Electric power prediction method and system, computer equipment and computer readable storage medium
By constructing seasonal differences from returning to the average model of sliding, combined with historical and real -time data, the accuracy and reliability of existing power prediction methods have been solved, and more accurate power prediction and better energy management support have been achieved.
Patent Information
- Application Number
- CN202411444689.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-16
- Publication Date
- 2025-01-28
AI Technical Summary
The existing power forecasting methods have single reference data, low prediction accuracy, poor prediction reliability, and cannot effectively support peak load prediction, energy saving management and power emergency management.
By obtaining historical power data and environmental data, pre -processing, regulatory processing, and feature extraction are performed, and seasonal differences are constructed from returning to the average model of sliding, and electricity prediction is performed in combination with real -time data.
It improves the accuracy and reliability of power prediction, can better predict peak loads, and support energy conservation management and power emergency management.
Smart Images
Figure CN119377572A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of electric power technology, and in particular to an electric power prediction method, system, computer device and computer-readable storage medium. Background Art
[0002] In parks or large-scale equipment, real-time monitoring and prediction of power consumption is the key to improving energy efficiency and reducing costs. At present, it mainly relies on simple sensors to collect power data, and conducts real-time monitoring and prediction based on the collected power data. However, the current monitoring and prediction methods, especially the prediction methods, generally have a single reference data, low prediction accuracy, and poor prediction reliability, and cannot provide effective technical support for peak load prediction, energy-saving management, and power emergency management. Summary of the invention
[0003] Based on this, it is necessary to provide a power prediction method, system, computer device and computer-readable storage medium to address the above technical problems.
[0004] In a first aspect, the present application provides a power prediction method, the power prediction method comprising:
[0005] Acquiring historical data, wherein the historical data includes historical power data and historical environmental data;
[0006] Preprocessing the historical data to remove noise and outliers in the historical data;
[0007] Normalize the preprocessed historical data;
[0008] Perform feature extraction based on normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features, and historical trend features;
[0009] Building a power forecasting model based on the historical key features, the power forecasting model comprising a seasonal difference autoregressive moving average model;
[0010] Acquiring real-time data, wherein the real-time data includes real-time power data and real-time environmental data;
[0011] Power forecasting is performed on power demand based on the real-time data and the power forecasting model.
[0012] In one embodiment, obtaining historical data includes:
[0013] Collect the historical data using sensors and smart meters;
[0014] The historical data is received based on message queue technology.
[0015] In one embodiment, after receiving the historical data based on the message queue technology and before preprocessing the historical data, the method further includes:
[0016] Performing data verification and preliminary filtering on the historical data;
[0017] Outliers in the historical data are marked.
[0018] In one embodiment, the constructing of the power forecasting model based on the historical key features includes:
[0019] Visualizing the historical key features to obtain a time series graph, and determining the length of the seasonal cycle based on the time series graph;
[0020] Determine the order of non-seasonal differences and the order of seasonal differences based on the time series plot or using a unit root test;
[0021] Based on the historical key features, using autocorrelation plots and partial autocorrelation plots to determine the order of the non-seasonal autoregressive term, the order of the non-seasonal moving average term, the order of the seasonal autoregressive term, and the order of the seasonal average term;
[0022] The power forecasting model is obtained by fitting based on the length of the seasonal cycle, the order of the non-seasonal difference, the order of the seasonal difference, the order of the non-seasonal autoregressive term, the order of the non-seasonal moving average term, the order of the seasonal autoregressive term and the order of the seasonal average term.
[0023] In one embodiment, after obtaining the power prediction model, the method further includes:
[0024] Checking whether the model residual of the power prediction model is white noise;
[0025] If not, the parameters of the power prediction model are adjusted.
[0026] In one embodiment, the power forecasting of power demand based on the real-time data and the power forecasting model includes:
[0027] Preprocessing the real-time data to remove noise and outliers in the real-time data;
[0028] Normalize the preprocessed real-time data;
[0029] Perform feature extraction based on normalized real-time data to obtain real-time key features;
[0030] The real-time key features are input into the power prediction model to perform power prediction on power demand.
