Method, device and computer equipment for forecasting electricity spot price

Abstract

The application relates to a power spot price prediction method, device and computer equipment. The method comprises the following steps: obtaining historical power loads of each historical moment in a historical period corresponding to a current moment, and meteorological data of the current moment; wherein the historical period and a current period to which the current moment belongs satisfy a preset time relationship; determining a target load interval to which a predicted power load of the current moment belongs according to the meteorological data and the historical power loads; and determining a predicted power spot price of the current moment according to the target load interval. By using the method, the predicted power spot price of the current moment can be determined differently according to the target load interval to which the predicted power load of the current moment belongs, so that the prediction accuracy of the power spot price can be improved.

Description

Technical Field

[0001] This application relates to the field of electricity price forecasting technology, and in particular to a method, apparatus and computer equipment for forecasting spot electricity prices. Background Technology

[0002] In the electricity spot market environment, electricity spot prices are determined by market supply and demand, exhibiting high volatility, nonlinearity, and multi-period characteristics. Accurate electricity spot price forecasting is the core basis for market participants such as power generation companies, electricity sales companies, and new energy storage to formulate trading strategies and optimize resource allocation.

[0003] In related technologies, a unified mapping relationship between power load and electricity spot price is typically used for electricity spot price forecasting. However, with the rapid increase in the proportion of renewable energy installations, the relationship between power load and electricity spot price in the current power system has changed.

[0004] Therefore, improving the accuracy of electricity spot price forecasts is a problem worthy of attention. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, and computer equipment for predicting electricity spot prices that can improve the accuracy of electricity spot price prediction, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for predicting spot electricity prices, including:

[0007] Obtain the historical power load for each historical time period within the historical time period corresponding to the current time, as well as the meteorological data for the current time; wherein, the historical time period and the current time period to which the current time belongs satisfy a preset time relationship;

[0008] Based on meteorological data and historical power loads, determine the target load range to which the current predicted power load belongs;

[0009] Determine the current spot electricity price based on the target load range.

[0010] In one embodiment, determining the target load interval to which the predicted power load belongs at the current moment based on meteorological data and historical power loads includes: determining at least one candidate load interval based on historical power loads; determining the predicted power load at the current moment based on meteorological data and historical power loads; and determining the target load interval to which the predicted power load belongs from the candidate load intervals.

[0011] In one embodiment, at least one candidate load interval includes a first load interval, a second load interval, and a third load interval. Determining the predicted spot electricity price for the current moment based on the target load interval includes: if the target load interval is the first load interval, determining a preset spot electricity price as the predicted spot electricity price for the current moment; if the target load interval is the second load interval and the current moment is within a preset time window, determining the predicted spot electricity price for the current moment based on the power load impact factor corresponding to the current moment and the historical spot electricity prices for each historical moment; if the target load interval is the second load interval and the current moment is outside the preset time window, or if the target load interval is the third load interval, determining a reference moment from each historical moment based on the predicted power load for the current moment and the historical power load for each historical moment; and determining the predicted spot electricity price for the current moment based on the historical spot electricity price for the reference moment; wherein each power load in the first load interval is less than each power load in the second load interval; and each power load in the second load interval is less than each power load in the third load interval.

[0012] In one embodiment, determining the predicted electricity spot price for the current moment based on the electricity load impact factor corresponding to the current moment and the historical electricity spot price for each historical moment includes: selecting at least one target historical moment from each historical moment based on the electricity load impact factor corresponding to the current moment; wherein each target historical moment is located within a historical time period and within a historical time window corresponding to a preset time window, and the electricity load impact factor corresponding to each target historical moment matches the electricity load impact factor corresponding to the current moment; and determining the predicted electricity spot price for the current moment based on the historical electricity spot price for each target historical moment.

[0013] In one embodiment, determining the predicted electricity spot price for the current moment based on the historical electricity spot price for each target historical moment includes: grouping each target historical moment into at least one first time group based on the historical electricity load for each target historical moment; determining the first time group to which the current moment belongs based on the predicted electricity load for the current moment, to obtain a target time group; determining the median of the historical electricity spot prices for each target historical moment belonging to the target time group, and determining the determined median as the predicted electricity spot price for the current moment.

[0014] In one embodiment, a reference time is determined from the historical times based on the predicted power load at the current time and the historical power load at each historical time, including: determining the similarity between the predicted power load at the current time and the historical power load at each historical time; and selecting each historical time with a similarity greater than a similarity threshold to obtain the reference time.

[0015] In one embodiment, there are multiple reference times; determining the predicted electricity spot price for the current time based on the historical electricity spot price of the reference times includes: grouping each reference time according to the target time period to which each reference time belongs to obtain at least one second time group; for each second time group, grouping each reference time in the second time group according to the historical electricity load of each reference time in the second time group to obtain at least one time subgroup corresponding to the second time group; and determining the median of the historical electricity spot price of each reference time in each time subgroup corresponding to the second time group; and determining the predicted electricity spot price for the current time based on the determined medians.

[0016] In one embodiment, determining the predicted spot electricity price at the current time based on the determined medians includes: for each second time group, determining the correspondence between the spot electricity price and the electricity load corresponding to the second time group based on the medians corresponding to the second time group; and determining the predicted spot electricity price at the current time based on the second time group to which the current time belongs, the predicted electricity load at the current time, and the correspondence between the spot electricity price and the electricity load corresponding to each second time group.

