Electricity load prediction and energy-saving dispatching control method and system for electric heating table

Abstract

The application discloses an electric heating table-oriented power load prediction and energy-saving scheduling control method and system, belongs to the technical field of energy-saving scheduling control, and is based on load time sequence data representing electric heating table power consumption behaviors to implement corresponding time sequence feature interference situation analysis, to determine whether the conditions for entering the electric heating table load prediction link are met; if the determination result is that the entering conditions are met, the model with the time sequence dependency relationship capturing function is used to predict the electric heating table energy use, and the power consumption scheduling control strategy is implemented accordingly; if the determination result is that the entering conditions are not met, the time sequence feature interference strength is taken as the adjustment basis, the capturing accuracy of the model to the time sequence dependency relationship presented by the load time sequence data is focused on, targeted adjustment is implemented, and whether to enter the electric heating table load prediction link is determined again after the adjustment is completed, which helps to solve the problems of high time deviation degree of the electric heating table power load based on time sequence prediction and imbalance of energy-saving scheduling control.

Description

Technical Field

[0001] This invention relates to the field of energy-saving dispatch and control technology, and in particular to a method and system for predicting and controlling the power load of electric heating tables. Background Technology

[0002] In the wave of energy transition and smart technology, accurate prediction of household appliance energy consumption has gone far beyond simply "knowing how much electricity is used." It has become a core technological cornerstone for achieving efficient, safe, and economical operation of energy systems. The random starting and stopping of a massive number of household appliances is a source of disturbance to the power grid. Through prediction and coordinated control, disordered electricity consumption can be transformed into "dispatchable" flexible loads, providing ancillary services to the power grid and enhancing its resilience and frequency stability. Electric heating tables, as a commonly used high-power, seasonal, and thermally inertial heating appliance, have an even more unique and urgent significance for load prediction during the winter heating season, especially during peak evening electricity consumption.

[0003] Electric heating tables typically consist of the following structure: a metal frame, a tabletop, and electric heating elements, such as heating wires or carbon fiber heating bodies, usually located on the underside of the tabletop, around the base, or inside a dedicated heating box; a control module (including a control box, electronic board, SCR, and heat sink, used to regulate the table's temperature and power); a heat sink; and data acquisition equipment (usually a room temperature sensor at the bottom of the tabletop and a temperature sensor attached to the heat sink fins). The existing scheduling process for electric heating tables involves the following: real-time data collection using sensors (such as room temperature sensors, power sensors, and temperature sensors) to predict loads, including power, room temperature, tabletop and under-table heating temperatures, user status, usage time, and ambient temperature and humidity. This data is then cleaned, normalized, and features extracted to form standardized time-series load data. A time-series forecasting model, such as LSTM (Long Short-Term Memory), is then employed. The system uses data such as Long Short-Term Memory (LSTM) networks to input historical load time-series data, regional temperature (e.g., desktop area, non-desktop area), human body induction signals, and weather factors. It outputs hourly load predictions for the next 24 hours or shorter timescales. Based on these predictions, and with energy conservation as a constraint, it dynamically plans the start-up and shutdown times, temperature settings, and power levels of electric heating tables (e.g., heating levels can be set via knobs, or power modulation of the heating circuit can be achieved using SCRs). The system executes scheduling commands through an IoT controller and monitors actual load and room temperature. Utilizing a Proportional-Integral-Derivative (PID) controller, it provides feasible paths for home energy management and demand-side response.

[0004] For example, Chinese invention patent CN116300626B discloses a control system for a height-adjustable heating table and its motor drive circuit, including: a heating control module, comprising an array heating unit and multiple heat direction adjustment units; wherein the array heating unit is used to generate real-time heat flow through a built-in electric ceramic stove, and the heat direction adjustment units are used to adjust the direction of the real-time heat flow; an energy-saving control module, comprising a human body sensing module and a temperature sensing module; wherein the temperature sensing module is used to adjust the real-time heat flow generated by the array heating unit according to the ambient temperature; a height-adjustable motor control module, comprising a desktop height-adjustable control unit and a whole-machine height-adjustable control unit; wherein the desktop height-adjustable control unit adjusts the desktop height by controlling the forward and reverse connection of the motor, and the whole-machine height-adjustable control unit adjusts the whole-machine height by controlling the forward and reverse connection of the motor; and a main control MCU, used to connect the heating control module, the height-adjustable motor control module, and the energy-saving control module.

[0005] Although the existing technology proposes an integrated control system for lifting and heating tables that achieves integrated control of heating and energy-saving sensing, optimizing the user experience and basic energy-saving effect, it has shortcomings in the design of interference adaptability and load prediction accuracy under high-power operation scenarios of electric heating tables.

[0006] For example, Chinese invention patent CN121028616A discloses a control system and method for a lifting and heating table, which includes: a control mode determination module for collecting real-time basic parameters and determining whether to execute an intelligent control mode; a control cycle determination module for extracting real-time basic parameters and determining the control cycle; a comprehensive data acquisition module for collecting lifting and heating data; a model simulation determination module for establishing a virtual simulation model, simulating the simulation state, and determining whether to issue a stop command; and a control strategy execution module for formulating and executing a control strategy.

[0007] The existing method, by using basic parameter acquisition and fixed-mode simulation, does not consider the interference of local thermal radiation on sensor data, resulting in significant shortcomings in data validity and prediction adaptability under actual complex working conditions.

[0008] The above-mentioned technology has at least the following technical problems:

[0009] In the energy usage prediction process for electric heating tables, interference from local thermal radiation (such as during peak electricity consumption in winter evenings when the electric heating table operates at full power continuously, causing the temperature of the heat sink and metal frame to remain high for extended periods, potentially affecting sensors near the frame) and the potential for PID scheduling response delays due to local thermal radiation interference (thermal radiation amplifies the physical delay in temperature feedback) may introduce time-series characteristic distortion. The reasons are as follows: Firstly, local thermal radiation interference causes core parameters collected by the sensors (such as room temperature) to deviate from the actual environmental conditions, resulting in spurious fluctuations or deviations in the load time-series data. Secondly, the delayed PID scheduling response behavior can cause distortions in the power time-series sequence of the electric heating table. Lagging power fluctuations; and because existing methods typically use fixed-dimensional feature inputs (using only fixed historical power, ambient temperature and humidity, etc., which cannot accurately characterize the load change mechanism under disturbances), the models may be unable to capture short-term load fluctuations and lack adaptability to lagging fluctuations. This directly leads to the deviation between the load forecast and the actual electricity load exceeding the acceptable threshold. The formulation of energy-saving dispatch strategies may lack accurate data and the degree of deviation between the dispatch logic and the actual electricity consumption scenario may increase. This may further cause errors in the dispatch instructions (temperature setting, power level adjustment, etc.) and the prediction of grid load adjustment, resulting in the problem of high time deviation of the electricity load of electric heating tables based on time-series prediction and imbalance of energy-saving dispatch control. Summary of the Invention

[0010] This invention provides a method and system for predicting and controlling the power load of electric heating tables, which can improve the accuracy of energy-saving scheduling control for electric heating tables. The technical solution provided by this application is as follows:

[0011] Firstly, a method for electricity load forecasting and energy-saving scheduling control for electric heating tables is provided. The specific implementation of this method is as follows: Within a set electricity load forecasting scenario, based on the load time-series data characterizing the electricity consumption behavior of the electric heating table, a corresponding time-series characteristic interference analysis is performed to determine whether the conditions for entering the load forecasting stage of the electric heating table are met. If the determination result is that the entry conditions are met, the energy usage of the electric heating table is predicted through a model with the function of capturing time-series dependencies, and an electricity scheduling control strategy is implemented accordingly. If the determination result is that the entry conditions are not met, the intensity of time-series characteristic interference is used as the adjustment basis, focusing on the accuracy of the model in capturing the time-series dependencies presented by the load time-series data, and targeted adjustments are implemented. After the adjustment is completed, a new determination is made as to whether to enter the load forecasting stage of the electric heating table.