[0031] In a second aspect, the present application further provides a power prediction system, the power prediction system comprising:
[0032] A data acquisition module, used to acquire historical data and real-time data; the historical data includes historical power data and historical environmental data; the real-time data includes real-time power data and real-time environmental data;
[0033] A preprocessing module, used for at least preprocessing the historical data to remove noise and abnormal values in the historical data;
[0034] A normalization processing module, used for performing normalization processing on at least the pre-processed historical data;
[0035] A feature extraction module is used to extract features based on the normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features and historical trend features;
[0036] A model building module is used to build an electricity forecasting model based on the historical key features, and the electricity forecasting model includes a seasonal difference autoregressive sliding average model; the electricity forecasting model is used to make electricity forecasts on electricity demand based on the real-time data.
[0037] In one of the embodiments, the data acquisition module includes: a data acquisition unit and a data receiving unit;
[0038] The data collection unit includes a plurality of sensors and a smart meter, which are used to collect the historical data and the real-time data;
[0039] The data receiving unit receives the historical data and the real-time data based on a message queue technology.
[0040] In a third aspect, the present application further provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the power prediction method described in the first aspect are implemented.
[0041] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the power prediction method described in the first aspect are implemented.
[0042] In the above-mentioned power forecasting method, system, computer equipment and computer-readable storage medium, the historical data and real-time data obtained include not only power data but also environmental data. The extracted historical key features also include temperature-related features, holiday features, seasonal indicator features and historical trend features. The power forecasting model constructed based on this, and the power forecasting of power demand based on the power forecasting model, fully consider the seasonal characteristics and holiday characteristics, have high prediction accuracy and good prediction reliability, and can provide effective technical support for peak load prediction, energy-saving management and power emergency management. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 A schematic diagram of a flow chart of a power prediction method is provided in one embodiment of the present application;
[0044] Figure 2 A structural block diagram of a power prediction system provided in another embodiment of the present application;
[0045] Figure 3 The internal structure diagram of a computer device provided in yet another embodiment. DETAILED DESCRIPTION
[0046] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0047] In one embodiment, see Figure 1 , the present application provides a power prediction method, which may include the following steps: S10~S70.
[0048] Step S10: Acquire historical data, where the historical data includes historical power data and historical environmental data.
[0049] As an example, step S10 may include the following steps S101 to S102.
[0050] S101: Collect the historical data using sensors and smart meters.
[0051] S102: Receive the historical data based on message queue technology.
[0052] As an example, you can choose current sensors, voltage sensors, and power sensors with high precision and high stability. According to the layout and power consumption characteristics of each electrical equipment in the park, reasonably plan the installation location of each sensor to ensure that all sensors can fully cover key electrical equipment and power lines to ensure the comprehensiveness of data collection. When installing each sensor, it is necessary to consider the impact of environmental factors on sensor readings. If necessary, take corresponding protective measures, such as installing dustproof and waterproof casings on each sensor.
[0053] As an example, the smart meter can be installed at the power inlet of each power-consuming device to record the power consumption data of each power-consuming device in real time. The smart meter needs to have a remote communication function and can support wireless or wired network connection to facilitate data transmission.
[0054] As an example, a time synchronization mechanism may be used to ensure that each of the smart meters and each of the sensors collects data at the same time point, so as to reduce the impact of time deviation on data analysis.
[0055] As an example, the collection frequency of each sensor and each smart meter can be set according to actual needs. For example, for key power-consuming equipment, it can be set to collect once every minute, and for general equipment, it can be set to collect once every hour.
[0056] As an example, the historical data collected by the sensor and the smart meter can be transmitted to a central data processing platform; specifically, the historical data can be transmitted based on a wireless network. The central data processing platform can be provided with an efficient data receiving module, which can process data streams from multiple smart meters and sensors in parallel; the data receiving module can receive the historical data based on message queue technology, such as Kafka or RabbitMQ. The use of message queue technology to receive data can buffer or asynchronously process a large amount of real-time data to ensure that data is not lost, and ensure the stability and high throughput of data reception.
[0057] As an example, after step S10 and before step S20, the following steps may also be included: performing data verification and preliminary filtering on the historical data; marking abnormal values in the historical data. Specifically, a threshold may be set and abnormal values in the historical data may be marked based on the threshold to facilitate subsequent cleaning of abnormal values.
