User power consumption behavior prediction method and system based on peak-valley time-of-use electricity price
By using a classification and independent forecasting model based on peak-valley time-of-use pricing, the problem of inaccurate load forecasting caused by differences in user electricity consumption behavior is solved, thus achieving grid stability and user electricity reliability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUOTOU ELECTRIC POWER HOLDINGS CO LTD
- Filing Date
- 2026-02-09
- Publication Date
- 2026-05-29
Smart Images
Figure CN122118670A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of smart grid control technology, specifically to a method and system for predicting user electricity consumption behavior based on peak-valley time-of-use pricing. Background Technology
[0002] In the intelligent management and operation of power systems, accurate prediction of user electricity consumption behavior is fundamental to achieving power supply-demand balance and optimizing energy dispatch. With the advancement of smart grid construction, demand response mechanisms have become an important means of regulating grid load distribution. By analyzing users' electricity consumption habits and load characteristics, the grid side can formulate more reasonable dispatch strategies, guiding users to optimize their electricity consumption patterns, thereby improving the overall operating efficiency and stability of the power system.
[0003] In existing engineering practices, power grid dispatch centers typically rely on intelligent measurement systems to collect historical load data from users and use statistical models or machine learning algorithms to model the overall electricity consumption trends of users. These conventional methods mostly treat users' historical electricity consumption data as a continuous time series and directly input it into the prediction model. By analyzing the fluctuation patterns of the data on the time axis, they infer future load demand and use this as data support for formulating peak shaving and valley filling strategies or issuing dispatch instructions.
[0004] However, this traditional approach often overlooks the inherent differences in user behavior patterns under peak-valley pricing mechanisms. Peak hours typically correspond to peak production or daily life activities, resulting in highly planned and regular electricity consumption. In contrast, valley hours involve charging energy storage devices or appliance standby, exhibiting significant randomness and dispersion. Existing prediction models fail to differentiate between these two distinct behavioral drivers, instead mixing data from all time periods. This leads to interference between ordered peak-hour characteristics and random valley-hour characteristics during model training, making it difficult to extract high-purity behavioral features from the mixed data. Consequently, this limits the accuracy of load forecasting and the targeted nature of subsequent grid regulation. Summary of the Invention
[0005] This application relates to the field of smart grid control technology and aims to solve the technical problems in the prior art, which are low load forecast accuracy and grid regulation lag or oscillation due to the neglect of the differences in user electricity consumption patterns at different times and the coupling relationship between peak and valley loads.
[0006] The first aspect of this invention provides a method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing, the method comprising: Historical electricity consumption data of users is obtained and classified according to peak-valley time-of-use pricing to obtain a first dataset; the first dataset is preprocessed to obtain a second dataset; the second dataset is input into a pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods; and the power supply is dynamically adjusted based on the user's electricity consumption behavior.
[0007] In one specific implementation, after acquiring users' historical electricity consumption data, the method involves establishing a time index table and allocating two independent logical storage units as target storage areas in the data processing system. By traversing the collected historical electricity consumption data, the timestamp of each data entry is read and compared with the time index table. If the timestamp falls within the peak time window, it is labeled as peak time; if the timestamp falls within the off-peak time window, it is labeled as off-peak time. The labeled data is then stored in the target storage area to form the first dataset. This classification step decouples users' behavioral patterns under different electricity price periods at the data level, extracting ordered behavioral characteristics driven by production and daily life during peak hours and random behavioral characteristics driven by energy storage standby during off-peak hours.
[0008] In one specific implementation, the preprocessing of the first dataset includes data cleaning and normalization, removing zero or invalid values with statistical bias, and converting different physical quantities into dimensionless pure numerical values. Subsequently, the normalized values of the first dataset corresponding to the target time node are input into a preset mapping table. The preset mapping table pre-stores baseline state values corresponding to standard time nodes. Based on the preset mapping table, discrete sampling points are calibrated into time-axis aligned standard input vectors to generate the second dataset, thereby eliminating alignment errors in the original data over time.
[0009] In one specific implementation, the user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model and a second user electricity consumption behavior prediction model, which are used to process peak and off-peak data, respectively.
[0010] For the first user's electricity consumption behavior prediction model, the processing logic includes: inputting relevant data information of the peak-hour electricity price corresponding to the time period from the second dataset into the model; weighting the normalized load ratio using the difference between the fitted values of the first correction function and the electricity consumption data information to obtain the first output result. The first correction function is used to quantify the temporal distribution characteristics of electricity consumption behavior within a time period, reflecting the user's behavioral inertia within a specific time period; the difference between the fitted values of the electricity consumption data information is used to characterize instantaneous volatility, capturing sudden changes in user behavior by calculating the deviation between the current actual value and the historical long-term fitted trend. The model ultimately outputs a score for the first user's electricity consumption behavior.
