A summer cooling load peak forecasting method, system, device and storage medium

Abstract

The application discloses a summer cooling load peak prediction method, system, device and storage medium, comprising the following steps: calculating a reference load according to typical daily load data, and calculating a cooling load peak value according to the reference load; performing time sequence coupling on the cooling load peak value to obtain a cooling load peak time sequence prediction value; establishing a cooling load peak calculation model, inputting the cooling load peak value and the cooling load peak time sequence prediction value into the cooling load peak calculation model, and obtaining a time sequence of a load increment factor; inputting the time sequence into a load increment factor prediction model to obtain a load increment factor prediction value; and modifying the cooling load peak calculation model according to the load increment factor prediction value to obtain a cooling load peak prediction model. The application modifies the prediction result of the time sequence cooling load peak through the load increment factor, improves the accuracy of the prediction result, and thus helps to optimize the operation and planning of the power system.

Description

Technical Field

[0001] This invention relates to the field of power system operation planning technology, and in particular to a method, system, device and storage medium for predicting summer cooling load peaks. Background Technology

[0002] In recent years, deep-seated contradictions such as low power system operating efficiency and insufficient complementarity among various power sources have become increasingly prominent, urgently requiring comprehensive optimization. The new power system, primarily based on new energy sources, presents significant challenges to medium- and long-term operation simulation and planning due to the strong volatility and low rotational inertia brought about by the massive integration of new energy sources. Furthermore, with the integration of flexible DC transmission, large-scale offshore wind power, distributed photovoltaic grid connection, solar thermal power plants, and electric vehicles and charging stations into the power system, the dimensions of power system operation and planning issues will be significantly increased, providing new approaches to solving problems in building new power systems, tapping the flexible potential of the load side, and constructing demand response systems. Therefore, accurately predicting user load, establishing load characteristic models, and analyzing user load characteristics are crucial prerequisites for constructing demand response systems. This helps to promote the achievement of carbon peaking and carbon neutrality goals and is a key core link in optimizing the coordinated configuration of power sources, grids, loads, and storage, reducing investment waste, and lowering carbon emissions from the power system.

[0003] With the rapid development of my country's economy, summer cooling loads are constantly increasing, accounting for over 40% in some cities. Peak loads on the power grid are continuously breaking records, the peak-to-valley difference is widening, and load spikes caused by short-term summer cooling loads will create significant gaps in power supply. To better formulate reasonable and efficient demand response management measures for summer cooling loads and further optimize power system planning and operation, it is necessary to conduct predictive analysis of summer cooling loads. Traditional summer cooling load forecasting is based on historical data, using typical daily load curves from spring and autumn as load baselines to assess summer cooling loads. This method can preserve the temporal correlation of cooling loads to some extent, but in reality, summer cooling loads are not only related to daily meteorological conditions but also exhibit a significant cumulative effect. Existing methods cannot reflect the impact of this cumulative load effect on the forecast results, affecting their accuracy. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a method, system, device, and storage medium for predicting peak summer cooling loads. This invention proposes a load increment factor to characterize the cumulative effect of cooling loads. By using the load increment factor, the prediction results are amplified under the condition of continuous accumulation of load increments, thereby improving the accuracy of predicting user loads.

[0005] In a first aspect, embodiments of the present invention provide a method for predicting peak summer cooling loads, including:

[0006] Typical day load data are selected from historical load data based on the typical day selection criteria.

[0007] Based on the typical daily load data, the baseline load is calculated, and based on the baseline load, the peak value of the cooling load is calculated.

[0008] The peak cooling load was temporally coupled using a data fitting method to obtain the temporal prediction value of the peak cooling load.

[0009] Based on the load accumulation effect, a calculation model for cooling load peak based on load increment factor is established. The peak value of cooling load and the time series prediction value of cooling load peak are input into the calculation model to obtain the time series of the load increment factor.

[0010] The time series is input into a pre-trained load increment factor prediction model to obtain the predicted value of the load increment factor;

[0011] The cooling load peak calculation model is corrected based on the predicted value of the load increment factor to obtain the cooling load peak prediction model, and the cooling load peak prediction value is obtained based on the cooling load peak prediction model.