[0012] Secondly, a power load forecasting and energy-saving dispatching control system for electric heating tables is provided, including: an electric heating table time-series interference characterization module, which analyzes the corresponding time-series interference based on the load time-series data characterizing the power consumption behavior of the electric heating table within a set power load forecasting scenario, and determines whether the conditions for entering the load forecasting stage of the electric heating table are met; an electric heating table dispatching control module, which, if the determination result is that the entry conditions are met, predicts the energy usage of the electric heating table through a model with the function of capturing time-series dependencies, and implements a power dispatching control strategy accordingly; and an electric heating table time-series adjustment module, which, if the determination result is that the entry conditions are not met, adjusts the system based on the intensity of time-series interference, focuses on the accuracy of the model in capturing the time-series dependencies presented by the load time-series data, implements targeted adjustments, and re-determines whether to enter the load forecasting stage of the electric heating table after the adjustment is completed.

[0013] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0014] 1. Based on the load time-series data characterizing the electricity consumption behavior of electric heating tables, corresponding time-series characteristic interference analysis is performed to determine whether the conditions for entering the load forecasting stage for electric heating tables are met. This helps to effectively screen the input data for load forecasting and avoid prediction deviations caused by invalid data from the source. If the determination result is that the entry conditions are met, the energy consumption of electric heating tables is predicted by a model with the function of capturing time-series dependencies, and power dispatching control strategies are implemented accordingly. This helps to improve the matching degree between the electricity consumption of electric heating tables and the power grid load based on accurate load forecasting peak shaving and valley filling. If the determination result is that the entry conditions are not met, the intensity of time-series characteristic interference is used as the adjustment basis. The focus is on the accuracy of the model in capturing the time-series dependencies presented by the load time-series data, and targeted adjustments are implemented. After the adjustment is completed, the determination of whether to enter the load forecasting stage for electric heating tables is re-evaluated. This helps to ensure that the predicted data conforms to the actual electricity consumption pattern, provides a reliable basis for subsequent power adjustment, and avoids frequent switching of heating table power due to improper scheduling on the user side.

[0015] 2. By monitoring the time-series characteristic distortion value, and when the time-series characteristic distortion value exceeds the predefined time-series characteristic distortion value, it is possible to choose whether to mark the load time-series data under local thermal radiation interference or the load time-series data under abnormal temperature adjustment frequency. Compared with the shortcomings of existing technologies, such as indiscriminate marking of abnormal load data and inability to distinguish the type of interference source, this method is more conducive to further realizing the refined classification and marking of abnormal load data, clarifying the influence mechanism of interference source on load time-series characteristics, thereby improving the pertinence of subsequent abnormal data correction and reducing the misjudgment rate of load prediction model.

[0016] 3. In high-power steady-state operation scenarios, thermal radiation interference becomes the dominant factor. Based on the time-series characteristic distortion value obtained under high-power steady-state operation scenarios, the system determines whether to enter the load prediction stage of the electric heating table. When the monitored time-series characteristic distortion value is not greater than the predefined qualified threshold, the system enters the load prediction stage of the electric heating table. Compared with existing technologies, which usually have poor adaptability to thermal radiation-dominated interference under high-power steady-state conditions, the model confuses false load characteristics with real electricity consumption patterns, which helps to accurately remove distorted data caused by thermal radiation interference under high-power steady-state conditions, thereby improving the ability of the time-series prediction model to capture the real electricity consumption patterns of the electric heating table.

[0017] 4. By marking load time-series data under abnormal temperature adjustment frequency during the emergency peak shaving dispatch phase of the power grid, and monitoring whether the number of temperature adjustments during the predicted abnormal period of power load exceeds the corresponding adjustment peak, a fixed power is used as the power value of the electric heating table under the current state when the peak is exceeded, and a fault reminder is pushed at the same time. Compared with the shortcomings of existing methods that only passively respond to power limits during the emergency dispatch phase and cannot actively identify abnormal adjustment behavior, this method helps to realize the active monitoring and intervention of the power consumption behavior of electric heating tables in emergency peak shaving scenarios, avoid the aggravation of power grid load fluctuations caused by abnormal adjustments, and thus ensure the effective execution of power grid peak shaving dispatch instructions and improve the dispatch response efficiency of the electric heating table group control system.

[0018] 5. By performing load forecasting on electric heating tables, load forecast values ​​are obtained, and when the load forecast value is not less than the upper limit of the pre-stored load forecast range, the load status is dynamically adjusted. Compared with the shortcomings of existing technologies that passively limit power after the load exceeds the limit and lack advance prediction and dynamic adjustment, this helps to reflect the synergistic mechanism of load forecasting and active control, realize the forward-looking management of electricity load, thereby improving the energy efficiency of electric heating tables and the stability of grid-side load, and reducing the power supply pressure caused by sudden over-limit. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a general flowchart of the power load prediction and energy-saving scheduling control method for electric heating tables provided in the embodiments of the present invention;

[0021] Figure 2 This is a schematic diagram of the process for marking load time-series data under abnormal temperature feedback provided in an embodiment of the present invention;

[0022] Figure 3 This is a schematic diagram of load prediction training provided in an embodiment of the present invention;

[0023] Figure 4 This is a schematic diagram of the power load prediction and energy-saving scheduling control system for electric heating tables provided in an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the implementation methods of this application will be further described in detail below with reference to the accompanying drawings.

[0025] Before providing a detailed explanation of the embodiments of this application, the application scenarios of these embodiments will be described first.

[0026] Example 1: This embodiment of the invention provides a method for predicting and controlling the power load of an electric heating table. The processing flow of this method may include the following steps:

[0027] Electric heating table timing interference characterization: Within a set electricity load forecasting scenario, based on the load timing data characterizing the electricity consumption behavior of the electric heating table, corresponding timing characteristic interference analysis is performed. Through this analysis, it is determined whether the conditions for entering the load forecasting stage for the electric heating table are met. By characterizing the timing interference of the electric heating table, it is helpful to accurately determine the validity of the load timing data, filtering out distorted data affected by interference such as local heat radiation and PID scheduling response delay from the source, and providing a reliable data foundation for subsequent load forecasting and scheduling control.

[0028] Electric heating table scheduling control: If the determination result is that the entry condition is met, the energy usage of the electric heating table is predicted by a model with the function of capturing time-series dependencies, and the power dispatch control strategy is implemented accordingly. By carrying out electric heating table scheduling control, it is helpful to balance the user's heating demand and the power grid load regulation requirements, and avoid scheduling command errors caused by prediction deviations.

[0029] Electric heating table time series adjustment: If the judgment result is that the entry conditions are not met, the adjustment measures are based on the intensity of time series characteristic interference. The focus is on the accuracy of the model in capturing the time series dependence relationship presented by the load time series data, and targeted adjustments are implemented. After the adjustment is completed, it is re-determined whether to enter the load prediction stage of the electric heating table. By performing electric heating table time series adjustment, it is helpful to achieve targeted correction of distorted load time series data, thereby helping to restore the original time series pattern of the data, improve the usability of invalid data, and reduce the direct rejection of the prediction stage due to data distortion.