[0058] In one implementation, taking a key power-consuming equipment in the park as an example, a smart meter with remote communication function is installed at the power inlet of the key power-consuming equipment, and current sensors, voltage sensors and power sensors are installed at key nodes of the power line; the smart meter and each sensor are set to collect data once a minute, and the collected data is transmitted to the central data processing platform through the wireless network; after the central data processing platform receives the data based on the message queue technology, it first performs time alignment, then performs data verification and preliminary filtering on the data, and marks abnormal values.
[0059] Through the above scheme, it can be ensured that the historical power data and historical environmental data of each electrical equipment in the park are collected in real time, accurately and comprehensively, providing reliable basic data for subsequent data processing and analysis.
[0060] Step S20: preprocessing the historical data to remove noise and abnormal values in the historical data.
[0061] As an example, noise and outliers in the historical data can be identified and removed based on an automated cleaning algorithm. The automated cleaning algorithm can include a Z-score or IQR (interquartile range) test based on statistical methods to identify outliers in the historical data.
[0062] As an example, corresponding to the detected outliers, a deletion, replacement or correction strategy can be adopted according to the specific situation.
[0063] In one example, for outliers that are beyond the normal range, the average value of the preceding and following data points may be used for replacement.
[0064] Step S30: normalizing the preprocessed historical data.
[0065] As an example, the normalization method may include minimum-maximum normalization or Z-score normalization; through normalization, the influence of different dimensions and sizes on the subsequent power forecasting model training can be eliminated. Through normalization, the pre-processed historical data can be scaled to the [0, 1] interval for subsequent processing.
[0066] Step S40: performing feature extraction based on the normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features and historical trend features.
[0067] As an example, the Fast Fourier Transform (FFT) may be used, but is not limited to, to extract the periodic features of the data, or to construct a time window to calculate a moving average to capture trend features.
[0068] As an example, in the process of extracting the time window features, daily electricity consumption, weekly electricity consumption and monthly electricity consumption can be extracted; for daily electricity consumption, the daily average electricity consumption on weekdays and weekends and holidays can be calculated; for weekly electricity consumption, the total weekly electricity consumption can be calculated, that is, the sum of weekly electricity consumption; for monthly electricity consumption, the historical average electricity consumption in the same month can be calculated. Since seasonal changes are often reflected in these periodic time windows, the time window features can capture the periodic fluctuations in electricity consumption. The above-mentioned electricity consumption can be the electricity consumption corresponding to each electrical equipment in the park, or it can be the electricity consumption corresponding to all electrical equipment in the park.
[0069] As an example, in the process of extracting the temperature-related features, the average temperature, maximum temperature and minimum temperature of the park can be extracted as temperature-related features. Since temperature has a significant impact on electricity consumption, especially in areas with significant seasonal changes, temperature-related features are a crucial feature. Specifically, the average temperature over the past few days and the maximum temperature corresponding to the peak electricity consumption can be extracted as the temperature-related features.
[0070] As an example, during holidays, the electricity consumption of the park will usually change compared to weekdays, so we can create a holiday feature that indicates whether a specific date is a holiday. Specifically, we can create a binary variable that is 1 if a certain day is a holiday and 0 otherwise.
[0071] As an example, seasonal indicator features, such as seasonal index, can be directly extracted to represent seasonal changes. Specifically, the average historical electricity consumption in the same month of each year can be calculated, and the corresponding seasonal index can be calculated using the obtained historical electricity consumption average; how to use the obtained historical electricity consumption average to calculate the corresponding seasonal index is known to those skilled in the art and will not be described here.
[0072] As an example, the historical power consumption trend in the same time period in the past few years can be extracted as the historical trend feature. Specifically, for each prediction point, the power consumption on the same date in the past three years can be extracted, and the historical power consumption trend can be obtained based on the extracted power consumption.
[0073] In one embodiment, assuming that the electricity consumption in July is being predicted, the average temperature in June and July, the total electricity consumption in June, the electricity consumption in July last year, and the average electricity consumption in July in previous years can be extracted. In addition, if there are holidays in July, the corresponding dates can also be used as a special holiday feature.
[0074] Through the above feature extraction method, the subsequent power forecasting model can better capture seasonal changes.
[0075] Step S50: constructing a power forecasting model based on the historical key features, wherein the power forecasting model includes a Seasonal Autoregressive Integrated Moving Average (SARIMA) model.