[0012] The cooling load peak prediction model is expressed by the following formula:

[0013]

[0014] In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0015] Further, the step of calculating the peak cooling load based on the reference load includes:

[0016] Acquire summer daily load data, calculate the cooling load value using the benchmark load comparison method based on the summer daily load data and the benchmark load, and calculate the peak value of the cooling load based on the cooling load value;

[0017] The reference load is calculated using the following formula:

[0018]

[0019] The cooling load value is calculated using the following formula:

[0020] L TCL,ij=max{L ij -L0,0}

[0021] The peak value of the cooling load is calculated using the following formula:

[0022] L RTCL,i =max{L TCL,ij}

[0023] In the formula, L represents the daily load data for summer. TCL L is the cooling load value. RTCL For the peak value of cooling load, subscript i is the date index, subscript j is the hour index, L0 is the baseline load, N is the number of typical days, and Ltp is the typical daily load data.

[0024] Furthermore, the step of using a data fitting method to perform time-series coupling on the peak cooling load to obtain the time-series predicted value of the peak cooling load includes:

[0025] The peak value of the cooling load was fitted using the least squares method to obtain the time-series correlation coefficient.

[0026] Based on the time-series correlation coefficient and the preset time-series correlation degree, the peak value of the cooling load is time-series coupled to obtain the time-series predicted value of the cooling load peak.

[0027] The predicted peak cooling load time series value is calculated using the following formula:

[0028]

[0029] In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n The peak value of the cooling load on day tn.

[0030] Furthermore, the step of establishing a calculation model for cooling load peaks based on the load increment factor according to the load accumulation effect includes:

[0031] The difference between the predicted peak cooling load time series value and the peak cooling load value is calculated to obtain the load accumulation term. The load increment factor is then multiplied by the load accumulation term to obtain the load accumulation correction term.

[0032] The peak cooling load time-series prediction value is added to the load accumulation correction term to obtain the peak cooling load calculation model;

[0033] The calculation model for the peak cooling load is expressed by the following formula:

[0034]

[0035] In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0036] Furthermore, a load increment factor prediction model is established based on an LSTM neural network, and the MSE function and Adam optimizer are used as the loss function and optimizer of the load increment factor prediction model, respectively.

[0037] Secondly, embodiments of the present invention provide a summer cooling load peak prediction system, comprising:

[0038] The load data acquisition module is used to select typical day load data from historical load data based on typical day selection criteria.

[0039] The baseline load comparison module is used to calculate the baseline load based on the typical daily load data, and to calculate the peak value of the cooling load based on the baseline load.

[0040] The timing coupling module is used to perform timing coupling on the peak cooling load using a data fitting method to obtain the timing prediction value of the peak cooling load.

[0041] The cooling load peak calculation module is used to establish a cooling load peak calculation model based on the load increment factor according to the load accumulation effect. The peak value of the cooling load and the time series prediction value of the cooling load peak are input into the cooling load peak calculation model to obtain the time series of the load increment factor.

[0042] The incremental factor prediction module is used to input the time series into a pre-trained load incremental factor prediction model to obtain the load incremental factor prediction value.

[0043] The cooling load peak prediction module is used to correct the cooling load peak calculation model based on the predicted value of the load increment factor to obtain the cooling load peak prediction model, and to obtain the cooling load peak prediction value based on the cooling load peak prediction model.

[0044] The cooling load peak prediction model is expressed by the following formula:

[0045]

[0046] In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0047] Furthermore, the timing coupling module includes:

[0048] The coefficient calculation module is used to perform data fitting on the peak value of the cooling load using the least squares method to obtain the time-series correlation coefficient.

[0049] The time-series prediction value calculation module is used to perform time-series coupling on the peak cooling load based on the time-series correlation coefficient and the preset time-series correlation degree to obtain the time-series prediction value of the peak cooling load.

[0050] The predicted peak cooling load time series value is calculated using the following formula:

[0051]

[0052] In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n The peak value of the cooling load on day tn.