[0030] Prior to the design of the power load prediction and energy-saving scheduling control method for electric heating tables provided in this application, a database with standardized data storage and management functions has been constructed. This database serves as the data foundation of the entire methodology, providing solid support for subsequent accurate prediction and intelligent scheduling. It includes, but is not limited to, pre-stored thermal radiation level values. The power load prediction and energy-saving scheduling control method for electric heating tables proposed in this application can fully utilize data resources to achieve accurate power load prediction and intelligent energy-saving scheduling control, providing strong guarantees for the energy-saving operation of electric heating tables and convenient use by users.

[0031] like Figure 1 The diagram shows the overall flowchart of the power load prediction and energy-saving scheduling control method for electric heating tables provided in the embodiments of this invention: Based on the load time-series data characterizing the power consumption behavior of the electric heating table, the corresponding time-series characteristic interference analysis is performed, and the time-series characteristic distortion value is obtained. When the monitored time-series characteristic distortion value is not greater than the predefined time-series characteristic distortion value, the load prediction stage of the electric heating table is entered; otherwise, the influence degree analysis of the time-series characteristics is performed to obtain the local thermal radiation influence value and the abnormal temperature feedback influence value. When the local thermal radiation influence value is greater than the upper limit of the influence range defined by the abnormal temperature feedback influence value, the load time-series data under local thermal radiation interference is marked. When the local thermal radiation influence value is within the influence range defined by the abnormal temperature feedback influence value, the thermal radiation-regulation coordinated interference regulation is entered. When the local thermal radiation influence value is less than the lower limit of the influence range defined by the abnormal temperature feedback influence value, the load time-series data under abnormal temperature feedback is marked.

[0032] In this embodiment, by characterizing the timing interference of electric heating tables, and exploring the interaction and influence between the scheduling control and timing adjustment of electric heating tables, it is helpful to construct a closed-loop processing mechanism from data validity determination to accurate prediction and scheduling, and then to the correction of distorted data. This forms a linkage system of interference identification, prediction execution, and deviation correction, thereby improving the accuracy of electric heating table power load prediction and the stability of energy-saving scheduling control, and solving the core problems of excessive prediction deviation and scheduling imbalance in the prior art.

[0033] Preferably, the analysis of the interference of the corresponding time series characteristics is carried out based on the load time series data characterizing the electricity consumption behavior of the electric heating table, as follows: The analysis of the time series characteristic distortion of the load time series data of the electric heating table is carried out. The number of abrupt change points of the load time series data monitored by the counter within the electricity load prediction period is taken as the time series characteristic distortion value. The abrupt change point is the preset sampling point when the power change between the preset sampling point monitored by the power sensor and the adjacent sampling point is greater than a preset multiple of the rated power (such as 20%).

[0034] To control the validity of data input to the load forecasting stage of electric heating tables from the source, and to accurately identify and distinguish between undistorted valid data and distorted data affected by localized heat radiation and abnormal temperature feedback, thus preventing distorted data from directly entering the forecasting stage and causing increased forecast deviations and scheduling strategy failures, it is necessary to first determine whether the time-series characteristic distortion value is not greater than a predefined time-series characteristic distortion value. If so, the corresponding load time-series data is marked as compliant load time-series data and enters the load forecasting stage of the electric heating table. If not, an impact analysis of the time-series characteristics is performed, where the predefined time-series characteristic distortion value is represented by the average value of time-series characteristic distortion values ​​over a historical period.

[0035] The specific process for analyzing the impact of execution time-series characteristics is as follows: First, an analysis is conducted on the interference of local thermal radiation within the preset area of ​​the electric heating table on the load time-series data. The average value of the thermal radiation flux density monitored by the thermal radiation meter in each preset area is used as the local thermal radiation-time-series deviation value. This value is an indicator of the interference intensity of local thermal radiation on the time-series characteristics of the load time-series data. Second, an analysis is conducted on the interference of temperature feedback delay within the preset area of ​​the electric heating table on the load time-series data. The time difference between the moment the electric heating table receives the temperature adjustment command and the moment the sensor collects the corresponding temperature, as monitored by the timer, is used as the abnormal temperature feedback-time-series deviation value. This value is an indicator of the interference intensity of temperature feedback delay on the time-series characteristics of the load time-series data.

[0036] To accurately quantify the respective contributions of local thermal radiation interference and abnormal temperature feedback interference to the time-series characteristic distortion of electric heating tables, and to identify the dominant factors among the two types of interference, the acquired local thermal radiation-time-series deviation values, abnormal temperature feedback-time-series deviation values, and time-series characteristic distortion values ​​are input into a predefined time-series characteristic distortion influence weight table. The output values ​​are the degree to which the local thermal radiation-time-series deviation values ​​and abnormal temperature feedback-time-series deviation values ​​affect the time-series characteristic distortion of electric heating tables, respectively, and are denoted as the local thermal radiation influence value and the abnormal temperature feedback influence value. The magnitudes of the local thermal radiation influence value and the abnormal temperature feedback influence value are then compared.

[0037] It should be noted that within the scope of research on the predictive power load of electric heating tables, the primary step is to construct a data architecture system with multi-dimensional mapping characteristics, supporting two distinct forms of mapping relationships: one-to-one precise mapping and many-to-one composite mapping. This ensures accurate data conversion across different dimensional spaces, laying a solid data foundation for subsequent power load prediction work. The plan is for personnel to focus on diverse information within specific historical interaction time periods, conducting comprehensive and systematic data collection. The scope of the original electric heating table time-series information collected is extremely broad, specifically encompassing parameters such as local thermal radiation-time-series deviation values, abnormal temperature feedback-time-series deviation values ​​and time-series characteristic distortion values, data sets of local thermal radiation-time-series deviation values, abnormal temperature feedback-time-series deviation values ​​and time-series characteristic distortion values, data combinations of the highest-level time-series determination value and the signal-to-noise ratio of load time-series data, load prediction exceedance values, combinations of initial electric heating table power and initial temperature values, and combinations of local thermal radiation-time-series deviation values ​​and time-series characteristic distortion values ​​under high-power steady-state operation scenarios.

[0038] The raw time-series information of electric heating tables collected undergoes a rigorous preprocessing process, including data cleaning to remove noise interference and data normalization to unify data units, thereby improving data quality and usability. Subsequently, this preprocessed historical key information is input into a machine learning model with feature importance assessment capabilities. Among numerous machine learning models, the random forest model, due to its superior feature selection ability and accurate classification prediction ability, has been widely used in this field. By constructing multiple decision trees and combining the prediction results of each decision tree, the random forest model can accurately screen out feature variables that have a key impact on the prediction of electric heating table electricity load, thus providing strong support for subsequent predictive analysis. Through continuous iterative optimization of the location of cluster centers and the range of cluster boundaries, data within the same cluster exhibit a higher degree of similarity in features, while data between different clusters show more significant differences in features. This helps to accurately extract data groups with similar characteristics from massive amounts of data, uncover the potential patterns and rules hidden behind the data, and thus obtain a time-series mapping set of electric heating tables.

[0039] The electric heating table time-series mapping set includes a combination of local thermal radiation influence values ​​and abnormal temperature feedback influence values, a preset number of sampling points, the number of decomposition layers corresponding to the load time-series data, a combination of qualified temperature values ​​and electric heating table power reduction values, and local thermal radiation interference level values, etc.

[0040] The system will perform a key data association and matching operation, which will accurately associate and match the raw data obtained from historical monitoring with the parameters output by the model after iterative optimization, to form a time series table or set of electric heating tables.