[0017] Secondly, this application also provides a device for predicting electricity spot prices, comprising:

[0018] The acquisition module is used to acquire the historical power load of each historical time within the historical time period corresponding to the current time, as well as the meteorological data of the previous time; wherein, the historical time period and the current time period to which the previous time period belongs satisfy a preset time relationship;

[0019] The first determining module is used to determine the target load range to which the predicted power load belongs at the current moment based on meteorological data and historical power loads.

[0020] The second determining module is used to determine the predicted spot price of electricity at the current moment based on the target load range.

[0021] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0022] Obtain the historical power load for each historical time period within the historical time period corresponding to the current time, as well as the meteorological data for the current time; wherein, the historical time period and the current time period to which the current time belongs satisfy a preset time relationship;

[0023] Based on meteorological data and historical power loads, determine the target load range to which the current predicted power load belongs;

[0024] Determine the current spot electricity price based on the target load range.

[0025] The aforementioned method, apparatus, and computer equipment for predicting electricity spot prices acquire historical electricity load data for each historical time period within the historical time frame corresponding to the current time, as well as meteorological data for the current time. The historical time periods and the current time period satisfy a preset time relationship. Based on the meteorological data and the historical electricity load data, the target load range to which the predicted electricity load belongs at the current time is determined. Based on the target load range, the predicted electricity spot price for the current time is determined. In this process, by using the target load range to which the predicted electricity load belongs obtained through prior prediction, the predicted electricity spot price for the current time can be determined differently for different load ranges. This avoids the inaccuracy of prediction results caused by using a uniform mapping relationship between electricity load and electricity spot price, and can improve the accuracy of electricity spot price prediction. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart illustrating a method for predicting spot electricity prices in one embodiment;

[0028] Figure 2 This is a flowchart illustrating the steps for determining the target load range in one embodiment;

[0029] Figure 3 This is a flowchart illustrating the steps for determining the predicted spot price of electricity in one embodiment.

[0030] Figure 4 This is a flowchart illustrating the steps for determining the predicted spot price of electricity in another embodiment;

[0031] Figure 5 This is a flowchart illustrating a method for predicting spot electricity prices in another embodiment;

[0032] Figure 6 This is a structural block diagram of a device for predicting spot electricity prices in one embodiment;

[0033] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0035] The method for predicting spot electricity prices provided in this application can be applied to computer equipment, which can be a terminal or a server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services.

[0036] In one exemplary embodiment, a method for predicting spot electricity prices is provided, illustrated using an application of this method to a server as an example. See also Figure 1 The method for predicting spot electricity prices, as shown, includes the following steps:

[0037] S110, obtain the historical power load of each historical time within the historical time period corresponding to the current time, as well as the meteorological data of the current time; wherein, the historical time period and the current time period to which the current time belongs satisfy a preset time relationship.

[0038] The current moment can be understood as the moment when the spot price of electricity is to be predicted.

[0039] Here, the current time period can be understood as the time period during which the spot electricity price to be predicted is to be determined. The current time period may include at least one moment during which the spot electricity price to be predicted is to be determined.

[0040] The historical time period corresponding to the current moment may include at least one historical moment.

[0041] In some embodiments, at least one historical moment can be obtained by dividing a historical period into preset durations. For example, the preset duration can be 15 minutes.

[0042] Historical power load can be understood as the power consumed by electrical equipment.

[0043] Historical power load refers to the total power load within the target dispatch range, excluding tie-line power load and renewable energy generation load. Tie-line power load can be understood as the power received from the external power grid through inter-regional transmission tie points. Renewable energy generation load can be understood as the power generated by renewable energy sources.

[0044] Meteorological data can be understood as predicted meteorological data.

[0045] Optionally, the preset time relationship between the historical time period and the current time period can be that the historical time period precedes the current time period and the historical time period is continuous with the current time period. For example, the current time period can include the current forecast date and one day in the future, and the historical time period can include seven consecutive historical days preceding the current forecast date.

[0046] Alternatively, the preset time relationship between the historical time period and the current time period can be that the historical time period precedes the current time period, and the historical time period and the current time period are not consecutive. For example, the current time period can include the current forecast date and one day in the future, and the historical time period can include seven consecutive historical days that are two days apart from the current forecast date and precede the current forecast date.

[0047] In some embodiments, the historical power load of each historical time within the historical period corresponding to the current time can be obtained from the power trading platform.

[0048] In some embodiments, meteorological data at the current moment can be obtained from a meteorological data service platform. Optionally, the meteorological data may include at least one of temperature, radiation intensity, wind speed, and cloud cover.

[0049] In some embodiments, the historical power load data at each historical time point and the meteorological data at the current time point may be preprocessed as follows: the historical power load data at each historical time point and the meteorological data at the current time point are timestamped; at least one of outlier handling and missing value handling is performed on the historical power load data at each historical time point; at least one of outlier handling and missing value handling is performed on the meteorological data at the current time point. The preprocessed historical power load data at each historical time point and the meteorological data at the current time point can be used to perform subsequent steps.

[0050] S120, based on meteorological data and historical power loads, determines the target load range to which the current predicted power load belongs.

[0051] The target load range can be understood as the target load range to which the predicted power load belongs at the current moment.

[0052] In some embodiments, the predicted power load at the current moment can be determined based on meteorological data and historical power loads; and the target load range to which the predicted power load belongs can be determined.

[0053] In some embodiments, meteorological data can be feature extracted to obtain meteorological features; historical power loads can be feature extracted to obtain load time-series features; meteorological features and load time-series features can be fused to obtain fused features; and the predicted power load at the current moment can be determined based on the fused features.