[0011] For the second user electricity consumption behavior prediction model, the processing logic includes: inputting relevant data information of the time period corresponding to the off-peak electricity price in the second dataset into the model, using the difference between the second correction function and the fitted value of the off-peak electricity consumption data to perform weighted calculation of the normalized load ratio, and combining the off-peak weight factor and the benchmark electricity consumption to obtain the second output result, thereby determining the second user electricity consumption behavior score.
[0012] In one specific implementation, during the dynamic adjustment of power supply based on user electricity consumption behavior, a correlation function is introduced to calculate the correlation value between a first user's electricity consumption behavior score and a second user's electricity consumption behavior score. The correlation value characterizes the coupling degree between peak-hour and off-peak behavior. The calculation logic of the correlation value is based on the absolute value of the ratio of the difference between the first user's electricity consumption behavior score and the second user's electricity consumption behavior score to their sum, combined with a coupling coefficient representing the degree of interaction and a time coupling weight. The correlation value reflects the degree of influence of a user's peak-hour electricity consumption behavior on their off-peak electricity consumption behavior and the synchronicity between the two.
[0013] In one specific implementation, a hierarchical hysteresis control strategy is used for the dynamic adjustment of power supply. When the detected correlation value is between a first preset threshold and a second preset threshold, the system determines that the power grid is in a controllable state. It calculates the product of the difference between the correlation value and the first preset threshold using a first adjustment coefficient to obtain a first parameter adjustment value, and performs linear fine-tuning of the current power supply. When the detected correlation value is greater than or equal to the second preset threshold, the system determines that there is a risk of load imbalance. It calculates the product of the difference between the correlation value and the second preset threshold using a second adjustment coefficient, and adds the adjustment amount corresponding to the upper limit of the mild adjustment range to obtain a second parameter adjustment value, and performs a significant adjustment of the current power supply.
[0014] A second aspect of the present invention provides a user electricity consumption behavior prediction system based on peak-valley time-of-use pricing, the system comprising: The first dataset acquisition module is used to acquire users' historical electricity consumption data, establish a time index table, compare the traversed timestamps with time windows and then label them with peak or valley time tags, and classify the historical electricity consumption data based on peak and valley time-of-use electricity prices to obtain the first dataset. The second dataset acquisition module is used to preprocess the first dataset by cleaning and normalizing the data, and mapping the results to a preset mapping table to generate a standard input vector to obtain the second dataset. The user electricity consumption behavior prediction module is used to input the second dataset into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use electricity pricing periods. The dynamic adjustment module is used to calculate the correlation value of peak and valley behavior based on the user's electricity consumption behavior, and to perform power supply adjustment in stages according to the threshold range to which the correlation value belongs.
[0015] A third aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method described in the first aspect above.
[0016] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect above.
[0017] The technical solution of this invention decouples historical electricity consumption data into peak and off-peak dimensions at the logical level by establishing a time-axis mapping, and constructs prediction models for each to adapt to the behavioral differences in different time periods. During the prediction process, a correction function and the difference between fitted values are introduced, which corrects the smoothing error caused by the average value while retaining sensitivity to sudden fluctuations. Furthermore, by calculating the correlation value between peak and off-peak behavior, the coupling relationship between the two independent time periods is re-established, and a graded hysteresis regulation strategy is implemented based on this correlation value. This mechanism avoids frequent system oscillations caused by small fluctuations and can respond promptly when large deviations occur, realizing dynamic interactive control of grid energy and improving the reliability of user electricity consumption and the stability of grid operation.
[0018] This invention provides a method and system for predicting user electricity consumption behavior based on peak-valley time-of-use pricing. It has the following beneficial effects: 1. This invention acquires users' historical electricity consumption data and uses a time index table for logical layering, accurately dividing continuous time series data into two independent first datasets: peak hours and valley hours. In the preprocessing stage, a preset mapping table is used to calibrate discrete sampling points into standard input vectors aligned with the time axis. This data-level decoupling effectively separates the ordered behavioral characteristics driven by production and daily life during peak hours from the random behavioral characteristics driven by energy storage standby during valley hours, eliminating interference between different dimensional indicators and achieving the effect of providing a high-purity, low-noise data foundation with clear feature orientation for subsequent prediction models.
[0019] 2. This invention constructs independent prediction models for peak-hour orderliness and valley-hour randomness, and introduces a correction function that describes the time distribution characteristics within a time period and a fitting value difference parameter that characterizes instantaneous fluctuations into the computational logic. The correction function quantifies the user's electricity consumption inertia trend within a specific time window, and the difference parameter captures the sudden deviation of the current actual load from the historical long-term fitting trend. This dual-track modeling method, which includes a dynamic correction factor, overcomes the shortcomings of traditional static statistical models that are insensitive to short-term fluctuations, and achieves the effect of accurately capturing and multi-dimensionally representing the steady-state laws and transient changes in user electricity consumption behavior.