[0041] Among them, the time series tables or sets for electric heating tables include a weight table for the influence of time series feature distortion, a selection set for smoothing processing, a selection set for wavelet decomposition layer number, a temperature-power value adjustment table, and a weight table for the influence of thermal radiation, etc.

[0042] Once the timing information of the electric heating table is captured by the system, this information is then input into a pre-built timing table or set for the electric heating table. Given that the constructed data architecture has multi-dimensional mapping characteristics, this characteristic supports two different forms of mapping relationships: one-to-one precise mapping and many-to-one composite mapping.

[0043] Under this feature, the input information will be accurately matched and transformed in the data architecture according to the established mapping rules, and the system can quickly and accurately extract the corresponding electric heating table time-series mapping set, etc.

[0044] Specifically, when the local thermal radiation impact value exceeds the upper limit of the impact range defined by the abnormal temperature feedback impact value, the load time series data under local thermal radiation interference is marked. Among them, the abnormal temperature feedback impact value is within the impact range defined by the abnormal temperature feedback impact value set by preset personnel based on historical experience. This helps to improve the identification accuracy of distorted data dominated by PID scheduling response delay, provides accurate data support for optimizing the temperature feedback mechanism and shortening the response delay, and ensures the data repair effect under delay interference.

[0045] When the local thermal radiation impact value is less than the lower limit of the impact range defined by the abnormal temperature feedback impact value, the load time series data under the abnormal temperature feedback is marked; after marking the load time series data, it is determined whether to enter the load prediction stage of the electric heating table; this helps to improve the identification accuracy of distorted data dominated by PID scheduling response delay, provides accurate data support for optimizing the temperature feedback mechanism and shortening the response delay, and ensures the data repair effect under delay interference.

[0046] When the local thermal radiation impact value is within the influence range defined by the abnormal temperature feedback impact value, it directly enters the thermal radiation-regulation coordinated interference adjustment, which helps to improve the anti-interference capability of the double interference superposition scenario, simultaneously suppresses the superposition effect of thermal radiation and temperature feedback delay, and efficiently restores the periodicity and correlation of load time series data; after the thermal radiation-regulation coordinated interference adjustment is completed, it is determined whether the re-acquired time series characteristic distortion value is not greater than the predefined time series characteristic distortion value. If so, it enters the load prediction stage of the electric heating table; if not, the corresponding load time series data is removed.

[0047] The reason for comparing the range of abnormal temperature feedback influence corresponding to the values ​​of local thermal radiation influence and abnormal temperature feedback influence, rather than directly comparing the values ​​of local thermal radiation influence and abnormal temperature feedback influence, is as follows: Directly comparing the magnitudes of the two values ​​could easily lead to the mistaken identification of the interference source with the "slightly higher value" as the sole dominant factor, ignoring the superimposed influence of the other interference. By using the range of abnormal temperature feedback influence as a reference, we can accurately classify three scenarios: "thermal radiation dominant," "temperature feedback dominant," and "double superposition," avoiding severing the coupling relationship between interferences. When the thermal radiation influence exceeds the upper limit of the range, thermal radiation (dominant interference) is addressed first. When it is below the lower limit of the range, temperature feedback delay is addressed first. When it is within the range, coordinated adjustment is initiated to avoid single adjustment deviations caused by direct comparison, ensuring that the adjustment strategy accurately matches the interference type, while avoiding frequent switching of adjustment measures and ensuring the stability of system operation.

[0048] It should be added that thermal radiation-regulation coordinated interference regulation is used to reduce prediction bias and subsequent energy-saving scheduling imbalance caused by the composite interference formed by the superposition of local thermal radiation interference and abnormal temperature feedback, and to reduce the degree of time series characteristic distortion. The specific process of thermal radiation-regulation coordinated interference regulation is as follows: the data set of local thermal radiation-time series deviation values, abnormal temperature feedback-time series deviation values, and time series characteristic distortion values ​​are input into a predefined smoothing selection set, and a preset number of sampling points are output based on the predefined smoothing selection set; a sliding window with a sampling length of the preset number of sampling points (such as 5-10 sampling points) is obtained. The median of the collected parameters (such as room temperature and power) within the sliding window is taken as the replacement index. This median is used to replace outliers in the sliding window that exceed the standard deviation of the collected parameters by a preset multiple (such as 2 times). This helps to improve the purification effect and stability of load time series data when the time series distortion value is greater than the predefined time series distortion value and the local thermal radiation influence value is within the influence range defined by the abnormal temperature feedback influence value. This helps to reduce the degree of time series distortion under the superposition of double interference, so that the corrected data meets the admission conditions of the prediction process, improves the accuracy of load prediction in this scenario, and ensures the reliability of energy-saving scheduling control.

[0049] In this embodiment, based on the load time-series data characterizing the electricity consumption behavior of the electric heating table, corresponding time-series characteristic interference analysis is performed to obtain the time-series characteristic distortion value. The time-series characteristic distortion value is used as a feedback condition. When the time-series characteristic distortion value is not greater than a predefined time-series characteristic distortion value, the load prediction stage of the electric heating table is entered; otherwise, the influence degree analysis of the time-series characteristics is performed. This helps to control the effectiveness of the load time-series data from the source, accurately distinguish between directly predictable data and distorted data that requires intervention, and avoid invalid data entering the prediction stage, which would lead to an expansion of the deviation. Through the interaction and interrelation between the time-series characteristic interference analysis, the load prediction stage of the electric heating table, and the influence degree analysis of the time-series characteristics, the accuracy of the electric heating table's electricity load prediction and its interference adaptability are improved. This provides solid data and technical support for the formulation of power grid energy-saving dispatch strategies, the guarantee of users' heating needs, and the stable operation of the electric heating table.

[0050] Preferably, in order to avoid the accidental fluctuation of the local thermal radiation-time sequence deviation value within a single power load prediction period causing the load time sequence data to be marked incorrectly, the load time sequence data under local thermal radiation interference is marked. The specific process is as follows: the average value of the local thermal radiation-time sequence deviation value within a preset number of power load prediction periods is marked as the average thermal radiation interference level value.

[0051] For example, when it is necessary to reduce the impact of random fluctuations in a single time period on the assessment of interference level and focus on the overall interference trend, the average value of the local thermal radiation-time sequence deviation values ​​within a predetermined number of randomly selected electricity load prediction time periods can be marked as the average thermal radiation interference level value. This helps to avoid assessment bias caused by systematic errors in continuous time periods and improves the objectivity and comprehensiveness of the determination of thermal radiation interference level.

[0052] When it is necessary to accurately capture the continuous changing trend of local thermal radiation interference (such as the continuous increase or decrease of interference intensity under high power steady-state operation), the average value of the local thermal radiation-time sequence deviation values ​​within the predicted time period of a continuous preset number of power loads is marked as the average thermal radiation interference level value. This helps to track the dynamic changes of interference intensity in real time and identify the cumulative effect of interference in a timely manner.

[0053] The average thermal radiation interference level is used as the criterion for determining whether to mark the load time series data. If the average thermal radiation interference level is greater than the pre-stored thermal radiation level value, the corresponding load time series data is marked as data to be corrected, and the highest level of time series interference correction is initiated. Otherwise, the regular time series interference correction is initiated. The pre-stored thermal radiation level value is represented by the average value of a preset number of local thermal radiation-time series deviation values ​​randomly selected during the historical electricity load prediction period. After the load time series data under local thermal radiation interference is marked, it is determined whether the re-acquired time series characteristic distortion value is not greater than the predefined time series characteristic distortion value. If so, the load prediction stage of the electric heating table is entered. If not, a local thermal radiation interference abnormality alarm is pushed.