[0054] In some embodiments, at least one power load forecasting model can be used to determine the forecast power load at the current moment based on meteorological data and historical power loads.

[0055] Specifically, for each power load forecasting model, meteorological data and historical power loads can be input into the model to obtain the predicted power load for the current moment. After obtaining the predicted power loads for the current moment from each model, a weighted sum can be performed to obtain the predicted power load for the current moment.

[0056] The power load forecasting model can include a meteorological feature extraction network, a load time series feature extraction network, a feature fusion network, and a forecasting network.

[0057] A meteorological feature extraction network is used to extract features from meteorological data to obtain meteorological features; a load time series feature extraction network is used to extract features from each historical power load to obtain load time series features; a feature fusion network is used to fuse meteorological features and load time series features to obtain fused features; and a prediction network is used to predict the current power load based on the fused features.

[0058] The power load forecasting model can be implemented based on a deep learning model, and this application does not impose any limitations on the specific network structure of the power load forecasting model. For example, the power load forecasting model can be implemented based on a Long Short-Term Memory (LSTM) neural network, a Temporal Fusion Transformer (TFT), a Gated Recurrent Unit (GRU), or a Temporal Convolutional Network (TCN).

[0059] In some embodiments, the power load forecasting model can be trained as follows: the actual power load corresponding to the sample forecast time is obtained; historical power load samples at each historical time within the sample historical period corresponding to the sample forecast time are obtained; and meteorological data samples at the sample forecast time are obtained. Using the historical power load samples at each historical time within the sample historical period corresponding to the sample forecast time and the meteorological data samples at the sample forecast time as training samples, and using the actual power load at the corresponding sample forecast time as the label, the network parameters of the pre-constructed power load forecasting model are adjusted until the training cutoff condition is met. The training cutoff condition may include at least one of the following: the number of training samples reaches a preset threshold, the number of iterations of model training reaches a preset threshold, the model accuracy reaches a preset accuracy threshold, and the model tends to converge. The preset thresholds for the number of training samples, the preset number of iterations, and the preset accuracy threshold can be set or adjusted by technicians according to their needs or experience, or determined repeatedly through numerous experiments; this application does not impose any limitations on these settings.

[0060] In some embodiments, when acquiring training samples, a sliding window mechanism can be used to slide sampling across multiple historical time points with a target number of time points as the step size, generating a sample time sequence with a first number of historical time points as sample historical time points and a second number of historical time points as sample prediction time points. The sum of the first and second numbers is a multiple of the target number.

[0061] In some embodiments, when training the power load forecasting model, in order to eliminate the numerical discontinuity at the cycle boundary, periodic sine and cosine encoding can be used to encode the historical sample time and the predicted sample time in the sample time sequence into a unit circle, so that the distance between the beginning and end of the sample time in the feature space tends to be consistent. This allows the power load forecasting model to learn the cyclical pattern of the sample time in the cycle, while avoiding the introduction of false numerical order that would affect the forecasting results.

[0062] In some embodiments, when training the power load forecasting model, the actual power load corresponding to the sample forecast time can be standardized using a robust standardization tool (RobustScaler) to remove the dimensions of the actual power load. Based on meteorological characteristics, load time-series characteristics, and the standardized actual power load, the power load forecasting model is trained to obtain the load forecast results corresponding to the sample forecast time. After obtaining the load forecast results corresponding to the sample forecast time, the load forecast results can be inversely transformed to restore the predicted power load with dimensions.

[0063] S130, determine the forecasted spot electricity price for the current moment based on the target load range.

[0064] In some embodiments, a reference electricity spot price corresponding to at least one reference load range can be obtained. Based on the target load range, the reference electricity spot price corresponding to the target load range is queried from the reference electricity spot prices corresponding to at least one reference load range, and the queried reference electricity spot price corresponding to the target load range is used as the predicted electricity spot price at the current time.

[0065] In the above-mentioned method for predicting electricity spot prices, by using the target load range to which the predicted electricity load belongs at the current moment, the predicted electricity spot price at the current moment can be determined differently for different load ranges. This avoids the inaccuracy of prediction results caused by using a uniform mapping relationship between electricity load and electricity spot price, and can improve the accuracy of electricity spot price prediction.

[0066] Based on the technical solutions of the above embodiments, this application also provides another optional embodiment, in which the target load range determination step of S120 is refined.

[0067] See Figure 2 The steps for determining the target load range shown include:

[0068] S210, determine at least one candidate load range based on each historical power load.

[0069] In some embodiments, at least one load threshold may be determined based on each historical power load; and at least one candidate load interval may be determined based on the at least one load threshold.

[0070] In some embodiments, at least one load threshold may include a first load threshold and a second load threshold. The first load threshold is less than the second load threshold. Electrical loads less than the first load threshold may be defined as a first load range, electrical loads not less than the first load threshold and less than the second load threshold may be defined as a second load range, and electrical loads not less than the second load threshold may be defined as a third load range.

[0071] In some embodiments, the first load threshold can be determined according to the following steps: obtaining the historical spot price of electricity corresponding to each historical power load; selecting multiple first candidate load thresholds from the first load range; traversing each first candidate load threshold in ascending order of load value, and performing the following steps: for the current first candidate load threshold, selecting at least one first reference power load not greater than the first candidate load threshold from each historical power load; determining a first ratio between the number of first reference power loads with a corresponding negative historical spot price and the total number of at least one first reference power load; if the first ratio is greater than or equal to the first ratio threshold, using the first candidate load threshold as the first load threshold.