[0020] 3. This invention quantifies the coupling degree of load at different time periods by calculating the correlation value between peak-hour and valley-hour behavior scores using a correlation function. Based on this correlation value, a hierarchical hysteresis control strategy including mild and emergency regulation is constructed. When the correlation value is detected to be in different threshold ranges, linear fine-tuning or large-scale adjustment of superimposed regulation amount is performed respectively. This feedback mechanism based on behavior coupling degree establishes a direct mapping relationship between user-side behavior and grid-side regulation, effectively avoiding frequent system oscillations caused by small load fluctuations, and achieving the effect of dynamic balance of grid energy supply and demand and proactive interactive regulation under the premise of ensuring normal power consumption experience for users. Attached Figure Description
[0021] Figure 1 This is an application environment diagram of a user electricity consumption behavior prediction method based on peak-valley time-of-use pricing in one embodiment. Figure 2 This is a flowchart illustrating a user electricity consumption behavior prediction method based on peak-valley time-of-use pricing in one embodiment. Figure 3 This is a block diagram of a user electricity consumption behavior prediction system based on peak-valley time-of-use pricing in one embodiment; Figure 4 This is an internal structural diagram of a computer device in one embodiment.
[0022] Among them, 102 is the terminal; 104 is the server. Detailed Implementation
[0023] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0024] The user electricity consumption behavior prediction method based on peak-valley time-of-use pricing provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with a data processing platform set on server 104 via a network. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. Server 104 can be implemented as a standalone server or a server cluster consisting of multiple servers.
[0025] In one embodiment, such as Figure 2 As shown, a method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing is provided, and this method is applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps: S1: Obtain the user's historical electricity consumption data and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain the first dataset.
[0026] The user's historical electricity consumption data includes at least average electricity consumption, electricity load, electricity consumption percentage, and electricity consumption growth rate. In practice, data classification is not a simple physical segmentation, but a logical layering based on timeline mapping. This step aims to decouple user behavior patterns under different electricity price periods at the data level, so that subsequent models can capture "peak-hour production / life-driven" behavioral characteristics and "off-peak-hour energy storage / standby-driven" behavioral characteristics respectively.
[0027] In some specific implementations, before acquiring historical electricity consumption data and classifying the historical electricity consumption data based on peak-valley time-of-use pricing to obtain the first dataset, the method further includes: acquiring the time period corresponding to the peak-valley time-of-use pricing for the target region. For example, the peak time price for a certain city corresponds to the time period from 8:00 to 21:00, and the valley time price corresponds to the time period from 21:00 to 8:00 the next day; the time period is marked by time-of-use to generate a target storage area (e.g., the time period from 8:00 to 21:00 is marked as 1, the time period from 21:00 to 8:00 the next day is marked as 2, etc.). Specifically, the system reads the preset peak-valley time period configuration, establishes a time index table, and allocates two independent logical storage units in the data processing system as the target storage area.
[0028] In some specific implementations, acquiring historical electricity consumption data of users and classifying the historical electricity consumption data based on peak-valley time-of-use pricing to obtain a first dataset includes: acquiring historical electricity consumption data of users; classifying the historical electricity consumption data of users according to the time period corresponding to the peak-valley time-of-use pricing based on the collection time corresponding to the historical electricity consumption data of users, to obtain the first dataset. This process involves traversing the collected historical data, reading the timestamp of each data point and comparing it with a time index table. If the timestamp of the data falls within the peak time window, it is labeled as peak time; otherwise, it is labeled as valley time. Finally, the first dataset is stored in the target storage area.
[0029] S2: Preprocess the first dataset to obtain the second dataset.
[0030] The specific steps of this scheme are as follows: First, the first dataset is cleaned, and then the cleaned dataset is normalized. Specifically, data cleaning refers to removing zero or invalid values with statistical biases from the original power grid data due to line maintenance or data acquisition terminal outages, to ensure the continuity of the input model. Normalization is used to eliminate the dimensional differences between different indicators such as "average electricity consumption" (unit: kWh) and "electricity growth rate" (unit: %), converting different physical quantities into dimensionless pure numerical values, so that the model can process each indicator with equal weight.
[0031] Subsequently, the normalized values of the first dataset corresponding to the target time node are input into a preset mapping table, and the second dataset is generated based on the preset mapping table. This preset mapping table includes at least one mapping relationship between a time node and a normalized value. The preset mapping table acts as a feature standardization container, pre-storing the baseline state values corresponding to standard time nodes. Mapping the cleaned data to the table calibrates discrete, unaligned sampling points into time-axis aligned standard input vectors, i.e., the second dataset.