[0076] As an example, the seasonal difference autoregressive moving average model is a classic statistical model for time series forecasting, which is particularly suitable for electricity demand forecasting with seasonal characteristics.
[0077] As an example, the seasonal difference autoregressive moving average model can generally be expressed as SARIMA(p, d, q)(P, D, Q, S), where p is the order of the non-seasonal autoregressive term, d is the order of the non-seasonal difference, q is the order of the non-seasonal moving average term, P is the order of the seasonal autoregressive term, D is the order of the seasonal difference, Q is the order of the seasonal moving average term, and S is the length of the seasonal cycle; where p, d, and q are the non-seasonal parameters of the seasonal difference autoregressive moving average model, and P, D, Q, and S are the seasonal parameters of the seasonal difference autoregressive moving average model.
[0078] As an example, the formula for the seasonal difference autoregressive moving average model can be expressed as:
[0079] (1-φ1L-φ2L 2 -...-φ p L P )(1-Φ1L S -Φ2L 2S -...-Φ p L PS )(1-L) d (1-L S ) D yt
[0080] =c+q1L+q2L 2 +...+q q L q +Q1L S +Q2L 2S ...+Q Q L QS +εt
[0081] Among them, 1-φ1L-φ2L 2 -...-φ p L P is a non-seasonal autoregressive term, 1-Φ1L S -Φ2L 2S -...-Φ p L PSis the seasonal autoregressive term, (1-L) d is the non-seasonal difference, (1-L S ) D is the seasonal difference; φi is a coefficient or parameter of the seasonal difference autoregressive moving average model, which is related to time series analysis, φ1, φ2, …, φ p Represent different coefficients respectively; L is the lag operator, which is used to represent the lag term in the time series; Φ i Similar to φ i , are the coefficients in the seasonal difference autoregressive moving average model, which can be used for the seasonal lag terms in the seasonal time series, Φ1, Φ2, …, Φ p They represent different coefficients respectively; S is the length of the seasonal cycle; d is the order of non-seasonal differences; D is the order of seasonal differences; yt is the observed value of the time series at time point t; c is the constant term in the seasonal difference autoregressive moving average model; θ i are the coefficients in the seasonal difference autoregressive moving average model, which are related to the lagged terms of the non-seasonal part, θ1, θ2, θ q are different coefficients; Θ i Similar to θ i , coefficients that can be used for the seasonal part, Θ1, Θ2, Θ Q are different coefficients; q is the order of the non-seasonal moving average term; Q is the order of the seasonal moving average term; εt is the error term, that is, the unexplained random fluctuations in the seasonal difference autoregressive moving average model.
[0082] As an example, step S50 may include the following steps: S501 to S504.
[0083] S501: Visualize the historical key features to obtain a time series graph, and determine the length S of the seasonal cycle based on the time series graph. Specifically, the historical key features are time series data, and by visualizing them, the trend, seasonality and periodicity can be observed; if the historical key features show an annual periodicity, the length S of the seasonal cycle = 12.
[0084] S502: Based on the time series graph or using a unit root test (such as an ADF test), determine the order d of the non-seasonal difference and the order D of the seasonal difference. Through this step, the sequence can be made stable.
[0085] S503: Based on the historical key features, the order p of the non-seasonal autoregressive term, the order q of the non-seasonal moving average term, the order P of the seasonal autoregressive term and the order Q of the seasonal average term are determined using the autocorrelation diagram (ACF) and the partial autocorrelation diagram (PACF).
[0086] S504: The power forecasting model is obtained by fitting parameters such as the length S of the seasonal cycle, the order d of the non-seasonal difference, the order D of the seasonal difference, the order p of the non-seasonal autoregressive term, the order q of the non-seasonal moving average term, the order P of the seasonal autoregressive term and the order Q of the seasonal average term.
[0087] As an example, after step S50, the following step may also be included: evaluating the parameters of the seasonal difference autoregressive moving average model by maximizing the likelihood function.
[0088] As an example, after step S50, the following steps may also be included: checking whether the model residual of the power prediction model is white noise; if not, adjusting the parameters of the power prediction model.