[0072] In some embodiments, if the first ratio corresponding to at least two consecutive first candidate load thresholds is less than the first ratio threshold, the last first candidate load threshold is used as the first load threshold.

[0073] In some embodiments, the second load threshold can be determined according to the following steps: obtaining the historical spot price of electricity corresponding to each historical power load; selecting multiple second candidate load thresholds from the second load range; traversing each second candidate load threshold in ascending order of load value, and performing the following steps: for the current second candidate load threshold, selecting at least one second reference power load from each historical power load that is not less than the second candidate load threshold; determining a second ratio between the number of second reference power loads with a corresponding negative historical spot price and the total number of at least one second reference power load; if the second ratio is less than or equal to a second ratio threshold, using the second candidate load threshold as the second load threshold.

[0074] In some embodiments, the first load threshold and the second load threshold can be preset values. The first load threshold and the second load threshold can be set according to actual needs or experience.

[0075] S220 determines the predicted power load for the current moment based on meteorological data and historical power loads.

[0076] The process of determining the predicted power load can be referred to in the previous text, and will not be repeated here.

[0077] S230: Determine the target load interval to which the predicted power load belongs from each candidate load interval.

[0078] In some embodiments, the target load interval to which the predicted power load belongs can be determined based on a comparison between the predicted power load and the load values ​​within each candidate load interval.

[0079] Specifically, for any candidate load interval, if the predicted power load is not greater than the maximum value within the candidate load interval and not less than the minimum value within the candidate load interval, then the candidate load interval is determined as the target load interval to which the predicted power load belongs.

[0080] In this embodiment, by using meteorological data and historical power loads, the predicted power load at the current moment can be determined relatively accurately. This allows for the accurate determination of the target load range to which the predicted power load belongs from each candidate load range, laying the foundation for accurate prediction of electricity spot prices in the future.

[0081] Based on the technical solutions of the above embodiments, this application also provides another optional embodiment. In this optional embodiment, at least one candidate load interval is refined into a first load interval, a second load interval, and a third load interval. Each power load in the first load interval is less than each power load in the second load interval; each power load in the second load interval is less than each power load in the third load interval. Correspondingly, the step of determining the predicted electricity spot price in S130 is refined. Specifically, the step of determining the predicted electricity spot price includes:

[0082] 1) When the target load range is the first load range, the preset electricity spot price is determined as the predicted electricity spot price at the current moment.

[0083] The preset spot electricity price can be set according to actual needs or experience. For example, the preset spot electricity price can be a floor price.

[0084] 2) If the target load range is the second load range and the current time is within the preset time window, determine the predicted electricity spot price for the current time based on the power load impact factor corresponding to the current time and the historical electricity spot price for each historical time.

[0085] The preset time window can be understood as a preset time period. For example, if the current time is in January, February, November, or December, the preset time window is 9:00-17:30; if the current time is in March or October, the preset time window is 8:30-18:00; and if the current time is in April, May, June, July, August, or September, the preset time window is 8:00-18:30.

[0086] Among them, the power load influencing factor can be understood as the factors that affect the power load.

[0087] Optionally, the power load impact factor may include at least one of holiday attributes and environmental radiation level.

[0088] The holiday attribute can include both public holidays and non-public holidays.

[0089] The environmental radiation level can be determined based on the environmental radiation level at the current moment and at least one radiation threshold.

[0090] In some embodiments, when the ambient radiation level is not greater than a first radiation threshold, the ambient radiation level is determined to be a first radiation level; when the ambient radiation level is greater than the first radiation threshold but not greater than a second radiation threshold, the ambient radiation level is determined to be a second radiation level; and when the ambient radiation level is greater than the second radiation threshold, the ambient radiation level is determined to be a third radiation level.

[0091] In some embodiments, the following types of power load impact can be determined based on holiday attributes and environmental radiation levels: first radiation level - holidays; first radiation level - holidays and non-holidays; second radiation level - holidays; second radiation level - holidays and non-holidays; third radiation level - holidays; third radiation level - holidays and non-holidays.

[0092] In some embodiments, the predicted electricity spot price for the current moment can be determined based on the type of electricity load impact corresponding to the current moment and the historical electricity spot price for each historical moment.

[0093] In some embodiments, the predicted spot electricity price can be determined by the following steps: a weighted sum of historical spot prices at each historical time point is obtained to arrive at the base spot electricity price; the price influence coefficient for the current time point is determined based on the electricity load influence factor; and the predicted spot electricity price for the current time point is obtained based on the price influence coefficient and the base spot electricity price.

[0094] Among them, the price influence coefficient is a dimensionless coefficient.

[0095] In some embodiments, the price influence coefficient at the current moment can be determined based on the power load influence factor at the current moment and the preset price influence coefficient corresponding to each power load influence factor.

[0096] Specifically, if there is only one power load influencing factor, the preset price influence coefficient corresponding to that power load influencing factor can be used as the price influence coefficient at the current moment. If there are two power load influencing factors, the maximum preset price influence coefficient among the preset price influence coefficients corresponding to each power load influencing factor can be used as the price influence coefficient at the current moment.

[0097] In some embodiments, the product of the base electricity spot price and the electricity spot price influence coefficient can be used as the predicted electricity spot price at the current moment.

[0098] 3) If the target load range is the second load range and the current time is outside the preset time window, or if the target load range is the third load range, determine the reference time from each historical time based on the predicted power load at the current time and the historical power load at each historical time; determine the power spot price at the current time based on the historical power spot price at the reference time.