[0032] S3: Input the second dataset into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods.
[0033] It should be noted that the user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model and a second user electricity consumption behavior prediction model.
[0034] The reason for building two separate models is that peak behavior usually has strong orderliness, while trough behavior has strong randomness. Modeling them separately can avoid interference between the two very different mathematical distributions.
[0035] In some specific implementations, inputting the second dataset into a pre-built user electricity consumption behavior prediction model to determine user electricity consumption behavior during peak-valley time-of-use pricing periods includes: inputting relevant data information of the peak-hour pricing time period from the second dataset into a first user electricity consumption behavior prediction model to obtain a first output result. The first user electricity consumption prediction model includes: in, This indicates the first output result. This represents the normalized value of estimated electricity consumption during peak hours. Indicates the number of peak time periods. , Both represent the normalized load ratio, which is the ratio of the actual load to the baseline load. This represents the first correction function, also known as the original time-period correction coefficient, used to describe the temporal distribution characteristics of electricity consumption behavior within a time period. For example, if users start high-power equipment as soon as they arrive at work, the load curve will be highest at the beginning of the period and then gradually decrease, resulting in a value of 2 at the beginning and 1 at the end, indicating that early electricity consumption has a greater impact. If users start equipment close to their lunch break, the load curve will be low at the beginning of the period and reach a peak before the end, resulting in a value of 0.5 at the beginning and 1 at the end, and so on. Indicates time, , For a certain period of time, the start and end times are specified. express Time's up The difference between the fitted values of electricity consumption data at any given time, that is, the difference between the peak electricity consumption on that day and the moving average of the peak electricity consumption over N days. Indicates the time-sharing marker number. Indicates the weighting factor for time period type. This represents the baseline electricity consumption, such as the average daily electricity consumption, where n is the duration of the time period. Based on the first output result, determine the user's electricity consumption behavior during the peak electricity price period; The relevant data information of off-peak electricity prices for the corresponding time periods in the second dataset is input into the second user electricity consumption behavior prediction model to obtain the second output result. The second user electricity consumption behavior prediction model includes: in, This indicates the second output result. Indicates the number of valley time periods. Indicates time, , For a certain period of time, the start and end times are specified. , Both represent the normalized load ratio, which is the ratio of the actual load to the baseline load. The second correction function is represented as above. express Time's up The difference between the fitted values of electricity consumption data at any given time, that is, the difference between the off-peak electricity consumption of the day and the moving average of the peak electricity consumption over N days. This represents the estimated electricity consumption during peak hours, and D represents the off-peak hour weighting factor. As the baseline electricity consumption, The duration of the time period; Based on the second output result, determine the user's electricity consumption behavior during off-peak electricity pricing periods; In the above model, the electricity consumption data information is all normalized value (such as electricity load value, etc.). The fitted value of the electricity consumption data can be obtained through a linear function. For example, y=ax+by+cz, where y represents the fitted value, a, b, and c are all weight coefficients, and x, y, and z represent the average electricity consumption, electricity consumption ratio, and electricity growth rate, respectively.
[0036] In some specific implementations, determining the user's electricity consumption behavior during peak-hour pricing periods based on the first output result includes: Based on the first output result and the first preset mapping table, the user's electricity consumption behavior during peak electricity price periods is determined. The first preset mapping table includes at least one mapping relationship between a model output value and the user's electricity consumption behavior and the user's electricity consumption behavior score. Based on the second output result, the user's electricity consumption behavior during off-peak electricity pricing periods is determined to include: Based on the second output result and the second preset mapping table, the user's second user electricity consumption behavior during peak electricity price periods is determined. The second preset mapping table includes at least one mapping relationship between model output value and second user electricity consumption behavior and second user electricity consumption behavior score.
[0037] S4: Based on the user's electricity consumption behavior, dynamically adjust the power supply.
[0038] It should be noted that this step specifically includes: Using a correlation function, the correlation value between the electricity consumption behavior of the first user and the electricity consumption behavior of the second user is calculated. The correlation function includes: in, Indicates associated value, Represents an association function. The coupling coefficient is determined by the degree of interaction between the user and the power grid, and its value ranges from 0 to 1. The higher the degree of interaction, the larger the coupling coefficient. This represents the score for the first user's electricity consumption behavior. This indicates the second user's electricity consumption behavior score. Indicates the time-coupled weights. Represents the smoothing constant. This represents the system adjustment parameters, where J is the total number of time period types and q is the qth time period. Based on the correlation value, the power supply is dynamically adjusted, including: in response to detecting that the correlation value is greater than a first preset threshold and less than a second preset threshold, the current power supply is dynamically adjusted according to a first parameter adjustment value, wherein the first preset threshold and the second preset threshold can be set according to actual needs; at this time, the system determines that the power grid is in a sub-healthy but controllable state, and the first parameter adjustment value is a linear or small-amplitude correction amount, which aims to fine-tune the power supply strategy without affecting the normal user experience; In response to the detection that the associated value is greater than or equal to a second preset threshold, the current power supply is dynamically adjusted according to the second parameter adjustment value. At this time, the system determines that there is a risk of load imbalance, and the activated second parameter adjustment value contains a larger adjustment factor. This hierarchical mechanism constructs a control loop with hysteresis characteristics to prevent the system from frequently and significantly adjusting due to small fluctuations.