[0053] Furthermore, the cooling load peak calculation module includes:

[0054] The load accumulation correction term calculation module is used to calculate the difference between the predicted value of the peak cooling load and the peak value of the cooling load to obtain the load accumulation term, and multiply the load increment factor by the load accumulation term to obtain the load accumulation correction term;

[0055] A peak cooling load calculation model is used to add the predicted peak cooling load time series value to the cumulative load correction term to obtain the peak cooling load calculation model;

[0056] The calculation model for the peak cooling load is expressed by the following formula:

[0057]

[0058] In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0059] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0060] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0061] This invention provides a method, system, device, and storage medium for predicting summer cooling load peaks. Compared with existing technologies, this invention uses data fitting to predict time-series cooling load peaks, defines a cooling load increment factor considering the load accumulation effect, and uses a prediction model based on an LSTM neural network to predict the load increment factor. The prediction results of time-series cooling load peaks are corrected by the load increment factor, thereby obtaining more accurate prediction results, which helps to optimize power system operation and planning. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the summer cooling load peak prediction method provided in an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the cooling load on a certain day of a certain year, provided in an embodiment of the present invention;

[0064] Figure 3 This is a schematic diagram of the peak cooling load in a certain year provided in an embodiment of the present invention;

[0065] Figure 4 This is a schematic diagram comparing the predicted and actual peak cooling load values ​​obtained by the method provided in this invention.

[0066] Figure 5 This is a schematic diagram of the summer cooling load peak prediction system provided in an embodiment of the present invention;

[0067] Figure 6 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation

[0068] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] Please see Figure 1 The first embodiment of the present invention proposes a method for predicting peak summer cooling loads, comprising steps S10 to S60:

[0070] Step S10: Select typical day load data from historical load data according to the typical day selection criteria.

[0071] Step S20: Calculate the baseline load based on the typical daily load data, and calculate the peak value of the cooling load based on the baseline load.

[0072] Cooling load refers to the portion of the power grid load that grows rapidly in a short period of time. For example, due to the hot summer weather, the large-scale use of cooling equipment such as air conditioners and fans increases the power grid load. Therefore, cooling load has become an unavoidable part of the summer power planning process. At present, there are many methods for calculating cooling load. Among them, the benchmark load comparison method is a relatively universal method. Therefore, this invention adopts the benchmark load comparison method to calculate cooling load.

[0073] First, historical load data needs to be preprocessed to select typical daily load data. Following the conventional approach of benchmark load comparison, sample days are usually selected from spring and autumn of previous years, taking into account meteorological information such as temperature and humidity. At the same time, it is necessary to remove samples with obvious abnormalities caused by force majeure factors, such as special dates like holidays, as well as weather conditions such as typhoons and rainy seasons. Therefore, sunny weekdays are usually selected as typical days, and about 30 days of typical daily load data are selected as the data for the next step of processing. Of course, the selection of typical days can also be flexibly chosen according to the actual situation. This is just to provide an optimal method rather than a specific limitation.

[0074] After selecting typical daily load data, the baseline load can be calculated using the baseline load calculation formula:

[0075]

[0076] In the formula, L0 is the baseline load, N is the number of typical days, and Ltp is the typical daily load data. Next, the cooling load value is calculated based on the baseline load comparison method:

[0077] L TCL,ij =max{L ij -L0,0}

[0078] In the formula, L represents the daily load data for summer. TCL The value represents the cooling load, with subscript i being the date index and subscript j being the hour index.

[0079] L ijThese are actual load data measured in summer, typically calculated by time period within a specific date, such as hourly load data over a 24-hour period. In other words, the cooling load value L... TCL,ij In reality, it is a set of cooling load data within a certain date. This value is non-negative because if the difference is calculated to be negative, it means that no clear cooling load was generated during that period, so only zero needs to be taken for that period.

[0080] After obtaining the cooling load value, the peak value of the cooling load can be calculated as follows:

[0081] L RTCL,i =max{L TCL,ij}

[0082] The peak cooling load value for a given date is the value of the maximum cooling load across all time periods of the day.