[0089] In one embodiment, if it is determined through analysis that the seasonal cycle of electricity demand data is 12 months, the series is stable after one non-seasonal difference and one seasonal difference, and the ACF graph and the PACF graph indicate that the order p of the non-seasonal autoregressive term is 2, the order q of the non-seasonal moving average term is 1, and the order P of the seasonal autoregressive term and the order of the seasonal moving average term are both 1, then a SARIMA (2,1,1) (1,1,1,12) model can be fitted.
[0090] Step S60: Acquire real-time data, where the real-time data includes real-time power data and real-time environmental data.
[0091] As an example, the real-time data can be collected based on the sensors and smart meters in step S10. The method for collecting and transmitting the real-time data is similar to the method for collecting and transmitting the historical data. Please refer to step S10 for details, which will not be repeated here.
[0092] Step S70: Perform power forecasting on power demand based on the real-time data and the power forecasting model.
[0093] As an example, step S70 may include the following steps: S701 to S704.
[0094] S701: preprocessing the real-time data to remove noise and abnormal values in the real-time data;
[0095] S702: normalizing the preprocessed real-time data;
[0096] S703: extracting features based on the normalized real-time data to obtain real-time key features;
[0097] S704: Input the real-time key features into the power prediction model to make power prediction on power demand.
[0098] As an example, in step S701, the method for preprocessing the real-time data may refer to the method for preprocessing the historical data in step S20, which will not be described again here.
[0099] As an example, in step S702, the method for normalizing the preprocessed real-time data can refer to the method for normalizing the preprocessed historical data in step S30, which will not be repeated here.
[0100] As an example, in step S703, the specific method for extracting features based on normalized real-time data to obtain real-time key features can refer to the method for extracting features based on normalized historical data in step S40, which will not be repeated here.
[0101] As an example, the real-time key feature has the same format as the historical key feature.
[0102] As an example, in step S704, the real-time key features are input into the power prediction model to perform power prediction on the power demand in a future period of time.
[0103] In one embodiment, at fixed intervals (for example, 5 minutes), the smart meter and the sensor transmit the collected real-time data to a central data processing platform. The received real-time data is quickly cleaned to exclude obviously erroneous outliers. The real-time data after quick cleaning is normalized so that it is in the same numerical range as the normalized historical data. According to the input requirements of the seasonal difference autoregressive moving average model, the real-time key features in the real-time data are extracted, such as timestamps and power consumption, etc. The real-time key features are input into the seasonal difference autoregressive moving average model for prediction. For example, the real-time data input into the seasonal difference autoregressive moving average model may include power consumption at the last few time points, and the seasonal difference autoregressive moving average model outputs predicted power consumption per hour in the next 24 hours, etc.
[0104] Through the above steps, accurate electricity demand forecasts can be generated in real time, providing an important basis for subsequent energy management and scheduling.
[0105] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0106] Based on the same inventive concept, the embodiment of the present application also provides a power forecasting system for implementing the power forecasting method involved above. The implementation scheme for solving the problem provided by the power forecasting system is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more power forecasting system embodiments provided below can refer to the limitations on the power forecasting method above, and will not be repeated here.
[0107] In another embodiment, see Figure 2 The present application also provides a power forecasting system, which may include: a data acquisition module 10, a preprocessing module 20, a normalization processing module 30, a feature extraction module 40 and a model construction module 50; wherein the data acquisition module 10 is used to acquire historical data and real-time data; the historical data includes historical power data and historical environmental data; the real-time data includes real-time power data and real-time environmental data; the preprocessing module 20 is used to at least preprocess the historical data to remove noise and abnormal values in the historical data; the normalization processing module 30 is used to at least normalize the preprocessed historical data; the feature extraction module 40 is used to extract features based on the normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features and historical trend features; the model construction module 50 is used to construct a power forecasting model based on the historical key features, and the power forecasting model includes a seasonal difference autoregressive sliding average model; the power forecasting model is used to forecast power demand based on the real-time data.
[0108] In the power forecasting system of the present invention, the acquired historical data and real-time data include not only power data but also environmental data. The extracted historical key features also include temperature-related features, holiday features, seasonal index features and historical trend features. The power forecasting model constructed based on this and the power forecasting of power demand based on the power forecasting model fully consider the seasonal characteristics and holiday characteristics, have high prediction accuracy and good prediction reliability, and can provide effective technical support for peak load prediction, energy-saving management and power emergency management.
