In the subsequent modeling process, it is possible to model S using relevant data that affects the power angle and frequency in the operation of the power grid (such as the proportion of new energy installed capacity), which is equivalent to constraining the model from the perspective of power grid operation, so that the prediction results are better. Similarly, T is the energy storage equipment divided according to the peak demand, and I is the equipment for the remaining part.

[0069] S4. Based on the historical data and the results of the time series decomposition, a prediction model is established respectively, and the prediction result is obtained by using the prediction model. In the embodiment of the present invention, the model is established respectively according to the results of the above time series decomposition:

[0070] T t =f1(T t-1 ,T t-2 ,L,T t-k ,Wind t-1 ,Wind t-2 ,L,Wind t-k ,Solar t-1 ,Solar t-2 ,L,Solar t-k )

[0071] S t =f2(S t-1 ,S t-2 ,L,S t-k ,Proportion t-1 )

[0072] I t =f3(I t-1 ,I t-2 ,L,I t-k )

[0073] The subscript k indicates the length of the historical data used; Wind t-1Represents the cumulative installed capacity of wind power generation at time t-1 in the past; Solar t-1 Represents the cumulative installed capacity of photovoltaic power generation at time t-1 in the past; Proportion t―1 Indicates the proportion of renewable energy power generation in total installed power generation capacity.

[0074] Using the time series of different components decomposed in the above steps, the RNN recurrent neural network is used to build a prediction model. The prediction model of component T is modeled in combination with the annual installed capacity data of wind power and photovoltaic power; the prediction model of component S is modeled in combination with the proportion of new energy installed capacity; the prediction model of component I is modeled using only a single time series. The above components are linearly added to obtain the prediction result of the total installed capacity of energy storage.

[0075] Figure 2 FIG. 4 is a schematic diagram showing the model prediction results of different components using the component prediction method provided by an embodiment of the present invention, such as Figure 2 As shown in the figure, after adopting component prediction, the high-dimensional model established by combining the different roles played by energy storage equipment in the operation of the power grid, the power grid peak demand component T, the power grid security and stability demand component S and the comprehensive energy balance optimization component I can more accurately obtain the trend of energy storage installed capacity changing over time. Among them, the prediction results of the power grid security and stability demand component S model are close to the true value; the prediction results of the power grid peak demand component T and the comprehensive energy balance optimization component I model are also in line with the actual trend.

[0076] According to some optional embodiments, the method further includes: performing feature fusion on the historical data according to the following formula:

[0077]

[0078] X S =[S t-1 S t-2 LS t-k |Proportion t-1 ]

[0079] X I =[I t-1 I t-2 LI t-k ]

[0080] Among them, X T represents the input vector after the characteristics of the power grid peak demand component prediction model are fused; X S represents the input vector after the characteristics of the power grid security and stability demand component prediction model are fused; X I Represents the input vector after feature fusion of the comprehensive energy balance optimization component prediction model.

[0081] Through the above feature fusion, the input vector corresponding to each component prediction model is obtained, and the input vector is used as the input of the prediction model for prediction, thereby obtaining the prediction result. Through the above feature fusion, compared with modeling based only on the time series value of the value to be predicted, more relevant information is integrated, so that the state space dimension of the established model is higher and more in line with the actual situation. T The feature fusion process enables the model in the subsequent modeling process to take into account the characteristics of the time series of the energy storage device itself during the training process, and to take into account the impact of wind power and photovoltaic installed capacity on peak load demand during the operation of the power grid. S The process of feature fusion mainly considers the impact of the new energy content of the power grid on the safe and stable operation of the system and the frequency regulation requirements. I ,Since this component has little impact on the actual power grid operation, only its own time series is considered.

[0082] S5. Correct the prediction result of the prediction model once to obtain a corrected result. In the embodiment of the present invention, the following formula is used to correct the prediction result once:

[0083]

[0084] The result of the first correction is:

[0085]

[0086] Where, ΔT t-1 It represents the increment of installed capacity of pumped storage equipment at time t-1 compared with the previous time; ΔS t-1 Represents the increment of installed capacity of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment at time t-1 compared with the previous time; η t-1,T represents the proportion of pumped storage equipment in the total energy storage at time t-1; η t-1,S It indicates the proportion of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage devices in the total energy storage at time t-1.