[0099] The reference time can be understood as a historical time when the similarity between the corresponding historical power load and the predicted power load at the current time is greater than a similarity threshold.

[0100] In some embodiments, the reference time can be obtained according to the following steps: determining the similarity between the predicted power load at the current time and the historical power load at each historical time; selecting each historical time with a similarity greater than a similarity threshold to obtain the reference time.

[0101] In some embodiments, each historical moment may include each first historical moment within a first historical time period corresponding to the current moment, each second historical moment within a second historical time period corresponding to the current moment, and each third historical moment within a third historical time period corresponding to the current moment. The duration of the first historical time period is shorter than the duration of the second historical time period, and the duration of the second historical time period is shorter than the duration of the third historical time period.

[0102] For example, each historical moment can include the first historical moments within 30 days before the current moment, the second historical moments within 395 to 336 days before the current moment, and the third historical moments within all historical dates before the current moment.

[0103] In some embodiments, a first number of reference times can be determined from each first historical moment within a first historical period, a second number of reference times can be determined from each second historical moment within a second historical period, and a third number of reference times can be determined from each third historical moment within a third historical period.

[0104] For example, 768 reference times (i.e., 8 days, 96 reference times per day) are determined from the first historical times within 30 days before the current time, 576 reference times (i.e., 6 days, 96 reference times per day) are determined from the second historical times within 395 to 336 days before the current time, and 384 reference times (i.e., 4 days, 96 reference times per day) are determined from the third historical times within all historical dates before the current time.

[0105] Optionally, the similarity can be determined based on the Manhattan distance between the predicted power load at the current moment and the historical power load at each historical moment.

[0106] The process described above, which compares the predicted power load at the current moment with the historical power load at each historical moment, can accurately determine the reference time, laying the foundation for subsequently determining the predicted spot price of electricity.

[0107] In some embodiments, the current spot electricity price can be obtained by the following steps: if there is only one reference time, the historical spot electricity price of the reference time is used as the current spot electricity price; if there are multiple reference times, the weighted result of the historical spot electricity prices of the multiple reference times is used as the current spot electricity price.

[0108] In the above embodiments, by determining the current electricity spot price differently for different load intervals based on the target load interval to which the predicted electricity load belongs, the inaccurate prediction results caused by using a uniform mapping relationship between electricity load and electricity spot price are avoided, thereby improving the prediction accuracy of electricity spot price.

[0109] Based on the technical solutions of the above embodiments, this application also provides another optional embodiment. In this optional embodiment, when the target load range is the second load range and the current time is within a preset time window, the steps for determining the predicted spot price of electricity at the current time are refined.

[0110] See Figure 3 The steps for determining the predicted spot price of electricity, as shown, include:

[0111] S310, based on the power load impact factor corresponding to the current time, select at least one target historical time from all historical times.

[0112] Among them, each target historical moment is located within the historical period and the corresponding historical time window of the preset time window, and the power load impact factor corresponding to each target historical moment matches the power load impact factor corresponding to the current moment.

[0113] Optionally, the power load impact factor corresponding to each target historical time is the same as the power load impact factor corresponding to the current time.

[0114] In some embodiments, at least one candidate historical moment can be selected from each historical moment whose corresponding power load influence factor matches the power load influence factor corresponding to the current moment, based on the power load influence factor corresponding to the current moment.

[0115] In practice, historical moments where the corresponding power load impact factor is the same as the current power load impact factor can be used as candidate historical moments.

[0116] In some embodiments, for each candidate historical moment, if the candidate historical moment is located within the same historical time window corresponding to a preset time window, the candidate historical moment is taken as the target historical moment.

[0117] For example, if the current time's power load impact factor is that the radiation level is at the first level and the current time is a holiday, and the current time is within a preset time window of 8:00-18:30, then the historical time when the corresponding radiation level is at the first level, the current time is a holiday, and the current time is within the historical time window of 8:00-18:30 can be used as the target historical time.

[0118] S320 determines the predicted spot price of electricity for the current moment based on the historical spot prices of electricity at each target historical time.

[0119] In some embodiments, the weighted sum of historical electricity spot prices at a target historical moment can be used as the predicted electricity spot price at the current moment.

[0120] In some embodiments, each target historical time can be grouped according to the historical power load of each target historical time to obtain at least one first time group; the first time group to which the current time belongs can be determined according to the predicted power load of the current time to obtain the target time group; the median of the historical electricity spot price of each target historical time belonging to the target time group can be determined, and the determined median can be determined as the predicted electricity spot price of the current time.

[0121] Each first time group includes at least one target historical time.

[0122] The target time group can be understood as the first time group to which the current time belongs.

[0123] In some embodiments, grouping each target historical time according to the historical power load of each target historical time to obtain at least one first time group may include: obtaining at least one first power load grouping range; for each target historical time, determining the first power load grouping range to which the historical power load of the target historical time belongs; and dividing the target historical times belonging to the same first power load grouping range into a first time group.

[0124] The maximum load difference within each first power load group is the same; for example, the maximum load difference within each first power load group is 1000 MW. Each first power load group can be understood as a bucket, with a width of 1000 MW. The process of determining the first power load group to which the historical power load of each target historical moment belongs can be understood as bucketing, that is, assigning the historical power load of each target historical moment to a corresponding bucket.

[0125] In some embodiments, determining the first time group to which the current time belongs based on the predicted power load at the current time to obtain the target time group may include: determining the target first power load group range to which the predicted power load belongs based on the predicted power load at the current time and the range of each first power load group; and taking the first time group to which the target historical time belongs within the target first power load group range as the target time group.