[0083] Step S30: The peak value of the cooling load is coupled in time using a data fitting method to obtain the time-series predicted value of the peak cooling load.

[0084] After obtaining the peak cooling load, the data fitting method can be used to calculate the time-series prediction of the peak cooling load. The specific steps are as follows:

[0085] Step S301: The least squares method is used to fit the peak value of the cooling load to obtain the time-series correlation coefficient;

[0086] Step S302: Based on the time-series correlation coefficient and the preset time-series correlation degree, perform time-series coupling on the peak value of the cooling load to obtain the time-series predicted value of the peak cooling load.

[0087] Assuming that there is a strong temporal coupling between the peak cooling load in the past n days and the current peak cooling load, n can usually be set to 7 days in weeks, but can be set to other values ​​according to requirements. n is the pre-set temporal correlation.

[0088] Then, data fitting methods such as the least squares method can be used to fit the measured peak cooling load to obtain its time-series correlation coefficient α, with n days corresponding to n time-series correlation coefficients α1, α2, ... α n The specific fitting process can be referred to the conventional calculation method of least squares, and will not be elaborated here.

[0089] Based on the obtained time-series correlation coefficient and time-series correlation degree, a polynomial structure can be used to characterize the time-series coupling information between the predicted peak value of the cooling load and the peak value of the cooling load:

[0090]

[0091] In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n The peak value of the cooling load on day tn.

[0092] Step S40: Based on the load accumulation effect, establish a cooling load peak calculation model based on the load increment factor, input the cooling load peak value and the cooling load peak time series prediction value into the cooling load peak calculation model to obtain the time series of the load increment factor.

[0093] In reality, summer cooling load is not only related to daily meteorological conditions but also exhibits a significant cumulative effect. For example, if the cooling load has been consistently high over a period of time, the future cooling load will show a significant increase under the influence of this continuous high cooling load. Therefore, considering the cumulative effect of summer loads, this invention defines a cooling load increment factor to characterize the cumulative load effect. This factor corrects the original load increment, meaning that the cumulative load effect is incorporated into the prediction results of cooling load peaks, establishing a cooling load peak calculation model based on the load increment factor.

[0094]

[0095] In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0096] As can be seen from the above formula, we have added a defined load increment factor to the difference between the original peak cooling load time-series forecast and the peak cooling load value. That is, a correction term related to the load increment is superimposed on the peak cooling load time-series forecast. If the load increment factor γ... t >0, and This indicates that due to continuous high temperatures, the cooling load has accumulated, and the increase in cooling load compared to the previous day has increased significantly.

[0097] Assuming day t is a future date, then the peak cooling load L RTCL,t It is unknown, and similarly, the load increment factor γ t It is also unknown, but based on the time-series coupling calculation steps described above, the predicted value of the cooling load peak time series can be obtained. Assuming that days t-1, t-2, ..., tn are historical dates, then the corresponding peak cooling load L can be measured. RTCL,t-1 LRTCL,t-2 ..., L RTCL,t-n And the corresponding peak cooling load time-series prediction value is obtained based on time-series coupling calculation. By substituting these known data into the peak cooling load calculation model, the load increment factors for days t-1, t-2, ..., tn can be calculated, thus obtaining the time series of the load increment factors:

[0098] Ξ={γ t-1 ,γ t-2 ,…,γ t-n}

[0099] Step S50: Input the time series into the pre-trained load increment factor prediction model to obtain the predicted value of the load increment factor.

[0100] Because LSTM neural networks effectively address the problem of gradient explosion or vanishing that occurs with increasing training time and network layers, making them unable to handle longer data sequences and thus unable to acquire information from long-distance data, this invention establishes a load increment factor prediction model based on LSTM neural networks. The loss function is defined as the MSE loss function, and the Adam optimizer is used to train the model. Hyperparameters are adjusted during validation to obtain the final load increment factor prediction model. The specific model establishment and training process can refer to the establishment and training of conventional LSTM neural network models, and will not be elaborated here. The calculated time series of the load increment factor is input into the trained load increment factor prediction model for prediction, thereby obtaining the predicted load increment factor value γ for day t. t * .