[0054] By labeling load time-series data under localized thermal radiation interference to obtain the average thermal radiation interference level, the highest level of time-series interference correction is initiated when the average thermal radiation interference level is greater than the pre-stored thermal radiation level value; otherwise, regular time-series interference correction is initiated. This helps improve the accuracy of graded correction under thermal radiation interference scenarios, avoids "over-correction" or "under-correction," and maximizes the preservation of the true electricity consumption patterns in the load time-series data, reduces data information loss during the correction process, and provides high-quality data for subsequent prediction stages.

[0055] like Figure 2 The diagram shows a flowchart illustrating the process of marking load time-series data under abnormal temperature feedback according to an embodiment of this invention: The average interference level of the abnormal temperature is obtained; when the average interference level of the abnormal temperature is greater than the upper limit of the pre-stored abnormal temperature feedback deviation interference range, the corresponding load time-series data is removed; when the average interference level of the abnormal temperature is within the pre-stored abnormal temperature feedback deviation interference range, the highest level of time-series interference correction is initiated; when the average interference level of the abnormal temperature is less than the lower limit of the pre-stored abnormal temperature feedback deviation interference range, conventional time-series interference correction is initiated.

[0056] On another level, to avoid labeling bias caused by random fluctuations in the abnormal temperature feedback-time series deviation values ​​within a single electricity load forecast period, the load time series data under abnormal temperature feedback is labeled. The specific process is as follows: The average value of the abnormal temperature feedback-time series deviation values ​​within a preset number of electricity load forecast periods is labeled as the average abnormal temperature interference level value; after obtaining the average abnormal temperature interference level value, it is compared and analyzed with the pre-stored abnormal temperature feedback deviation interference range set by preset personnel; if the average abnormal temperature interference level value is greater than the upper limit of the pre-stored abnormal temperature feedback deviation interference range, the corresponding load time series data is removed; if the average abnormal temperature interference level value is within the pre-stored abnormal temperature feedback deviation interference range, the data is removed. Within the disturbance range, the corresponding load time-series data is marked as data to be corrected, and the highest level of time-series disturbance correction is initiated. If the average disturbance level of abnormal temperature is less than the lower limit of the pre-stored abnormal temperature feedback deviation disturbance range, regular time-series disturbance correction is initiated. After marking the load time-series data under abnormal temperature feedback, the re-acquired time-series characteristic distortion quantification value is evaluated. The evaluation process includes: comparing the re-acquired time-series characteristic distortion value with the predefined time-series characteristic distortion value. If the comparison result is yes, the load prediction stage of the electric heating table is entered; if the comparison result is no, an alarm notification is pushed. Time-series characteristic re-comparison means judging whether the re-acquired time-series characteristic distortion value is not greater than the predefined time-series characteristic distortion value.

[0057] By marking the load time-series data under abnormal temperature feedback to obtain the average disturbance level value of abnormal temperature, and removing the corresponding load time-series data when the average disturbance level value of abnormal temperature exceeds the upper limit of the pre-stored abnormal temperature feedback deviation disturbance range, the purity of the input prediction model is improved, and the impact of severe distortion data caused by extreme delay interference on prediction accuracy is avoided. By initiating the highest level of time-series interference correction when the average disturbance level value of abnormal temperature is within the pre-stored abnormal temperature feedback deviation interference range, the deep purification of temperature feedback delay interference is achieved, and the time-series characteristic distortion caused by hysteretic power fluctuations is efficiently repaired. By initiating regular time-series interference correction when the average disturbance level value of abnormal temperature is less than the lower limit of the pre-stored abnormal temperature feedback deviation interference range, the correction efficiency under mild delay interference is improved, the system computing resource consumption is reduced, and the data correction effect and system operating efficiency are balanced, ensuring the smoothness of the overall processing flow.

[0058] It should be added that the specific process of the highest-level time-series interference correction is as follows: The highest-level time-series determination value and the signal-to-noise ratio data of the load time-series data monitored by the spectrum analyzer and oscilloscope are combined and input into the predefined wavelet decomposition level selection set. The decomposition level corresponding to the load time-series data is output, and the decomposition level is used as the allocation value of the wavelet decomposition process to decompose the load time-series data.

[0059] It should be added that the selection process for the highest-level time series determination value is as follows: If the monitored local thermal radiation influence value is greater than the upper limit of the influence range defined by the abnormal temperature feedback influence value, the average thermal radiation interference level value that is greater than the pre-stored thermal radiation level value is selected as the highest-level time series determination value; if the monitored local thermal radiation influence value is less than the lower limit of the influence range defined by the abnormal temperature feedback influence value, the average abnormal temperature interference level value within the pre-stored abnormal temperature feedback deviation interference range is selected as the highest-level time series determination value. The highest-level time series interference correction is used to adapt to high-intensity interference types and helps to repair severely distorted time series characteristics. Among them, by performing the highest-level time series interference correction to obtain the decomposition level corresponding to the load time series data, and using the decomposition level as the adjustment value of the wavelet decomposition process to decompose the load time series data, it helps to improve the separation accuracy of distorted components and true components in the load time series data, accurately remove the interference components caused by thermal radiation or temperature feedback delay, thereby improving the integrity of the time series regularity of the corrected load time series data and strengthening the ability of the time series prediction model to capture the real electricity consumption pattern.

[0060] Conventional time-series interference correction refers to the direct filtering of the data to be corrected based on conventional filtering algorithms (such as mean filtering, median filtering, etc.) to quickly eliminate small noise in load time-series data and maintain the basic stability of time-series characteristics.

[0061] When there is a small, continuous thermal radiation noise in the load time series data (such as desktop power fluctuating within the preset fluctuation range, but not deviating from the normal range, but the noise signal intensity obtained from continuous sampling points is not within the preset range), mean filtering is used to filter the data to be corrected (mean filtering can smooth continuous noise and retain the overall trend of the load data).

[0062] When load time series data has isolated abrupt changes due to a single adjustment and there is no continuous noise (such as desktop power exceeding the preset fluctuation range, and the noise signal strength obtained at continuous sampling points is within the preset range), median filtering is used to filter the data to be corrected (median filtering can accurately remove isolated outliers and avoid the trend distortion caused by excessive smoothing of abrupt changes by mean filtering).

[0063] In this embodiment, the marking of load time-series data under local thermal radiation interference and the marking of load time-series data under abnormal temperature feedback are targeted measures determined by the correspondence between the influence ranges defined by the influence values ​​of local thermal radiation and abnormal temperature feedback. The two are also specifically related. Based on this interrelationship, it helps to achieve accurate classification, control and targeted identification of dual interference sources, reduce processing deviations caused by interference type confusion, and thus ensure the adaptability and effectiveness of subsequent interference correction measures, providing clear directional guidance for load time-series data purification.

[0064] Preferably, the specific process of the load prediction stage of the electric heating table is as follows: The qualified load time-series data, the average temperature of the tabletop area, and the average temperature of the non-tabletop area (monitored by a temperature sensor) are used as input data for the time-series prediction model. Based on the time-series prediction model, a load prediction value is output, representing the total power consumption for a specified prediction period. The load prediction value is compared with a pre-stored load prediction range set by a preset user. If the load prediction value is less than the upper limit of the pre-stored load prediction range, a load qualification prompt is sent. If it is not less than the upper limit of the pre-stored load prediction range, the current load prediction value is recorded as an unqualified load prediction value, and dynamic adjustment of the load status is performed. After the dynamic adjustment of the load status is completed, the load prediction value is re-output. If the re-output load prediction value is still not less than the upper limit of the pre-stored load prediction range, a load warning is sent; otherwise, a load qualification prompt is sent.