[0126] In some embodiments, determining the median of the historical electricity spot price for each target historical moment belonging to a target time group includes: for each target time group, if the number of target historical moments in the target time group is odd, using the historical electricity spot price of the middle target historical moment as the median; if the number of target historical moments in the target time group is even, using the average of the historical electricity spot prices of the two middle target historical moments as the median.

[0127] In some embodiments, where there is no historical power load in at least one first power load grouping range, the median of the historical spot price of electricity corresponding to at least one target historical time can be determined as the predicted spot price of electricity at the current time.

[0128] In this embodiment, by determining the median of the historical electricity spot price of each target historical time in the target time group to which the current time belongs, the clearing pattern of the historical time period can be reflected more accurately, thereby the predicted electricity spot price at the current time can be determined more accurately.

[0129] Based on the technical solutions of the above embodiments, this application also provides another optional embodiment. In this optional embodiment, when the target load range is the second load range and the current time is outside the preset time window, or when the target load range is the third load range, the reference time is refined into multiple times. Accordingly, the steps for determining the predicted spot price of electricity at the current time are refined.

[0130] See Figure 4 The steps for determining the predicted spot price of electricity, as shown, include:

[0131] S410, group the reference times according to the target time period to which each reference time belongs, to obtain at least one second time group.

[0132] The target time period to which each reference time belongs can be determined from at least one preset time period. Optionally, the preset time period may include at least one of a morning time period and an afternoon time period.

[0133] In some embodiments, reference times belonging to the same target time period can be determined as the same second time group.

[0134] S420: For each second time group, based on the historical power load of each reference time in the second time group, group the reference times in the second time group to obtain at least one time subgroup corresponding to the second time group.

[0135] In some embodiments, at least one second power load grouping range may be obtained; for each reference time in each second time group, the second power load grouping range to which the historical power load of the reference time belongs may be determined; and reference times belonging to the same second power load grouping range may be divided into a time subgroup.

[0136] The maximum load difference within each second power load group is the same; for example, the maximum load difference within each second power load group is 2000MW. Each second power load group can be understood as a bucket, with a width of 2000MW. The process of determining the second power load group to which the historical power load of each reference time belongs for each reference time within each second time group can be understood as bucketing, that is, assigning the historical power load of each reference time to the corresponding bucket.

[0137] S430, determine the median of the historical spot electricity price at each reference time in each subgroup corresponding to the second time group.

[0138] In some embodiments, for each time subgroup, if the number of reference times within the time subgroup is odd, the historical spot price of electricity at the middle reference time is used as the median; if the number of reference times within the time subgroup is even, the average of the historical spot prices of electricity at the middle two reference times is used as the median.

[0139] S440 determines the predicted spot electricity price for the current moment based on the determined medians.

[0140] In some embodiments, the weighted sum of the medians can be used as the predicted spot price of electricity at the current moment.

[0141] In some embodiments, determining the forecasted spot electricity price at the current time based on the determined medians includes: for each second time group, determining the correspondence between the spot electricity price and the electricity load corresponding to the second time group based on the medians corresponding to the second time group; and determining the forecasted spot electricity price at the current time based on the second time group to which the current time belongs, the forecasted electricity load at the current time, and the correspondence between the spot electricity price and the electricity load corresponding to each second time group.

[0142] In some embodiments, for each second time group, linear interpolation and moving average smoothing can be performed on each median corresponding to the second time group to obtain the correspondence between the spot price of electricity and the electricity load corresponding to the second time group.

[0143] In some embodiments, the correspondence between the spot price of electricity and the load of electricity corresponding to each second time group can be determined based on the correspondence between the spot price of electricity and the load of electricity corresponding to each second time group. Then, the spot price of electricity corresponding to the predicted load of the current time can be found in the correspondence between the spot price of electricity and the load of electricity corresponding to the second time group to which the current time belongs, and used as the predicted spot price of electricity for the current time.

[0144] In this embodiment, by determining the predicted electricity spot price at the current moment based on the median of the historical electricity spot price at each reference moment in each subgroup corresponding to each second time moment group, the clearing pattern over historical time periods can be reflected more accurately, thereby allowing for a more accurate determination of the predicted electricity spot price at the current moment.

[0145] Based on the technical solutions of the above embodiments, this application also provides another optional embodiment, in which the method for predicting electricity spot prices is further refined.

[0146] See Figure 5 The methods for forecasting spot electricity prices shown include:

[0147] S501, obtain the historical power load of each historical time within the historical period corresponding to the current time, as well as the meteorological data of the current time.

[0148] Among them, the historical time period and the current time period to which the current moment belongs satisfy the preset time relationship.

[0149] S502, determine at least one candidate load range based on each historical power load.

[0150] Among them, at least one candidate load interval includes a first load interval, a second load interval, and a third load interval.

[0151] Among them, the power load in the first load interval is less than the power load in the second load interval; the power load in the second load interval is less than the power load in the third load interval.

[0152] S503 determines the predicted power load for the current moment based on meteorological data and historical power loads.

[0153] S504: Determine the target load interval to which the predicted power load belongs from each candidate load interval.

[0154] S505, when the target load range is the first load range, the preset electricity spot price is determined as the predicted electricity spot price at the current moment.

[0155] S506, if the target load interval is the second load interval and the current time is within the preset time window, select at least one target historical time from each historical time according to the power load influence factor corresponding to the current time.