[0087] The first correction mainly takes into account the friendliness of energy storage equipment to the power grid. Without considering the priority of resource scheduling, the energy storage equipment corresponding to the safe and stable operation of the power grid can also achieve the function of peak load regulation. Under ideal conditions, the proportion of the first and second types of equipment should be much higher than that of the third type of equipment. From this perspective, the results directly output by the model are corrected, and the correction is made according to the increments and proportions of the S and T components in previous years, and finally the correction result of the total installed capacity is obtained. The correction result is more in line with the development trend of the actual power grid in terms of the ratio of the T component and the S component, and the predicted value of the overall energy storage installed capacity is also closer to reality.

[0088] Figure 3 The comparison of the energy storage equipment demand forecasting results using different methods is shown in Figure 3 As shown in Figure 1, the prediction results of the model established by considering multiple types of energy storage are much more accurate than the prediction results of the model established by directly modeling the total installed capacity of energy storage. Figure 3 As shown in the figure, the absolute difference between the forecast result after one correction and the true value in 2021 is 0.81GW, which is a slight increase in error compared to the absolute difference of 0.61GW between the forecast result and the true value when only considering multiple types of energy storage, but the difference is not large; the absolute difference between the forecast result after one correction and the true value in 2022 is 2.9GW, which is a significant decrease in error compared to the absolute difference of 3.9GW between the forecast result and the true value when only considering multiple types of energy storage. In general, the forecast results are more accurate after one correction.

[0089] S6. Perform secondary correction based on the primary correction result to obtain the final prediction result. In the embodiment of the present invention, the secondary correction is performed using the following formula:

[0090]

[0091] The final prediction result is:

[0092] Y t =T t ″+S t ″+I t ″

[0093] Among them, λ t,S Represents the first correction coefficient for the predicted values ​​of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment; λ t,I The first correction coefficient and the second correction coefficient can be determined according to the following formula considering the influence of some macro factors:

[0094] λ t,S =1+Δe+Δc

[0095]

[0096]

[0097] λ t,I =1+Δp

[0098]

[0099] Among them, Δe represents the correction amount caused by economic factors; RGDP represents the economic growth rate of the previous year; Δc represents the correction amount caused by the change of energy storage cost; ΔRLCOS represents the decline rate of levelized electricity storage cost in the previous year compared with the previous year; Δp represents the correction amount caused by comprehensive energy optimization configuration planning; ΔPV represents the change amount of the planned amount of molten salt thermal storage in the previous year compared with the previous year. Through the secondary correction, the accuracy of the prediction results is further improved.

[0100] like Figure 3 As shown, the absolute difference between the forecast results after the second revision and the true value in 2021 is 0.19GW, and the absolute difference between the forecast results and the true value in 2022 is 1.11GW, which is an improvement in accuracy compared with other methods.

[0101] The actual values ​​of the total installed capacity of energy storage in 2021 and 2022, as well as the predicted values ​​based on direct modeling, consideration of multiple types of energy storage, primary correction, and secondary correction are shown in Table 1.

[0102] Table 1 True values ​​and predicted values ​​of each method

[0103]

[0104] In this embodiment, MAPE and RMSE are used as evaluation indicators of the prediction model:

[0105]

[0106] As shown in Table 2, the evaluation index results of different methods are as follows: RMSE obtained by direct modeling is 10.127GW, MAPE is 17.353%; considering multiple types of energy storage, RMSE is 2.729GW, MAPE is 3.844%; after one correction, RMSE is 2.135GW, MAPE is 3.311%; after two corrections, RMSE is 0.796GW, MAPE is 1.133%. After taking into account the role of different types of energy storage in the power grid, the safety and stability requirements of the power grid, the primary and secondary frequency regulation requirements, the peak regulation requirements, and the real-time policy impact, the prediction result finally obtained by the present invention has the highest accuracy, so that the energy storage installed capacity demand can be more accurately predicted.

[0107] Table 2 Comparison of evaluation indicators of different methods

[0108] Direct Modeling Consider multiple types of energy storage After one correction After the second correction RMSE / GW 10.127 2.729 2.135 0.796 MAPE / % 17.353 3.844 3.311 1.133

[0109] An embodiment of the present invention further provides a device for predicting joint demand of multiple types of energy storage devices, the device comprising:

[0110] A data acquisition module, used to acquire historical data related to various types of energy storage devices, renewable energy power generation and power grid operation within the scope of the power system to be predicted, wherein the renewable energy power generation includes at least wind power generation and photovoltaic power generation;

[0111] A classification module, used to classify the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices to be predicted during the energy storage charging and discharging process;

[0112] A time series decomposition module, used to establish prediction models based on the historical data and the results of the time series decomposition, and use the prediction models to perform predictions to obtain prediction results;

[0113] A model building module, used to build prediction models based on the original data and the results of time series decomposition;

[0114] A primary correction module, used to perform a primary correction on the prediction result of the prediction model to obtain a primary correction result;

[0115] The secondary correction module is used to perform secondary correction based on the primary correction result to obtain the final prediction result.