[0065] It needs further explanation that the dynamic adjustment of the load status is performed as follows: if the detected load forecast value is not less than the upper limit of the pre-stored load forecast range, the following processing is performed: the load forecast over-limit value, the initial electric heating table power (the average power of preset points monitored by power sensors), and the initial temperature value (the average temperature of preset points monitored by temperature sensors) are input into the pre-built temperature-power value adjustment table. The corresponding qualified temperature value and the electric heating table power reduction value are output, and a prompt is sent to adjust the temperature to the corresponding qualified temperature value. A prompt is also sent to the preset personnel to adjust the electric heating table power to the target electric heating table power, i.e., the electric heating table power reduction value, based on the initial electric heating table power. The load forecast over-limit value refers to the load forecast value that is not less than the upper limit of the pre-stored load forecast range.

[0066] It should be added that the historical load time series data (such as ambient temperature and humidity, power of the electric heating table, user status, and usage time of the electric heating table), the average temperature of the tabletop area and the average temperature of the non-tabletop area, and the preset load prediction value are combined as the training set of the prediction model. The training set is input into the prediction model (such as bidirectional LSTM) to train the model. The LSTM model's ability to capture the long-term and short-term dependencies of time series data is utilized to learn the correlation between historical load data and temperature distribution characteristics, and the preset load value (such as the time lag relationship between the increase in tabletop temperature and the increase in power, and the triggering pattern of load demand due to the decrease in ambient temperature and humidity). The model parameters (such as the number of hidden layer neurons, learning rate, and number of iterations) are optimized through iterative training to minimize the error between the predicted value and the actual load value, and finally a trained model with stable prediction ability is obtained. The combination of the newly obtained load time series data, the average temperature of the tabletop area and the average temperature of the non-tabletop area is input into the trained model to output the load prediction value.

[0067] like Figure 3 The diagram illustrates the load prediction training provided in an embodiment of this invention: Historical load time-series data, the average temperature of the tabletop area and the average temperature of the non-tabletop area of ​​an electric heating table, and a preset load prediction value (as a training set) are input into a bidirectional LSTM network for training. This captures the dependencies in the time-series data. σ represents the Sigmoid activation function, a key component of LSTM for achieving long-term time-series memory and filtering redundant information. During training, the model uses forward and backward LSTM structures to generate hidden states (current...) using historical and future data respectively. 、 Next period Previous period (etc.), and perform load forecasting, with input data including x t-1 x t x t+1 Equal time series data, where x t The load data (load time series data) represents the current time period. t+1 x represents the load data for the next time period. t-1 The hidden state of the model, representing the load data of the previous time period, is calculated based on the input of the current time period and historical data, reflecting the long-term dependencies in the time series data. Finally, based on the load forecast value of the previous time period (y... t-1 The model outputs the load forecast value for the current time period (y). t ), and further predict future periods (y t+1 ) load demand.

[0068] By adjusting the temperature to the corresponding qualified temperature value and adjusting the power of the electric heating table to the target power when the load forecast value is not less than the upper limit of the pre-stored load forecast range, it helps to improve the timeliness and accuracy of the control response in the case of overload, avoid disorderly power growth from aggravating the load pressure on the power grid, and thus help to achieve the organic unity between the power grid peak shaving and valley filling goals and the heating needs of users, taking into account both energy saving and practicality.

[0069] In this embodiment, by entering the load prediction stage of the electric heating table and obtaining the load prediction value, when the load prediction value is less than the upper limit threshold of the pre-stored load prediction range, a load qualification prompt is sent. If the load prediction value is not less than the upper limit threshold of the pre-stored load prediction range, dynamic adjustment of the load status is performed. This helps to achieve advanced prediction and precise control of the electric heating table's power load, and avoid the risk of grid pressure caused by the superposition of overload limits in advance. Through the connection between the load prediction stage of the electric heating table and the dynamic adjustment of the load status, it helps to further strengthen the closed-loop coordination capability of prediction and control, reduce the scheduling errors caused by the disconnect between prediction and control, and thus ensure the two-way adaptation of grid load operation stability and electric heating table safe operation.

[0070] Example 2, based on the fundamental method of Example 1, in order to adapt to the special operating conditions of electric heating tables under high-power steady-state operation scenarios (such as full-power operation in winter evenings) where local thermal radiation interference is dominant and abnormal temperature feedback interference is aggravated, and to avoid misjudging data validity due to the use of general scenario judgment standards, it is necessary to accurately control whether the load time-series data under this scenario meets the requirements for entering the prediction stage. Specifically, it is necessary to perform corresponding time-series characteristic interference analysis based on the load time-series data characterizing the electricity consumption behavior of the electric heating table. Specifically, it is determined whether the time-series characteristic distortion value obtained under the high-power steady-state operation scenario is not greater than a predefined qualified threshold, where the predefined qualified threshold is represented by the average value of the time-series characteristic distortion values ​​obtained under the high-power steady-state operation scenario over a historical period. If so, the corresponding load time-series data is marked as qualified load time-series data. The data is then fed into the load prediction stage of the electric heating table. If not, the combination of the acquired local thermal radiation-time series deviation value and the time series characteristic distortion value under the high-power steady-state operation scenario is input into a predefined thermal radiation influence weight table, and the corresponding local thermal radiation interference level value is output. If the local thermal radiation interference level value is within the corresponding predefined qualified range set by preset personnel, the load time series data is processed based on the median filtering algorithm, and a decision is made according to established rules as to whether to enter the load prediction stage of the electric heating table. That is, if the time series characteristic distortion value under the high-power steady-state operation scenario acquired after monitoring and processing is not greater than the predefined qualified threshold, the load prediction stage of the electric heating table is entered; otherwise, a time series characteristic abnormality alarm is sent. If the local thermal radiation interference level value is not within the corresponding predefined qualified range, the corresponding load time series data is removed.

[0071] In this embodiment, when the electric heating table operates at full power continuously, such as during peak electricity consumption in winter evenings, the thermal inertia of the heating element increases under full power operation, leading to aggravated abnormal temperature feedback interference. Based on the load time-series data characterizing the electric heating table's electricity consumption behavior, corresponding time-series characteristic interference analysis is performed to obtain the time-series characteristic distortion value under the high-power steady-state operation scenario. When the time-series characteristic distortion value under the high-power steady-state operation scenario is not greater than a predefined qualified threshold, the load prediction stage of the electric heating table is entered. Otherwise, the obtained local thermal radiation interference level value is used as a feedback indicator for further division. When the local thermal radiation interference level value is within the corresponding predefined qualified range, the load time-series data is processed based on the median filtering algorithm. Median filtering is a nonlinear signal processing algorithm. The core logic is: taking the sampling point of the time-series data as the center, a sliding window of fixed length (containing several continuous sampling points) is selected. All data in the window are sorted by size and the median value is taken. This median value is used to replace the original sampling point at the center of the window. The filtering process is completed by traversing the entire time-series data sequence.

[0072] Its advantages lie in its ability to effectively suppress impulse noise and isolated outliers (such as instantaneous temperature or power jumps caused by thermal radiation interference), while preserving the trend characteristics and abrupt change points of load time-series data to the greatest extent (such as real power regulation signals). It avoids the ambiguity of time-series patterns caused by linear filtering, which helps to improve the anti-interference capability of load time-series data in high-power steady-state scenarios, effectively filters out instantaneous outliers caused by thermal radiation, and thus restores the original time-series patterns of the data, ensuring the quality of data entering the prediction stage. When the local thermal radiation interference level is outside the corresponding predefined qualified range, the corresponding load time-series data is removed, which helps to improve the data purity of the input prediction model, eliminate the interference of heavily distorted data, thereby avoiding prediction deviations caused by invalid data and ensuring the accuracy of load prediction results in high-power steady-state scenarios.