[0156] Among them, each target historical moment is located within the historical period and the corresponding historical time window of the preset time window, and the power load impact factor corresponding to each target historical moment matches the power load impact factor corresponding to the current moment.

[0157] S507, based on the historical power load of each target historical moment, group each target historical moment to obtain at least one first moment group.

[0158] S508: Based on the predicted power load at the current moment, determine the first time group to which the current moment belongs, so as to obtain the target time group.

[0159] S509, determine the median of the historical electricity spot price for each target historical time belonging to the target time group, and use the determined median as the predicted electricity spot price for the current time.

[0160] S510, when the target load interval is the second load interval and the current time is outside the preset time window, or when the target load interval is the third load interval, determine the similarity between the predicted power load at the current time and the historical power load at each historical time.

[0161] S511, Select each historical moment with a similarity greater than the similarity threshold to obtain the reference moment.

[0162] S512, group the reference times according to the target time period to which each reference time belongs, to obtain at least one second time group.

[0163] S513, for each second time group, based on the historical power load of each reference time in the second time group, group each reference time in the second time group to obtain at least one time subgroup corresponding to the second time group.

[0164] S514, determine the median of the historical spot electricity price at each reference time in each subgroup corresponding to the second time group.

[0165] S515, for each second time group, determine the correspondence between the spot electricity price and the electricity load corresponding to the second time group based on the median of each second time group.

[0166] S516. Based on the second time group to which the current time belongs, the predicted power load at the current time, and the correspondence between the spot power price and the power load for each second time group, determine the predicted spot power price at the current time.

[0167] In some embodiments, the predicted power load and the predicted spot price of electricity at the current moment can be output.

[0168] In some embodiments, at least one of the following methods—mean absolute error, root mean square error, and mean absolute percentage error—can be used to assess the forecast accuracy based on the forecasted electricity load and the forecasted spot electricity price at the current moment.

[0169] In some embodiments, load / electricity price forecast comparison charts, forecast accuracy diagrams, load-electricity price mapping relationship diagrams, etc., can be output.

[0170] In some embodiments, result files can be exported and stored or reported based on the current forecast power load and forecast spot power price.

[0171] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0172] Based on the same inventive concept, this application also provides an electricity spot price prediction device for implementing the above-mentioned method for predicting electricity spot prices. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations in one or more embodiments of the electricity spot price prediction device provided below can be found in the limitations of the electricity spot price prediction method described above, and will not be repeated here.

[0173] In one exemplary embodiment, such as Figure 6 As shown, a device for predicting spot electricity prices is provided, comprising: an acquisition module 610, a first determination module 620, and a second determination module 630, wherein:

[0174] The acquisition module 610 is used to acquire the historical power load of each historical time within the historical time period corresponding to the current time, as well as the meteorological data of the previous time; wherein, the historical time period and the current time period to which the previous time period belongs satisfy a preset time relationship;

[0175] The first determining module 620 is used to determine the target load range to which the predicted power load at the current moment belongs, based on meteorological data and historical power loads.

[0176] The second determining module 630 is used to determine the predicted spot price of electricity at the current moment based on the target load range.

[0177] In one embodiment, the first determining module 620 is specifically configured to: determine at least one candidate load interval based on each historical power load; determine the predicted power load at the current moment based on meteorological data and each historical power load; and determine the target load interval to which the predicted power load belongs from each candidate load interval.

[0178] In one embodiment, the second determining module 630 is specifically configured to: determine the preset electricity spot price as the predicted electricity spot price at the current moment when the target load range is the first load range; determine the predicted electricity spot price at the current moment based on the electricity load influencing factor corresponding to the current moment and the historical electricity spot prices at each historical moment when the target load range is the second load range and the current moment is outside the preset time window, or when the target load range is the third load range; determine a reference moment from each historical moment based on the predicted electricity load at the current moment and the historical electricity load at each historical moment; and determine the predicted electricity spot price at the current moment based on the historical electricity spot price at the reference moment; wherein, each electricity load in the first load range is less than each electricity load in the second load range; and each electricity load in the second load range is less than each electricity load in the third load range.

[0179] In one embodiment, the second determining module 630 is specifically used to: select at least one target historical moment from each historical moment according to the power load impact factor corresponding to the current moment; wherein each target historical moment is located within the historical time period and the historical time window corresponding to the preset time window, and the power load impact factor corresponding to each target historical moment matches the power load impact factor corresponding to the current moment; and determine the predicted power spot price at the current moment according to the historical power spot price of each target historical moment.

[0180] In one embodiment, the second determining module 630 is specifically configured to: group each target historical time according to the historical power load of each target historical time to obtain at least one first time group; determine the first time group to which the current time belongs according to the predicted power load of the current time to obtain a target time group; determine the median of the historical electricity spot price of each target historical time belonging to the target time group, and determine the determined median as the predicted electricity spot price of the current time.

[0181] In one embodiment, the second determining module 630 is specifically used to: determine the similarity between the predicted power load at the current moment and the historical power load at each historical moment; and select each historical moment with a similarity greater than a similarity threshold to obtain a reference moment.

[0182] In one embodiment, there are multiple reference times; the second determining module 630 is specifically used for: grouping each reference time according to the target time period to which each reference time belongs, to obtain at least one second time group; for each second time group, grouping each reference time in the second time group according to the historical power load of each reference time in the second time group, to obtain at least one time subgroup corresponding to the second time group; and determining the median of the historical spot price of each reference time in each time subgroup corresponding to the second time group; and determining the predicted spot price of the current time based on the determined medians.