[0101] Step S60: Correct the cooling load peak calculation model according to the predicted value of the load increment factor to obtain the cooling load peak prediction model, and obtain the cooling load peak prediction value according to the cooling load peak prediction model.

[0102] After obtaining the predicted value of the load increment factor γ t * Then, combining the above-mentioned peak cooling load calculation model, the model can be corrected using the predicted value of the load increment factor to obtain the peak cooling load prediction model:

[0103]

[0104] In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t.RTCL,t-1 The peak value of the cooling load on day t-1.

[0105] Predicted peak cooling load Because the predicted values ​​have been corrected by the load increment factor and take into account the load accumulation effect, the prediction results are more accurate than those of conventional methods. This improves the scientificity and rationality of the operation and planning of the power system in summer, and helps to optimize the operation and planning of the power system.

[0106] The following example uses the method provided in this invention to predict peak cooling loads in a coastal province of my country. Based on the baseline load comparison method, the historical summer cooling loads of the province can be calculated. Please refer to [link / reference]. Figure 2 Taking a specific day in the summer of 2017 as an example, the cooling load for that day is calculated. Then, based on the cooling load, the peak cooling load is calculated, i.e., Figure 3 As shown.

[0107] Based on the obtained peak cooling load, prediction processing is performed according to the steps of the prediction method provided by the present invention, as described above, to obtain the following result: Figure 4 The chart shown compares the predicted and actual peak cooling load values ​​for a certain period in 2021. Figure 4 It can be clearly seen that the curves of the predicted values ​​and the actual values ​​obtained by the method provided by this invention are consistent, indicating that the prediction method of this invention can accurately predict the peak value of summer cooling load, thereby helping to optimize the operation and planning of the power system in summer.

[0108] This invention provides a method for predicting summer cooling load peaks. Compared to traditional methods that are overly simplistic, fail to fully utilize the cumulative effect of cooling loads, and ignore the impact of cooling load increments, this invention fully utilizes historical cooling load peak data. Based on the time-series information of cooling loads, it specifically considers the cumulative effect of historical cooling load peaks over a past period on future cooling loads, proposing a load increment factor to characterize the cumulative effect of cooling loads. Through the load increment factor, the prediction results are amplified when load increments continue to accumulate. The proposed solution helps to more accurately predict user loads, rationally address summer power grid peak loads caused by cooling loads, efficiently ensure peak summer power supply, and has significant advantages.

[0109] Please see Figure 5 Based on the same inventive concept, the second embodiment of the present invention provides a summer cooling load peak prediction system, comprising:

[0110] The load data acquisition module 10 is used to select typical day load data from historical load data according to typical day selection conditions;

[0111] The baseline load comparison module 20 is used to calculate the baseline load based on the typical daily load data, and to calculate the peak value of the cooling load based on the baseline load.

[0112] The timing coupling module 30 is used to perform timing coupling on the peak value of the cooling load using a data fitting method to obtain the timing prediction value of the peak cooling load.

[0113] The cooling load peak calculation module 40 is used to establish a cooling load peak calculation model based on the load increment factor according to the load accumulation effect. The peak value of the cooling load and the time series prediction value of the cooling load peak are input into the cooling load peak calculation model to obtain the time series of the load increment factor.

[0114] The incremental factor prediction module 50 is used to input the time series into a pre-trained load incremental factor prediction model to obtain the load incremental factor prediction value.

[0115] The cooling load peak prediction module 60 is used to correct the cooling load peak calculation model according to the predicted value of the load increment factor to obtain the cooling load peak prediction model, and to obtain the cooling load peak prediction value according to the cooling load peak prediction model.

[0116] The cooling load peak prediction model is expressed by the following formula:

[0117]

[0118] In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0119] Furthermore, the timing coupling module 30 includes:

[0120] The coefficient calculation module 301 is used to perform data fitting on the peak value of the cooling load using the least squares method to obtain the time-series correlation coefficient.

[0121] The time-series prediction value calculation module 302 is used to perform time-series coupling on the peak cooling load based on the time-series correlation coefficient and the preset time-series correlation degree to obtain the time-series prediction value of the peak cooling load.