[0073] Example 3, based on the basic method of Example 1, aims to overcome the limitations of judging the degree of interference in the special scenario of emergency peak shaving dispatch of the power grid. It uses the number of temperature adjustments to intuitively reflect the impact of abnormal temperature feedback interference on the actual operation of the electric heating table (the number of adjustments exceeding the peak value means that the temperature deviation caused by the PID dispatch response delay has led to frequent power adjustments, which will further exacerbate the overload pressure on the power grid). It accurately identifies high-fluctuation load scenarios requiring emergency control and marks the load time-series data under abnormal temperature feedback, specifically as follows: It monitors whether the number of temperature adjustments monitored by the counter within the power load prediction period exceeds the corresponding adjustment peak value set by preset personnel; if so, it immediately sends a prompt to pause active temperature adjustment and uses the fixed power set by preset personnel as the current power value of the electric heating table, while simultaneously pushing a fault reminder; otherwise, it enters the load prediction stage of the electric heating table.

[0074] In this embodiment, when the power grid is in an emergency peak shaving dispatch phase, such as when the regional power load is overloaded and a scenario requiring accurate load forecasting in conjunction with the power grid's peak shaving strategy is needed, the load time-series data under abnormal temperature feedback is marked and the number of temperature adjustments is obtained. When the number of temperature adjustments exceeds the corresponding adjustment peak, a fixed power is used as the power value of the electric heating table in the current state. This helps to quickly stabilize the electric heating table's power load, avoids power fluctuations caused by frequent temperature adjustments from further exacerbating the power grid load pressure, and thus helps to achieve the goal of emergency peak shaving dispatch and ensure the safe operation of the regional power grid. When the number of temperature adjustments does not exceed the corresponding adjustment peak, the load forecasting stage of the electric heating table is entered. This helps to accurately predict the subsequent power load of the electric heating table based on effective data, providing a reliable basis for the formulation of the power grid's peak shaving strategy and realizing the refinement and scientification of peak shaving dispatch.

[0075] Example 4: Based on the basic method of Example 1, in order to ensure that the load control effect meets the grid dispatching requirements and that the PID adjustment strategy accurately matches the degree of over-limit of the load forecast value, dynamic adjustment of the load status is required. Specifically, the dynamic adjustment of the load status is as follows: the absolute value of the difference between the load forecast value and the upper limit critical value of the pre-stored load forecast range is monitored and used as the input parameter of the PID; the PID outputs temperature and power values, and the output temperature and power values ​​are used as the qualified values ​​of the electric heating table, and the corresponding temperature and power are adjusted to the corresponding qualified values.

[0076] In this embodiment, when the grid load approaches the overload threshold, conventional power limiting or temperature regulation may not be able to balance grid safety and user thermal comfort. It is necessary to preemptively smooth out load peaks by precisely controlling the preheating time. By performing dynamic load status adjustment, the absolute value of the difference between the load forecast value and the upper limit threshold of the pre-stored load forecast range is used as the input parameter of the PID controller. Based on the output temperature and power values, the corresponding adjustment is made as the qualified value of the electric heating table. This helps to achieve early prediction of load peaks, precise peak shifting, and stable regulation, thereby further ensuring the operational safety and stability of the grid when it approaches the overload threshold, while ensuring that users' heating needs are not significantly affected, achieving dual protection of grid safety and user experience.

[0077] like Figure 4 The diagram shows a schematic of the power load prediction and energy-saving scheduling control system for electric heating tables provided in an embodiment of this invention. The system includes: an electric heating table time-series interference characterization module, which, within a set power load prediction scenario, performs corresponding time-series interference analysis based on load time-series data characterizing the power consumption behavior of the electric heating table, and determines whether the conditions for entering the load prediction stage of the electric heating table are met; an electric heating table scheduling control module, which, if the determination result is that the entry conditions are met, predicts the energy usage of the electric heating table using a model capable of capturing time-series dependencies, and implements a power scheduling control strategy accordingly; and an electric heating table time-series adjustment module, which, if the determination result is that the entry conditions are not met, uses the intensity of time-series interference as the basis for adjustment measures, focuses on the accuracy of the model in capturing the time-series dependencies presented by the load time-series data, implements targeted adjustments, and re-determines whether to enter the load prediction stage of the electric heating table after the adjustment is completed.

[0078] The above-disclosed embodiments are merely some examples of the present invention and should not be construed as limiting the scope of the present invention. Therefore, any equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for predicting and controlling the power load of electric heating tables, characterized in that, The method includes: Within the established electricity load forecasting scenario, based on the load time series data characterizing the electricity consumption behavior of the electric heating table, the corresponding time series characteristic interference analysis is performed to determine whether the conditions for entering the load forecasting stage of the electric heating table are met. If the determination result is that the entry condition is met, the energy usage of the electric heating table is predicted by a model with the function of capturing time-series dependencies, and an electricity dispatch control strategy is implemented accordingly. If the determination result is that the entry conditions are not met, the adjustment will be based on the intensity of time-series interference, focusing on the accuracy of the model in capturing the time-series dependence of the load time-series data, implementing targeted adjustments, and re-determining whether to enter the load prediction stage for electric heating tables after the adjustment is completed.

2. The method for predicting and controlling the power load of an electric heating table as described in claim 1, characterized in that, The analysis of time-series disturbances based on load time-series data characterizing the electricity consumption behavior of electric heating tables is as follows: The number of abrupt changes in the load time-series data monitored within the predicted electricity load period is used as the time-series characteristic distortion value. First, determine whether the temporal feature distortion value is not greater than a predefined temporal feature distortion value; If so, mark the corresponding load time series data as qualified load time series data and enter the load prediction stage of the electric heating table; If not, the impact analysis of execution timing characteristics follows the specific process as follows: The average value of thermal radiation flux density in each preset area is used as the local thermal radiation-time series deviation value, which is an index for measuring the interference intensity of local thermal radiation on the time series characteristics of load time series data. The time difference between the moment when the electric heating table receives the temperature adjustment command and the moment when the sensor collects the corresponding temperature is used as the abnormal temperature feedback-time series deviation value. The abnormal temperature feedback-time series deviation value is an indicator that measures the interference intensity of the temperature feedback delay on the time series characteristics of the load time series data. The acquired local thermal radiation-time deviation values, abnormal temperature feedback-time deviation values, and time-series characteristic distortion values ​​are input into a predefined time-series characteristic distortion influence weight table. The output values ​​are the degree to which the local thermal radiation-time deviation values ​​and abnormal temperature feedback-time deviation values ​​affect the time-series characteristic distortion of the electric heating table, respectively denoted as the local thermal radiation influence value and the abnormal temperature feedback influence value. When the local thermal radiation impact value is greater than the upper limit of the impact range defined by the abnormal temperature feedback impact value, the load time series data under the local thermal radiation interference is marked. When the local thermal radiation impact value is less than the lower limit of the impact range defined by the abnormal temperature feedback impact value, the load time series data under the abnormal temperature feedback is marked. After marking the load time series data, it is determined whether to enter the load prediction stage of the electric heating table. When the local thermal radiation influence value is within the influence range defined by the abnormal temperature feedback influence value, it directly enters the thermal radiation-regulation coordinated interference regulation. After the thermal radiation-regulation coordinated interference regulation ends, it is determined whether the re-acquired time series characteristic distortion value is not greater than the predefined time series characteristic distortion value. If so, it enters the load prediction stage of the electric heating table. If not, the corresponding load time series data is removed.

