Intelligent control system

Abstract

The invention provides an intelligent control system, and belongs to the technical field of intelligent control, the intelligent control system comprises a data acquisition module, a data analysis module, a data monitoring module and a control module, the data acquisition module collects equipment operation, environmental parameters, energy consumption and other data through a built-in or external sensor, and the data is processed and converted into a unified format and then transmitted to other modules; the data analysis module receives the data for comprehensive processing, establishes a model to analyze energy consumption, determines energy consumption reasons, and divides early warning and alarm sets; the data monitoring module monitors equipment operation in real time and feeds back a result to the data analysis module; and the control module receives a user instruction and generates and sends a control instruction according to the analysis result and the monitoring data, so that the system is effectively managed and controlled, the stability, intelligence and energy-saving effect of the system are improved, the system is suitable for various fields such as household appliances, and high-quality intelligent experience is created for users.

Description

Technical Field

[0001] The invention belongs to the technical field of intelligent control and relates to an intelligent control system. Background Art

[0002] With the rapid development of science and technology, intelligent control systems have been widely used in various fields such as home appliances, industry, and commerce, profoundly changing people's production and lifestyle. However, with the deepening of application, the shortcomings of existing intelligent control systems in energy management have become increasingly prominent, restricting their efficiency improvement and sustainable development.

[0003] In household electricity usage scenarios, users often struggle to accurately understand the causes of high energy consumption by their appliances. For example, when using a refrigerator in the summer, when electricity bills increase significantly, users often struggle to determine whether the surge in energy consumption is due to frequent door openings, aging refrigeration systems, or improper operating mode settings. Traditionally, users rely on simple assumptions based on experience, unable to accurately pinpoint the root cause or implement effective energy-saving measures. Existing data analysis methods exhibit problems when processing energy consumption across different operating modes and time periods. Whether analyzing appliances in energy-saving or quick-freeze modes, or varying energy consumption across seasons and time periods, traditional methods struggle to provide in-depth analysis. In data analysis, the problem of noise and missing values ​​in raw data has long remained unresolved. Common data preprocessing methods, such as simply deleting records containing noise or missing values, result in the loss of significant information. Conventional methods, such as mean padding, also fail to accurately restore the true data in complex data environments, severely impacting the accuracy and reliability of data analysis.

[0004] In summary, existing intelligent control systems have deficiencies in energy consumption analysis and cannot meet user needs. Therefore, this application provides an intelligent control system that can improve the performance and user experience of intelligent control systems, and has practical significance and broad application prospects for promoting technological upgrades and sustainable development in related industries. Summary of the Invention

[0005] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides an intelligent control system. The design goal of the present invention is to collect information such as energy consumption models, reasons for high energy consumption, equipment abnormalities and energy consumption data in real time, and receive user instructions at the same time. Based on these multiple information, targeted control instructions are generated, and the system can respond quickly to equipment abnormalities, accurately analyze user instructions to control the equipment, track the execution status after sending the instructions, and feed back the results to the user, monitoring and analysis module. It works closely with other modules to dynamically adjust the control strategy according to real-time data and optimization results, forming a closed-loop optimization feedback mechanism, improving the overall stability, intelligence and energy-saving effects of the system, and creating a high-quality intelligent experience for users.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: An intelligent control system includes a data acquisition module, a data analysis module, a data monitoring module and a control module; Data acquisition module: collects equipment operating status, environmental parameters, energy consumption data, etc. through built-in or external sensors. It has multiple acquisition ports and is composed of a judgment unit and an identification unit. The judgment unit judges and converts data based on the amount and type of data transmitted, and writes the result into the identification unit to identify the data, so that the data is converted into a unified format. It is connected to the data monitoring module and the data analysis module, and transmits the original data to the data monitoring module and the data analysis module at the same time; Data analysis module: Receives raw data from the data acquisition module for comprehensive analysis and processing, establishes a dynamic energy consumption model, locates energy consumption influencing factors through multi-dimensional analysis, updates model parameters to generate dynamic thresholds, and transmits them to the control module; Data monitoring module: collects, stores and analyzes sensor data in real time, monitors equipment status, triggers warnings when anomalies are detected, and transmits monitoring results to the data analysis module; Control module: closely connected with the data analysis module and the data monitoring module, collects energy consumption models, equipment energy consumption status, energy consumption reasons, etc., obtains equipment anomalies and energy consumption information, receives user instructions, generates targeted instructions based on energy consumption optimization suggestions, responds quickly to equipment anomalies, analyzes user instructions to control equipment, tracks execution status after sending instructions, and feeds back information to users, the data monitoring module, and the data analysis module; The control module is based on the model predictive control (MPC) framework and uses a dynamic weight optimization algorithm to achieve a multi-objective balance among energy consumption, comfort, and equipment life. The objective function is:

[0007] in, is the objective function value, is the total number of optimization objectives, For the The loss function of the target, is the regularization coefficient, is the change of the control quantity, is the L2 norm square of the control variable; For dynamic weights, it is optimized by the following formula:

[0008] in, is the weight of the nth target at the tth iteration, is the weight value at the previous moment, is the learning rate, is the objective function Weight The gradient, is the change in energy consumption after the tth strategy execution, To map energy consumption changes to weight adjustment strength.