[0039] The first and second parameter adjustment values mentioned above are obtained through the correlation value and the third preset mapping table. The third preset mapping table includes at least one mapping relationship between the correlation value and the parameter adjustment value. For example, the correlation value is in the first interval with no adjustment, and the correlation value is in the second and third intervals with mild adjustment and emergency adjustment, respectively. The calculation methods are AS1=r1(F-QG1) and AS2=r2(F-QG2)+AS1(QG2), where r1 and r2 are adjustment coefficients, AS1 and AS2 are adjustment values, QG1 is the first preset threshold, QG2 is the second preset threshold, and AS1(QG2) is the parameter adjustment value corresponding to the upper limit of the mild adjustment interval (i.e., the second preset threshold).
[0040] The aforementioned method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing includes: acquiring historical user electricity consumption data and classifying the historical data based on peak-valley time-of-use pricing to obtain a first dataset; preprocessing the first dataset to obtain a second dataset; inputting the second dataset into a pre-constructed user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods; and dynamically adjusting the power supply based on the user's electricity consumption behavior. This application enables dynamic energy interaction control and management, thereby improving the reliability of user electricity consumption.
[0041] It should be understood that, although Figure 2The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 2 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.
[0042] In one embodiment, such as Figure 3 As shown, a user electricity consumption behavior prediction system based on peak-valley time-of-use pricing is provided, including: a first dataset acquisition module, a second dataset acquisition module, a user electricity consumption behavior prediction module, and a dynamic adjustment module. These modules are not merely logical program units; in actual deployment, they can be manifested as distributed computing nodes running on a server cluster or functional units integrated into an edge computing gateway. Wherein: The first dataset acquisition module is used to acquire users' historical electricity consumption data and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain the first dataset. Specifically, during operation, this module is responsible for establishing a data channel with smart meters or energy management centers, pulling raw load data in real time or at regular intervals, and using an internally maintained time index table to physically segment or logically label continuous time series data, dividing it into peak data groups and valley data groups.
[0043] The second dataset acquisition module is used to preprocess the first dataset to obtain the second dataset. This module mainly undertakes the tasks of "cleaning" and "normalizing". By loading preset normalization parameters, it transforms the raw readings with inconsistent dimensions into standard input vectors that the model can recognize, and eliminates the interference of invalid data.
[0044] The user electricity consumption behavior prediction module is used to input the second dataset into a pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak and off-peak time-of-use pricing periods. Internally, this module runs two independent sets of computational logic in parallel or sequentially, performing feature extraction and trend extrapolation based on peak and off-peak data characteristics, respectively.
[0045] The dynamic adjustment module is used to dynamically adjust the power supply based on the user's electricity consumption behavior. This module, as the system's output execution unit, is responsible for calculating the correlation indicators of peak-valley behavior and issuing specific load adjustment commands to the power grid control layer based on the calculation results.
[0046] In a preferred embodiment of the present invention, the first dataset acquisition module is specifically used for: Obtain the time period corresponding to the peak-valley time-of-use electricity price for the target region; The time period is divided into time segments and a target storage area is generated.
[0047] In a preferred embodiment of the present invention, the first dataset acquisition module is further configured to: Obtain historical electricity consumption data of users, which includes at least average electricity consumption, electricity load, electricity consumption ratio and electricity consumption growth rate. Based on the collection time corresponding to the user's historical electricity consumption data, the user's historical electricity consumption data is classified according to the time period corresponding to the peak-valley time-of-use electricity price to obtain the first dataset; Store the first dataset in the target storage area.
[0048] In a preferred embodiment of the present invention, the second dataset acquisition module is specifically used for: The first dataset is cleaned, and the cleaned dataset is then normalized. The normalized values of the first dataset corresponding to the target time node are input into a preset mapping table, and the second dataset is generated based on the preset mapping table.
[0049] In a preferred embodiment of the present invention, the user electricity consumption behavior prediction module is specifically used for: The user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model. The second dataset is input into the pre-constructed user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of the peak electricity price corresponding to the time period in the second dataset into the first user electricity consumption behavior prediction model to obtain the first output result; Based on the first output result, determine the user's electricity consumption behavior during the peak electricity price period.