3. The method for predicting and controlling the power load of an electric heating table as described in claim 2, characterized in that, The specific process of thermal radiation-modulation coordinated interference regulation is as follows: By inputting the data set of local thermal radiation-time deviation value, abnormal temperature feedback-time deviation value and time-series characteristic distortion value into a predefined smoothing selection set, a preset number of sampling points are output based on the predefined smoothing selection set; A sliding window with a sampling length of a preset number of sampling points is obtained. The median of the collected parameters within the sliding window is obtained. The obtained median is used as a replacement index to replace outliers within the sliding window that exceed a preset multiple of the standard deviation of the collected parameters.

4. The method for predicting and controlling the power load of an electric heating table as described in claim 2, characterized in that, The specific process for marking the load time-series data under localized thermal radiation interference is as follows: The average value of the local thermal radiation-time sequence deviation values ​​within the preset number of electricity load prediction time periods is marked as the average thermal radiation interference level value. If the average thermal radiation interference level is greater than the pre-stored thermal radiation level, the corresponding load time series data is marked as data to be corrected, and the highest level time series interference correction is initiated; otherwise, the regular time series interference correction is initiated. After the load time series data under local thermal radiation interference is marked, it is determined whether the re-acquired time series characteristic distortion value is not greater than the predefined time series characteristic distortion value. If so, the load prediction stage of the electric heating table is entered; if not, a local thermal radiation interference abnormality alarm is pushed. The specific process for marking the load time-series data under abnormal temperature feedback is as follows: The average value of the abnormal temperature feedback-time deviation values ​​within the preset number of electricity load prediction time periods is marked as the average abnormal temperature disturbance level value. If the average disturbance level of abnormal temperature is greater than the upper limit of the pre-stored abnormal temperature feedback deviation disturbance range, the corresponding load time series data will be removed. If the average disturbance level of abnormal temperature is within the pre-stored abnormal temperature feedback deviation disturbance range, the corresponding load time series data will be marked as data to be corrected, and the highest level time series disturbance correction will be initiated. If the average interference level of abnormal temperature is less than the lower limit of the pre-stored abnormal temperature feedback deviation interference range, start the normal timing interference correction. After the load time series data under abnormal temperature feedback is marked, the re-acquired time series feature distortion quantization value is evaluated. The evaluation process includes: comparing the re-acquired time series feature distortion value with the predefined time series feature distortion value. If the comparison result is yes, the load prediction stage of the electric heating table is entered. If the comparison result is no, an alarm notification is pushed. The re-comparison of temporal features indicates whether the distortion value of the re-acquired temporal feature is not greater than the predefined temporal feature distortion value.

5. The method for predicting and controlling the power load of an electric heating table as described in claim 4, characterized in that, The specific process of the highest-level timing interference correction is as follows: The highest-level time series determination value and the signal-to-noise ratio of the load time series data are combined and input into a predefined set of wavelet decomposition level selection. The decomposition level corresponding to the load time series data is output and used as the allocation value of the wavelet decomposition process to decompose the load time series data. The process for selecting the highest-level timing determination value is as follows: If the monitored local thermal radiation impact value is greater than the upper limit of the impact range defined by the abnormal temperature feedback impact value, the corresponding average thermal radiation interference level value that is greater than the pre-stored thermal radiation level value is selected as the highest-level time sequence determination value. If the monitored local thermal radiation influence value is less than the lower limit of the influence range defined by the abnormal temperature feedback influence value, the average abnormal temperature interference level value within the pre-stored abnormal temperature feedback deviation interference range is selected as the highest-level time series determination value.

6. The method for predicting and controlling the power load of an electric heating table as described in claim 4, characterized in that, The specific process for load prediction of the electric heating table is as follows: The load time series data that meets the standard, the average temperature of the tabletop area and the average temperature of the non-tabletop area of ​​the electric heating table are used as input data for the time series prediction model, and the load prediction value is output based on the time series prediction model. Compare the load forecast value with the pre-stored load forecast range. If the load forecast value is less than the upper limit of the pre-stored load forecast range, send a load qualification prompt. If it is not less than the upper limit of the pre-stored load forecast range, record the current load forecast value as an unqualified load forecast value and perform dynamic adjustment of the load status. After the load status dynamic adjustment is completed, the load forecast value is re-outputted. If the re-output load forecast value is still not less than the upper limit of the pre-stored load forecast range, a load warning is sent; otherwise, a load qualified prompt is sent.

7. The method for predicting and controlling the power load of an electric heating table as described in claim 6, characterized in that, The dynamic adjustment of the execution load status specifically refers to: Input the predicted overload value, the initial electric heating table power, and the initial temperature value into the pre-built temperature-power value adjustment table, output the corresponding qualified temperature value and the electric heating table power reduction value, and send a prompt to adjust the temperature to the corresponding qualified temperature value. Send a prompt to the preset personnel to adjust the electric heating table power to the electric heating table power reduction value based on the initial electric heating table power. The load forecast over-limit value refers to a load forecast value that is not less than the upper limit critical value of the pre-stored load forecast range.

8. The method for predicting and controlling the power load of an electric heating table as described in any one of claims 1-7, characterized in that, The analysis of time-series disturbances based on load time-series data characterizing the electricity consumption behavior of electric heating tables is as follows: Determine whether the time-series feature distortion value obtained under high-power steady-state operation scenario is not greater than a predefined qualified threshold; If so, mark the corresponding load time series data as qualified load time series data and enter the load prediction stage of the electric heating table; If not, the combination of the obtained local thermal radiation-time deviation value and the time-series characteristic distortion value under the high-power steady-state operation scenario is input into the predefined thermal radiation influence weight table, and the corresponding local thermal radiation interference level value is output. If the local thermal radiation interference level is within the corresponding predefined qualified range, the load time series data is processed based on the median filtering algorithm, and a decision is made on whether to enter the load prediction stage of the electric heating table according to the established rules. If the local thermal radiation interference level is not within the corresponding predefined acceptable range, the corresponding load time series data will be removed.

9. The method for predicting and controlling the power load of an electric heating table as described in any one of claims 1-7, characterized in that, The specific steps for marking load time-series data under abnormal temperature feedback are as follows: Monitor whether the number of temperature adjustments during the predicted electricity load period exceeds the corresponding peak adjustment value; If yes, immediately send a prompt to pause active temperature adjustment, and use a fixed power as the current power value of the electric heating table, while also pushing a fault reminder; otherwise, enter the load prediction stage of the electric heating table.

10. A power load prediction and energy-saving dispatch control system for electric heating tables, characterized in that, include: The electric heating table timing interference characterization module performs corresponding timing characteristic interference analysis based on the load timing data characterizing the electric heating table's power consumption behavior within the set power load prediction scenario, and determines whether the conditions for entering the load prediction stage of the electric heating table are met. The electric heating table scheduling and control module is used to predict the energy usage of the electric heating table through a model with the function of capturing time-series dependencies if the determination result is that the entry conditions are met, and to implement the power consumption scheduling and control strategy accordingly. The electric heating table timing adjustment module is used to adjust the load prediction stage based on the intensity of timing characteristic interference if the determination result is that the entry conditions are not met. It focuses on the accuracy of the model in capturing the timing dependency relationship presented by the load timing data, implements targeted adjustment, and re-determines whether to enter the load prediction stage of the electric heating table after the adjustment is completed.

Images