[0116] The functional implementation process of each module in the joint demand prediction device for multiple types of energy storage devices provided in this embodiment of the present invention is the same as the steps in the joint demand prediction method for multiple types of energy storage devices provided in the above embodiment of the present invention, and its repeated description will be omitted here.

[0117] In summary, the embodiments of the present invention relate to a method and device for joint demand forecasting of multiple types of energy storage devices, the method comprising the steps of: obtaining historical data related to multiple types of energy storage devices, new energy generation and power grid operation within the scope of the power system to be forecasted; classifying the demand of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices during energy storage charging and discharging; performing time series decomposition according to the results of the demand classification and the role of each energy storage device in the power grid operation; establishing prediction models based on the historical data and the results of the time series decomposition, and using the prediction models to predict and obtain prediction results; making a correction to the prediction results of the prediction model to obtain a primary correction result; making a secondary correction based on the primary correction result to obtain a final prediction result. The technical solution provided by the embodiments of the present invention classifies the energy storage devices in combination with the differences in the demands of multiple types of energy storage in the operation of the power system, and then establishes prediction models respectively, which overcomes the inaccurate prediction of the demand for single-type energy storage and improves the accuracy of the demand prediction of multiple types of energy storage devices.

[0118] It should be understood that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present invention (including claims) is limited to these examples; under the idea of ​​the present invention, the technical features in the above embodiments or different embodiments may also be combined, the steps may be implemented in any order, and there are many other changes in different aspects of one or more embodiments of the present invention as described above, which are not provided in detail for the sake of simplicity. The above specific embodiments of the present invention are only used to illustrate or explain the principles of the present invention, and do not constitute a limitation of the present invention. Therefore, any modifications, equivalent substitutions, improvements, etc. made without departing from the spirit and scope of the present invention should be included in the scope of protection of the present invention. In addition, the claims attached to the present invention are intended to cover all changes and modifications that fall within the scope and boundaries of the attached claims, or the equivalent forms of such scope and boundaries.

Claims

1. A method for joint demand forecasting of multiple types of energy storage equipment, characterized in that: Includes steps: S1. Acquire historical data related to various types of energy storage equipment, renewable energy power generation and grid operation within the power system to be predicted, wherein the renewable energy power generation includes at least wind power generation and photovoltaic power generation; S2. Classifying the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices during energy storage charging and discharging; S3, performing time series decomposition according to the results of the demand classification and the role of each energy storage device in the operation of the power grid; S4, establishing prediction models based on the historical data and the results of time series decomposition, and using the prediction models to perform predictions to obtain prediction results; S5, correcting the prediction result of the prediction model to obtain a corrected result; This includes a correction using the following formula: The result of the first correction is: Where, ΔT t-1 It represents the increment of installed capacity of pumped storage equipment at time t-1 compared with the previous time; ΔS t-1 Represents the increment of installed capacity of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment at time t-1 compared with the previous time; η t-1,T represents the proportion of pumped storage equipment in the total energy storage at time t-1; η t-1,S Indicates the proportion of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment in the total energy storage at time t-1; Class indicates the demand classification of energy storage equipment. The value of 1 is the first type, indicating the stable control type; the value of 2 is the second type, indicating the primary and secondary frequency modulation type; the value of 3 is the third type, indicating the peak load standby type; the value of 4 is the fourth type, indicating the comprehensive energy balance optimization type; T t represents the peak load demand component of the power grid, S t represents the power grid security and stability demand component, I t represents the comprehensive energy balance optimization component, with and Y t Same dimension; T t Contains the third type of energy storage device; S t comprising a first type and a second type of energy storage device; I t Includes a fourth type of energy storage device required; S6. Perform secondary correction based on the primary correction result to obtain the final prediction result; including performing secondary correction using the following formula: The final prediction result is: Y t ″=T t ″+S t ″+I t ″ Among them, λ t,S Represents the first correction coefficient for the predicted values ​​of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment; λ t,I Represents the second correction coefficient for the predicted values ​​of other thermal storage and cold storage equipment.

2. The method according to claim 1, characterized in that The multiple types of energy storage devices include multiple types of supercapacitors, flywheel energy storage devices, lithium-ion batteries, fuel cell hydrogen storage devices and pumped storage devices.

3. The method according to claim 1, characterized in that The historical data include: the annual cumulative installed capacity of the various types of energy storage equipment, the annual cumulative installed capacity of wind power generation, the annual cumulative installed capacity of photovoltaic power generation, and the proportion of new energy power generation in the total installed capacity of power generation.