[0050] In a preferred embodiment of the present invention, the user electricity consumption behavior prediction module is further configured to: The user electricity consumption behavior prediction model further includes a second user electricity consumption behavior prediction model. The second dataset is input into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of the off-peak electricity price corresponding to the time period in the second dataset into the second user electricity consumption behavior prediction model; Based on the second output result, determine the user's electricity consumption behavior during off-peak electricity pricing periods.
[0051] In a preferred embodiment of the present invention, the user electricity consumption behavior prediction module is further configured to: Based on the first output result and the first preset mapping table, determine the user's electricity consumption behavior during peak electricity price periods; Based on the second output result, the user's electricity consumption behavior during off-peak electricity pricing periods is determined to include: Based on the second output result and the second preset mapping table, the user's second user electricity consumption behavior during peak electricity price periods is determined.
[0052] In a preferred embodiment of the present invention, the dynamic adjustment module is specifically used for: Using the correlation function, calculate the correlation value between the electricity consumption behavior of the first user and the electricity consumption behavior of the second user; Based on the aforementioned correlation value, the power supply is dynamically adjusted.
[0053] In a preferred embodiment of the present invention, the dynamic adjustment module is further configured to: In response to detecting that the associated value is greater than a first preset threshold and less than a second preset threshold, the current power supply is dynamically adjusted according to the first parameter adjustment value; In response to detecting that the associated value is greater than or equal to the second preset threshold, the current power supply is dynamically adjusted according to the second parameter adjustment value.
[0054] Specific limitations regarding the user electricity consumption behavior prediction system based on peak-valley time-of-use pricing can be found in the limitations of the user electricity consumption behavior prediction method based on peak-valley time-of-use pricing mentioned above, and will not be repeated here. Each module in the aforementioned user electricity consumption behavior prediction system based on peak-valley time-of-use pricing can be implemented entirely or partially through software, hardware, or a combination thereof. Specifically, each module can be stored as an independent source code file, dynamic link library, or executable program in a computer device; it can also be embedded as logic gate circuits in an FPGA (Field-Programmable Gate Array) or ASIC (Application-Specific Integrated Circuit) to improve the real-time performance of data processing.
[0055] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus.
[0056] Processor: Provides computational and control capabilities. In this embodiment, the processor is configured to perform complex matrix operations to support the inference process of the predictive model and is responsible for calculating the aforementioned correction function and correlation values.
[0057] Memory: Includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. Specifically, the storage media also includes a dedicated database area for storing time-sharing data from the "target storage area," pre-built user electricity consumption behavior prediction model files, and preset mapping tables. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage media.
[0058] Network interface: Used for communication with external terminals via network connection. Specifically, this network interface is used to receive raw electricity consumption data from the AMI (Advanced Metering Architecture) system and send dynamic adjustment commands to the control terminal on the distribution network side.
[0059] Display screen and input device: The display screen of the computer device can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs or touchpads set on the casing of the computer device, or external keyboards, touchpads or mice, etc., used to intuitively display the predicted user power consumption behavior curve and the current power supply adjustment status, and allow maintenance personnel to manually input or correct preset thresholds.
[0060] Those skilled in the art will understand that Figure 4 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.
[0061] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: S1: Obtain historical electricity consumption data of users, and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain the first dataset; S2: Preprocess the first dataset to obtain the second dataset; S3: Input the second dataset into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods; S4: Based on the user's electricity consumption behavior, dynamically adjust the power supply.
[0062] In one embodiment, the processor, when executing a computer program, also performs the following steps: Obtain the time period corresponding to the peak-valley time-of-use electricity price for the target region; The time period is divided into time segments and a target storage area is generated.
[0063] In one embodiment, the processor, when executing a computer program, also performs the following steps: Obtain historical electricity consumption data of users, which includes at least average electricity consumption, electricity load, electricity consumption ratio and electricity consumption growth rate. Based on the collection time corresponding to the user's historical electricity consumption data, the user's historical electricity consumption data is classified according to the time period corresponding to the peak-valley time-of-use electricity price to obtain the first dataset; Store the first dataset in the target storage area.
[0064] In one embodiment, the processor, when executing a computer program, also performs the following steps: The first dataset is cleaned, and the cleaned dataset is then normalized. The normalized values of the first dataset corresponding to the target time node are input into a preset mapping table, and the second dataset is generated based on the preset mapping table.
[0065] In one embodiment, the processor, when executing a computer program, also performs the following steps: The user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model. The second dataset is input into the pre-constructed user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of the peak electricity price corresponding to the time period in the second dataset into the first user electricity consumption behavior prediction model; Based on the first output result, determine the user's electricity consumption behavior during the peak electricity price period.