[0183] In one embodiment, the second determining module 630 is specifically used to: for each second time group, determine the correspondence between the spot electricity price and the electricity load corresponding to the second time group based on the median of each second time group; and determine the predicted spot electricity price at the current time based on the second time group to which the current time belongs, the predicted electricity load at the current time, and the correspondence between the spot electricity price and the electricity load corresponding to each second time group.

[0184] The modules in the aforementioned electricity spot price forecasting device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0185] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores historical power load data, meteorological data, predicted power load data, and predicted spot electricity prices for various historical periods. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When executed by the processor, the computer program implements a method for predicting spot electricity prices.

[0186] Those skilled in the art will understand that Figure 7 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 computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0187] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method for predicting spot electricity prices provided in any of the above embodiments.

[0188] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the electricity spot price prediction method provided in any of the above embodiments.

[0189] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method for predicting spot electricity prices provided in any of the above embodiments.

[0190] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0191] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0192] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for predicting spot electricity prices, characterized in that, The method includes: Obtain the historical power load for each historical time period within the historical time period corresponding to the current time, as well as the meteorological data for the current time; wherein the historical time period and the current time period to which the current time belongs satisfy a preset time relationship; Based on the meteorological data and the historical power loads, determine the target load range to which the predicted power load at the current moment belongs; Based on the target load range, determine the predicted spot electricity price for the current moment.

2. The method according to claim 1, characterized in that, The step of determining the target load range to which the predicted power load at the current moment belongs, based on the meteorological data and the historical power loads, includes: Based on the aforementioned historical power loads, at least one candidate load range is determined; Based on the meteorological data and the historical power loads, determine the predicted power load for the current moment; From each of the candidate load intervals, the target load interval to which the predicted power load belongs is determined.

3. The method according to claim 2, characterized in that, The at least one candidate load interval includes a first load interval, a second load interval, and a third load interval; determining the predicted spot electricity price at the current moment based on the target load interval includes: When the target load range is the first load range, the preset spot electricity price is determined as the predicted spot electricity price at the current moment; When the target load range is the second load range and the current time is within a preset time window, the predicted electricity spot price at the current time is determined based on the power load impact factor corresponding to the current time and the historical electricity spot price at each historical time. If the target load range is the second load range and the current time is outside the preset time window, or if the target load range is the third load range, a reference time is determined from each of the historical times based on the predicted power load at the current time and the historical power load at each of the historical times; and the predicted power spot price at the current time is determined based on the historical power spot price at the reference time. Wherein, each power load in the first load interval is less than each power load in the second load interval; and each power load in the second load interval is less than each power load in the third load interval.

4. The method according to claim 3, characterized in that, The step of determining the predicted electricity spot price at the current moment based on the electricity load impact factor corresponding to the current moment and the historical electricity spot prices at each of the historical moments includes: Based on the power load impact factor corresponding to the current time, at least one target historical time is selected from each of the historical times; wherein each target historical time is located within the historical time period and the historical time window corresponding to the preset time window, and the power load impact factor corresponding to each target historical time matches the power load impact factor corresponding to the current time. The predicted spot price of electricity at the current moment is determined based on the historical spot price of electricity at each of the target historical moments.

5. The method according to claim 4, characterized in that, The step of determining the predicted electricity spot price at the current moment based on the historical electricity spot prices at each of the target historical moments includes: Based on the historical power load of each target historical moment, the target historical moments are grouped to obtain at least one first moment group; Based on the predicted power load at the current moment, determine the first time group to which the current moment belongs, so as to obtain the target time group; The median of the historical spot electricity price for each target historical time belonging to the target time group is determined, and the determined median is used as the predicted spot electricity price for the current time.

6. The method according to claim 3, characterized in that, The step of determining a reference time from the historical times based on the predicted power load at the current time and the historical power load at each of the historical times includes: Determine the similarity between the predicted power load at the current moment and the historical power load at each of the historical moments; Select the historical moments with a similarity greater than the similarity threshold to obtain the reference moments.

7. The method according to claim 3 or 6, characterized in that, The reference times can be multiple; determining the predicted electricity spot price for the current time based on the historical electricity spot prices at the reference times includes: Based on the target time period to which each reference time belongs, the reference times are grouped to obtain at least one second time group; For each of the second time groups, based on the historical power load of each reference time in the second time group, the reference times in the second time group are grouped to obtain at least one time subgroup corresponding to the second time group; and, Determine the median of the historical spot electricity price at each reference time in each of the time subgroups corresponding to the second time group; The predicted spot electricity price for the current moment is determined based on the determined medians.

8. The method according to claim 7, characterized in that, The step of determining the predicted spot electricity price at the current moment based on the determined medians includes: For each second time group, the relationship between the spot electricity price and the electricity load corresponding to the second time group is determined based on the median values ​​of the second time group. The predicted electricity spot price for the current time is determined based on the second time group to which the current time belongs, the predicted electricity load for the current time, and the correspondence between the electricity spot price and the electricity load for each second time group.

9. A device for predicting spot electricity prices, characterized in that, The device includes: The acquisition module is used to acquire the historical power load of each historical time within the historical time period corresponding to the current time, as well as the meteorological data of the previous time; wherein the historical time period and the current time period to which the previous time belongs satisfy a preset time relationship; The first determining module is used to determine the target load range to which the predicted power load at the current moment belongs, based on the meteorological data and the historical power loads. The second determining module is used to determine the predicted spot electricity price at the current moment based on the target load range.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, 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 8.

Images