[0109] As an example, the data acquisition module 10 may include: a data acquisition unit (not shown) and a data receiving unit (not shown); the data acquisition unit includes multiple sensors and smart meters, and the sensors and the smart meters are used to collect the historical data and the real-time data; the data receiving unit receives the historical data and the real-time data based on message queue technology.
[0110] Each module in the above power forecasting system can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.
[0111] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 3As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface, the display unit and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a power prediction method is implemented. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse, etc.
[0112] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0113] In one embodiment, a computer device is further provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above-mentioned power prediction method embodiment when executing the computer program.
[0114] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned power prediction method embodiment are implemented.
[0115] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile 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), magnetoresistive 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. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0116] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, 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 specification.
[0117] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. A power forecasting method, characterized in that: include: Acquiring historical data, wherein the historical data includes historical power data and historical environmental data; Preprocessing the historical data to remove noise and outliers in the historical data; Normalize the preprocessed historical data; Perform feature extraction based on normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features, and historical trend features; Building a power forecasting model based on the historical key features, the power forecasting model comprising a seasonal difference autoregressive moving average model; Acquiring real-time data, wherein the real-time data includes real-time power data and real-time environmental data; Power forecasting is performed on power demand based on the real-time data and the power forecasting model.
2. The power forecasting method according to claim 1, characterized in that: The obtaining of historical data includes: Collect the historical data using sensors and smart meters; The historical data is received based on message queue technology.
3. The power forecasting method according to claim 2, characterized in that: After receiving the historical data based on the message queue technology and before preprocessing the historical data, the method further includes: Performing data verification and preliminary filtering on the historical data; Outliers in the historical data are marked.
4. The power forecasting method according to claim 1, characterized in that: The constructing of the power forecasting model based on the historical key features includes: Visualizing the historical key features to obtain a time series graph, and determining the length of the seasonal cycle based on the time series graph; Determine the order of non-seasonal differences and the order of seasonal differences based on the time series plot or using a unit root test; Based on the historical key features, using autocorrelation plots and partial autocorrelation plots to determine the order of the non-seasonal autoregressive term, the order of the non-seasonal moving average term, the order of the seasonal autoregressive term, and the order of the seasonal average term; The power forecasting model is obtained by fitting based on the length of the seasonal cycle, the order of the non-seasonal difference, the order of the seasonal difference, the order of the non-seasonal autoregressive term, the order of the non-seasonal moving average term, the order of the seasonal autoregressive term and the order of the seasonal average term.
5. The power forecasting method according to claim 4, characterized in that: After obtaining the power prediction model, the method further includes: Checking whether the model residual of the power prediction model is white noise; If not, the parameters of the power prediction model are adjusted.
6. The power forecasting method according to any one of claims 1 to 5, characterized in that: The performing power forecasting on power demand based on the real-time data and the power forecasting model comprises: Preprocessing the real-time data to remove noise and outliers in the real-time data; Normalize the preprocessed real-time data; Perform feature extraction based on normalized real-time data to obtain real-time key features; The real-time key features are input into the power prediction model to perform power prediction on power demand.
7. A power forecasting system, characterized in that: include: A data acquisition module, used to acquire historical data and real-time data; the historical data includes historical power data and historical environmental data; the real-time data includes real-time power data and real-time environmental data; A preprocessing module, used for at least preprocessing the historical data to remove noise and abnormal values in the historical data; A normalization processing module, used for performing normalization processing on at least the pre-processed historical data; A feature extraction module is used to extract features based on the normalized historical data to extract historical key features; the historical key features include: time window features, temperature-related features, holiday features, seasonal indicator features and historical trend features; A model building module is used to build an electricity forecasting model based on the historical key features, and the electricity forecasting model includes a seasonal difference autoregressive sliding average model; the electricity forecasting model is used to make electricity forecasts on electricity demand based on the real-time data.
8. The power forecasting system according to claim 7, characterized in that: The data acquisition module includes: a data acquisition unit and a data receiving unit; The data collection unit includes a plurality of sensors and a smart meter, which are used to collect the historical data and the real-time data; The data receiving unit receives the historical data and the real-time data based on a message queue technology.
9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the power prediction method according to any one of claims 1 to 6 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
Citation Information
Cited By
Energy scheduling system and method
CN120601395A