4. The method according to any one of claims 1 to 3, characterized in that: The step S2 includes classifying the requirements of the multiple types of energy storage devices according to the following formula: Among them, t response Indicates the response speed of the energy storage device; ΔT discharge Indicates the maximum time that the energy storage device can continuously discharge at rated power; t1, t2, t3 and ΔT min They represent the first speed threshold, the second speed threshold, the third speed threshold and the time threshold respectively.

5. The method according to claim 4, characterized in that In the step S3, time series decomposition is performed according to the result of the demand classification and the role of each energy storage device in the operation of the power grid, including: performing time series decomposition according to the following formula: Y t =T t +S t +I t Where y represents the installed capacity of the corresponding energy storage device; the subscript t represents the current time; Y t is the time series of the original total energy storage capacity.

6. The method according to claim 5, characterized in that Based on the results of the time series decomposition, prediction models are established respectively, including: T t =f1(T t-1 ,T t-2 ,…,T t-k ,Wind t-1 ,Wind t-2 ,…,Wind t-k ,Solar t-1 ,Solar t-2 ,…,Solar t-k ) (3) S t =f2(S t-1 ,S t-2 ,…,S t-k ,Proportion t-1 ) (4) I t =f3(I t-1 ,I t-2 ,…,I t-k ) (5) The subscript k indicates the length of the historical data used; Wind t-1 Represents the cumulative installed capacity of wind power generation at time t-1 in the past; Solar t-1 Represents the cumulative installed capacity of photovoltaic power generation at time t-1 in the past; Proportion t-1 Indicates the proportion of renewable energy power generation in total installed power generation capacity.

7. The method according to claim 6, characterized in that The method further comprises: The feature fusion of historical data is performed according to the following formula: X S =[S t-1 S t-2 …S t-k |Proportion t-1 ] X I =[I t-1 I t-2 …I t-k ] Among them, X T represents the input vector after the characteristics of the power grid peak demand component prediction model are fused; X S represents the input vector after the characteristics of the power grid security and stability demand component prediction model are fused; X I Represents the input vector after the features of the comprehensive energy balance optimization component prediction model are fused; The input vector is used as an input of a prediction model for prediction.

8. A joint demand forecasting device for multiple types of energy storage equipment, characterized in that: include: A data acquisition module, used to acquire historical data related to various types of energy storage devices, renewable energy power generation and power grid operation within the scope of the power system to be predicted, wherein the renewable energy power generation includes at least wind power generation and photovoltaic power generation; A classification module, used to classify the requirements of the multiple types of energy storage devices according to the power, capacity and response speed parameters of the multiple types of energy storage devices to be predicted during the energy storage charging and discharging process; A time series decomposition module, used for performing time series decomposition according to the result of the demand classification and the role of each energy storage device in the operation of the power grid; A model building module, used to build prediction models based on the historical data and the results of time series decomposition, and use the prediction models to perform predictions to obtain prediction results; A primary correction module, used to perform a primary correction on the prediction result of the prediction model to obtain a primary correction result; This includes a correction using the following formula: The result of the first correction is: Where, ΔT t-1 It represents the increment of installed capacity of pumped storage equipment at time t-1 compared with the previous time; ΔS t-1 Represents the increment of installed capacity of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment at time t-1 compared with the previous time; η t-1,T represents the proportion of pumped storage equipment in the total energy storage at time t-1; η t-1,S Indicates the proportion of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment in the total energy storage at time t-1; Class indicates the demand classification of energy storage equipment. The value of 1 is the first type, indicating the stable control type; the value of 2 is the second type, indicating the primary and secondary frequency modulation type; the value of 3 is the third type, indicating the peak load standby type; the value of 4 is the fourth type, indicating the comprehensive energy balance optimization type; T t represents the peak load demand component of the power grid, S t represents the power grid security and stability demand component, I t represents the comprehensive energy balance optimization component, with and Y t Same dimension; T t Contains the third type of energy storage device; S t comprising a first type and a second type of energy storage device; I t Includes a fourth type of energy storage device required; The secondary correction module is used to perform secondary correction based on the primary correction result to obtain the final prediction result; including using the following formula for secondary correction: The final prediction result is: Y t ″=T t ″+S t ″+I t ″ Among them, λ t,S Represents the first correction coefficient for the predicted values ​​of electrochemical energy storage, flywheel energy storage and supercapacitor energy storage equipment; λ t,I Represents the second correction coefficient for the predicted values ​​of other thermal storage and cold storage equipment.

Images