[0122] The predicted peak cooling load time series value is calculated using the following formula:

[0123]

[0124] In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n The peak value of the cooling load on day tn.

[0125] Furthermore, the cooling load peak calculation module 40 includes:

[0126] The load accumulation correction term calculation module 401 is used to calculate the difference between the predicted value of the peak cooling load and the peak value of the cooling load to obtain the load accumulation term, and multiply the load increment factor with the load accumulation term to obtain the load accumulation correction term;

[0127] Cooling load peak calculation model 402 is used to add the cooling load peak time series prediction value to the load accumulation correction term to obtain the cooling load peak calculation model;

[0128] The calculation model for the peak cooling load is expressed by the following formula:

[0129]

[0130] In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

[0131] The technical features and effects of the summer cooling load peak prediction system proposed in this embodiment of the invention are the same as those of the method proposed in this embodiment of the invention, and will not be repeated here. Each module in the above-mentioned summer cooling load peak prediction system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or it can be stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0132] Please see Figure 6The diagram illustrates the internal structure of a computer device in one embodiment. This computer device can specifically be a terminal or a server. The computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for predicting peak summer cooling loads. The display screen of the computer device can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0133] Those skilled in the art will understand that Figure 6 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than shown in the diagram, or combine certain components, or have the same component arrangement.

[0134] Furthermore, embodiments of the present invention also propose a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0135] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0136] In summary, this invention provides a method, system, device, and storage medium for predicting summer cooling load peaks. The prediction method involves selecting typical day load data from historical load data based on typical day selection criteria; calculating a baseline load based on the typical day load data; calculating the peak cooling load based on the baseline load; using a data fitting method to time-series couple the peak cooling load to obtain a time-series predicted value for the peak cooling load; establishing a peak cooling load calculation model based on a load increment factor based on the load accumulation effect; inputting the peak cooling load and the time-series predicted value for the peak cooling load into the peak cooling load calculation model to obtain a time series of the load increment factor; inputting the time series into a pre-trained load increment factor prediction model to obtain a predicted value of the load increment factor; correcting the peak cooling load calculation model based on the predicted value of the load increment factor to obtain a peak cooling load prediction model; and obtaining the peak cooling load prediction value based on the peak cooling load prediction model. This invention fully considers the impact of load accumulation on summer cooling loads, and corrects the prediction results of time-series cooling load peaks by using load increment factors, thereby improving the accuracy of the prediction results. This helps to optimize the operation and planning of the power system, and further improves the scientificity and rationality of the power system operation.

[0137] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0138] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A method for predicting peak summer cooling load, characterized in that, include: Typical day load data are selected from historical load data based on the typical day selection criteria. Based on the typical daily load data, the baseline load is calculated, and based on the baseline load, the peak value of the cooling load is calculated. The peak cooling load was temporally coupled using a data fitting method to obtain the temporal prediction value of the peak cooling load. Based on the load accumulation effect, a calculation model for cooling load peak based on load increment factor is established. The peak value of cooling load and the time series prediction value of cooling load peak are input into the calculation model to obtain the time series of the load increment factor. The time series is input into a pre-trained load increment factor prediction model to obtain the predicted value of the load increment factor; The cooling load peak calculation model is corrected based on the predicted value of the load increment factor to obtain the cooling load peak prediction model, and the cooling load peak prediction value is obtained based on the cooling load peak prediction model. The cooling load peak prediction model is expressed by the following formula: In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

2. The method for predicting peak summer cooling load according to claim 1, characterized in that, The step of calculating the peak cooling load based on the baseline load includes: Acquire summer daily load data, calculate the cooling load value using the benchmark load comparison method based on the summer daily load data and the benchmark load, and calculate the peak value of the cooling load based on the cooling load value; The reference load is calculated using the following formula: The cooling load value is calculated using the following formula: L TCL,ij <max{L ij -L0,0} The peak value of the cooling load is calculated using the following formula: L RTCL,i <max{L TCL,ij } In the formula, L represents the daily load data for summer. TCL L is the cooling load value. RTCL For the peak value of cooling load, subscript i is the date index, subscript j is the hour index, L0 is the baseline load, N is the number of typical days, and Ltp is the typical daily load data.