[0066] In one embodiment, the processor, when executing a computer program, also performs the following steps: The user electricity consumption behavior prediction model further includes a second user electricity consumption behavior prediction model. The second dataset is input into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of the off-peak electricity price corresponding to the time period in the second dataset into the second user electricity consumption behavior prediction model; Based on the second output result, determine the user's electricity consumption behavior during off-peak electricity pricing periods.
[0067] In one embodiment, the processor, when executing a computer program, also performs the following steps: Based on the first output result and the first preset mapping table, determine the user's electricity consumption behavior during peak electricity price periods; Based on the second output result, the user's electricity consumption behavior during off-peak electricity pricing periods is determined to include: Based on the second output result and the second preset mapping table, the user's second user electricity consumption behavior during peak electricity price periods is determined.
[0068] In one embodiment, the processor, when executing a computer program, also performs the following steps: Using the correlation function, calculate the correlation value between the electricity consumption behavior of the first user and the electricity consumption behavior of the second user; Based on the aforementioned correlation value, the power supply is dynamically adjusted.
[0069] In one embodiment, the processor, when executing a computer program, also performs the following steps: In response to detecting that the associated value is greater than a first preset threshold and less than a second preset threshold, the current power supply is dynamically adjusted according to the first parameter adjustment value; In response to detecting that the associated value is greater than or equal to the second preset threshold, the current power supply is dynamically adjusted according to the second parameter adjustment value.
[0070] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: S1: Obtain historical electricity consumption data of users, and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain the first dataset; S2: Preprocess the first dataset to obtain the second dataset; S3: Input the second dataset into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods; S4: Based on the user's electricity consumption behavior, dynamically adjust the power supply.
[0071] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: Obtain the time period corresponding to the peak-valley time-of-use electricity price for the target region; The time period is divided into time segments and a target storage area is generated.
[0072] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: Obtain historical electricity consumption data of users, which includes at least average electricity consumption, electricity load, electricity consumption ratio and electricity consumption growth rate. Based on the collection time corresponding to the user's historical electricity consumption data, the user's historical electricity consumption data is classified according to the time period corresponding to the peak-valley time-of-use electricity price to obtain the first dataset; Store the first dataset in the target storage area.
[0073] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: The first dataset is cleaned, and the cleaned dataset is then normalized. The normalized values of the first dataset corresponding to the target time node are input into a preset mapping table, and the second dataset is generated based on the preset mapping table.
[0074] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: The user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model. The second dataset is input into the pre-constructed user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of the peak electricity price corresponding to the time period in the second dataset into the first user electricity consumption behavior prediction model; Based on the first output result, determine the user's electricity consumption behavior during the peak electricity price period.
[0075] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: The user electricity consumption behavior prediction model further includes a second user electricity consumption behavior prediction model. The second dataset is input into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: Input the relevant data information of off-peak electricity prices in the second dataset into the second user electricity consumption behavior prediction model to obtain the second output result; Based on the second output result, determine the user's electricity consumption behavior during off-peak electricity pricing periods.
[0076] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: Based on the first output result and the first preset mapping table, determine the user's electricity consumption behavior during peak electricity price periods; Based on the second output result, the user's electricity consumption behavior during off-peak electricity pricing periods is determined to include: Based on the second output result and the second preset mapping table, the user's second user electricity consumption behavior during peak electricity price periods is determined.
[0077] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: Using the correlation function, calculate the correlation value between the electricity consumption behavior of the first user and the electricity consumption behavior of the second user; Based on the aforementioned correlation value, the power supply is dynamically adjusted.
[0078] In one embodiment, when the computer program is executed by a processor, it also performs the following steps: In response to detecting that the associated value is greater than a first preset threshold and less than a second preset threshold, the current power supply is dynamically adjusted according to the first parameter adjustment value; In response to detecting that the associated value is greater than or equal to the second preset threshold, the current power supply is dynamically adjusted according to the second parameter adjustment value.
[0079] 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. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0080] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing, characterized in that, Includes the following steps: Obtain users' historical electricity consumption data and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain the first dataset; The first dataset is preprocessed to obtain the second dataset; The second dataset is input into a pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods. The power supply is dynamically adjusted based on the user's electricity consumption behavior.
2. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 1, characterized in that, Before acquiring users' historical electricity consumption data and classifying the historical electricity consumption data based on peak-valley time-of-use pricing to obtain the first dataset, the method further includes: Obtain the time period corresponding to the peak-valley time-of-use electricity price for the target region, read the preset peak-valley time period configuration, and create a time index table; The time period is divided into time segments, and two independent logical storage units are allocated in the data processing system as target storage areas.
3. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 2, characterized in that, Obtain users' historical electricity consumption data and classify the historical electricity consumption data based on peak-valley time-of-use electricity prices to obtain a first dataset, including: Obtain historical electricity consumption data of users, which includes at least average electricity consumption, electricity load, electricity consumption ratio and electricity consumption growth rate. The system iterates through the user's historical electricity consumption data, reads the timestamp of each data item and compares it with the time index table. If the timestamp of the data falls within the peak time window, it is labeled with a peak time tag; if the timestamp of the data falls within the valley time window, it is labeled with a valley time tag. The tagged user historical electricity consumption data is stored in the target storage area to decouple the user's behavior patterns under different electricity price periods at the data level, thus obtaining the first dataset.
4. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 3, characterized in that, The first dataset is preprocessed to obtain the second dataset, which includes: The first dataset is cleaned to remove zero or invalid values with statistical bias, and the cleaned first dataset is normalized. The normalized value of the first dataset corresponding to the target time node is input into a preset mapping table, which pre-stores the baseline state value corresponding to the standard time node. Based on the preset mapping table, the discrete sampling points are calibrated into a time-axis aligned standard input vector to generate the second dataset.
5. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 4, characterized in that, The user electricity consumption behavior prediction model includes a first user electricity consumption behavior prediction model. The second dataset is input into the pre-constructed user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: The relevant data information of the peak electricity price corresponding to the time period in the second dataset is input into the first user electricity consumption behavior prediction model. The normalized load ratio is calculated by weighting the difference between the first correction function and the fitted value of the electricity consumption data information to obtain the first output result. The first correction function is used to describe the time distribution characteristics of electricity consumption behavior within the time period, and the difference between the fitted values of the electricity consumption data information is used to characterize the instantaneous volatility. Based on the first output result, determine the user's electricity consumption behavior during the peak electricity price period.
6. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 5, characterized in that, The user electricity consumption behavior prediction model further includes a second user electricity consumption behavior prediction model. The second dataset is input into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use pricing periods, including: The relevant data information of the off-peak electricity price corresponding to the time period in the second dataset is input into the second user electricity consumption behavior prediction model. The normalized load ratio is calculated by weighting the difference between the fitting value of the second correction function and the off-peak electricity consumption data information to obtain the second output result. Based on the second output result, determine the user's electricity consumption behavior during off-peak electricity pricing periods.
7. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 6, characterized in that, Based on the first output result, the user's electricity consumption behavior during peak-hour pricing periods is determined to include: Based on the first output result and the first preset mapping table, determine the user's first user electricity consumption behavior score during peak electricity price periods; Based on the second output result, the user's electricity consumption behavior during off-peak electricity pricing periods is determined to include: Based on the second output result and the second preset mapping table, the user's second user electricity consumption behavior score during off-peak electricity price periods is determined.
8. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 7, characterized in that, Dynamically adjusting the power supply based on the user's electricity consumption behavior includes: Using a correlation function, calculate the correlation value between the first user's electricity consumption behavior score and the second user's electricity consumption behavior score; The correlation value represents the coupling degree between peak-hour behavior and valley-hour behavior. Its calculation logic is as follows: it is calculated based on the ratio of the difference and the sum of the first user's electricity consumption behavior score and the second user's electricity consumption behavior score, combined with the coupling coefficient representing the degree of interaction and the time coupling weight. Based on the aforementioned correlation value, the power supply is dynamically adjusted.
9. The method for predicting user electricity consumption behavior based on peak-valley time-of-use pricing according to claim 8, characterized in that, Based on the aforementioned correlation value, dynamic adjustment of the power supply includes: In response to detecting that the associated value is greater than a first preset threshold and less than a second preset threshold, the product of the difference between the associated value and the first preset threshold is calculated using a first adjustment coefficient to obtain a first parameter adjustment value, and the current power supply is linearly fine-tuned accordingly. In response to the detection that the correlation value is greater than or equal to the second preset threshold, the product of the difference between the correlation value and the second preset threshold is calculated using the second adjustment coefficient, and the adjustment amount corresponding to the upper limit of the mild adjustment range is added to obtain the second parameter adjustment value, and the current power supply is dynamically adjusted accordingly.
10. A user electricity consumption behavior prediction system based on peak-valley time-of-use pricing is based on the user electricity consumption behavior prediction method based on peak-valley time-of-use pricing as described in any one of claims 1 to 9, characterized in that, include: The first dataset acquisition module is used to acquire users' historical electricity consumption data information and classify the historical electricity consumption data information based on peak-valley time-of-use electricity prices to obtain the first dataset. The second dataset acquisition module is used to preprocess the first dataset to obtain the second dataset; The user electricity consumption behavior prediction module is used to input the second dataset into the pre-built user electricity consumption behavior prediction model to determine the user's electricity consumption behavior during peak-valley time-of-use electricity pricing periods. The dynamic adjustment module is used to dynamically adjust the power supply based on the user's electricity consumption behavior.