3. The method for predicting peak summer cooling load according to claim 2, characterized in that, The step of using a data fitting method to perform time-series coupling on the peak cooling load to obtain the time-series predicted value of the peak cooling load includes: The peak value of the cooling load was fitted using the least squares method to obtain the time-series correlation coefficient. Based on the time-series correlation coefficient and the preset time-series correlation degree, the peak value of the cooling load is time-series coupled to obtain the time-series predicted value of the cooling load peak. The predicted peak cooling load time series value is calculated using the following formula: In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n This represents the peak cooling load on day tn.

4. The method for predicting peak summer cooling load according to claim 3, characterized in that, The steps for establishing a calculation model for cooling load peaks based on the load increment factor according to the load accumulation effect include: The difference between the predicted peak cooling load time series value and the peak cooling load value is calculated to obtain the load accumulation term. The load increment factor is then multiplied by the load accumulation term to obtain the load accumulation correction term. The peak cooling load time-series prediction value is added to the load accumulation correction term to obtain the peak cooling load calculation model; The calculation model for the peak cooling load is expressed by the following formula: In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

5. The method for predicting peak summer cooling load according to claim 1, characterized in that, The load increment factor prediction model is established based on the LSTM neural network, and the MSE function and Adam optimizer are used as the loss function and optimizer of the load increment factor prediction model, respectively.

6. A summer cooling load peak prediction system, characterized in that, include: The load data acquisition module is used to select typical day load data from historical load data based on typical day selection criteria. The baseline load comparison module is used to calculate the baseline load based on the typical daily load data, and to calculate the peak value of the cooling load based on the baseline load. The timing coupling module is used to perform timing coupling on the peak cooling load using a data fitting method to obtain the timing prediction value of the peak cooling load. The cooling load peak calculation module is used to establish a cooling load peak calculation model based on the load increment factor according to the load accumulation effect. The peak value of the cooling load and the time series prediction value of the cooling load peak are input into the cooling load peak calculation model to obtain the time series of the load increment factor. The incremental factor prediction module is used to input the time series into a pre-trained load incremental factor prediction model to obtain the load incremental factor prediction value. The cooling load peak prediction module is used to correct the cooling load peak calculation model based on the predicted value of the load increment factor to obtain the cooling load peak prediction model, and to obtain the cooling load peak prediction value based on the cooling load peak prediction model. The cooling load peak prediction model is expressed by the following formula: In the formula, γ is the predicted peak cooling load value for day t. t * This is the predicted value of the load increment factor. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

7. The summer cooling load peak prediction system according to claim 6, characterized in that, The timing coupling module includes: The coefficient calculation module is used to perform data fitting on the peak value of the cooling load using the least squares method to obtain the time-series correlation coefficient. The time-series prediction value calculation module is used to perform time-series coupling on the peak cooling load based on the time-series correlation coefficient and the preset time-series correlation degree to obtain the time-series prediction value of the peak cooling load. The predicted peak cooling load time series value is calculated using the following formula: In the formula, Let n be the predicted peak cooling load on day t, where n is the time series correlation degree, α is the time series correlation coefficient, and L is the time series correlation coefficient. RTCL,t-n The peak value of the cooling load on day tn.

8. The summer cooling load peak prediction system according to claim 7, characterized in that, The cooling load peak calculation module includes: The load accumulation correction term calculation module is used to calculate the difference between the predicted value of the peak cooling load and the peak value of the cooling load to obtain the load accumulation term, and multiply the load increment factor by the load accumulation term to obtain the load accumulation correction term; A peak cooling load calculation model is used to add the predicted peak cooling load time series value to the cumulative load correction term to obtain the peak cooling load calculation model; The calculation model for the peak cooling load is expressed by the following formula: In the formula, L RTCL,t γ represents the peak cooling load on day t. t Let be the load increment factor on day t. L is the predicted peak cooling load for day t. RTCL,t-1 The peak value of the cooling load on day t-1.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images