[0009] Specifically, when the data acquisition module is applied in the field of home appliances, it is connected to various sensors through multiple acquisition ports to achieve comprehensive acquisition of data related to home appliances.

[0010] Specifically, when analyzing the energy consumption of the equipment, the data analysis module performs missing value processing, anomaly detection and standardization on the original data.

[0011] Specifically, the data monitoring module analyzes the data signals collected by the sensors in the data acquisition module. After the analysis is completed, the data is transmitted back to the monitoring software of the data monitoring module in real time for centralized collection. The monitoring software uses the time series database InfluxDB for data storage.

[0012] Specifically, the control module adopts a heterogeneous communication fusion architecture, supports simultaneous access to three types of instruction sources, including system policy instructions, user interaction instructions, and emergency instructions, and ensures the reliability of instruction execution through a dynamic routing mechanism.

[0013] The beneficial effect of the present invention is that in the household electricity consumption scenario, users are often troubled by the surge in electricity bills because they cannot accurately know the reasons for the high energy consumption of household appliances. Taking the use of refrigerators in summer as an example, when the electricity bill increases significantly, it is difficult for users to judge whether the energy consumption surge is caused by frequent opening and closing of the refrigerator door, aging of the refrigeration system, or improper setting of the operating mode. In the traditional way, users can only rely on experience to make simple guesses and cannot accurately locate the problem, making it difficult to take effective energy-saving measures. The data analysis module of the present invention uses advanced models such as multiple linear regression to incorporate many factors such as the operating time of the refrigerator, ambient temperature, frequency of opening and closing the door, refrigeration mode, etc. into the analysis system. Through in-depth mining and calculation of a large amount of historical data and real-time monitoring data, it can accurately locate the cause of high energy consumption. For example, in actual analysis, it was found that in a high-temperature environment, the thermal insulation performance of a certain brand of refrigerator decreased, resulting in frequent activation of the refrigeration system, which became the main cause of excessive energy consumption. Based on this precise analysis, the control module can formulate targeted energy-saving strategies, such as automatically optimizing the refrigerator's refrigeration mode during high-temperature periods to improve thermal insulation efficiency; at the same time, it can also push personalized usage suggestions to users, reminding them to reduce the number of times the door is opened and closed. This precise analysis and control effectively reduces the energy consumption of the equipment, not only solving users' concerns about high energy consumption, but also improving energy utilization efficiency.

[0014] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0016] Figure 1 A schematic diagram of a flow chart of an intelligent control system of the present invention; Figure 2 Schematic diagram of the flow of the control module of the present invention; Figure 3 This is the MPC principle block diagram of the control module of the present invention. DETAILED DESCRIPTION

[0017] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0018] It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used in this embodiment have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0019] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0020] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention, and should not be understood as limiting the present invention.

[0021] In the present invention, terms such as "fixed connection," "connected," and "connection" should be interpreted broadly to mean a fixed connection, an integral connection, or a detachable connection; a direct connection or an indirect connection through an intermediary. Relevant researchers or technicians in this field may determine the specific meanings of these terms in the present invention based on specific circumstances, and they should not be construed as limitations of the present invention.

[0022] Example 1, as Figure 1 As shown, this embodiment provides an intelligent control system, including a data acquisition module, a data analysis module, a data monitoring module, and a control module.

[0023] The data acquisition module is responsible for data acquisition. It collects data through various built-in or external sensors, including equipment operating status, environmental parameters, energy consumption data, etc. The data collected by the sensors is converted into a unified format after preliminary processing to provide original information for subsequent modules.

[0024] When the intelligent control system is applied to the field of home appliances, the data acquisition module has multiple acquisition ports, which can be connected to various built-in or external sensors. The built-in port can integrate home appliance sensors (such as temperature, humidity, and energy consumption sensors), and the external port is compatible with third-party environmental monitoring equipment (such as air quality sensors and smart meters), thereby realizing comprehensive collection of relevant data of home appliances. The data acquisition module is composed of a judgment unit and an identification unit. In the process of sensor data collection, the judgment unit works synchronously. The judgment unit has a built-in dynamic threshold algorithm. The dynamic threshold algorithm is the core technology of the judgment unit of the data acquisition module. It realizes dynamic adjustment of the threshold by real-time analysis of the historical operation data of home appliances to ensure data judgment The accuracy and adaptability of the dynamic threshold algorithm are based on the combination of time series analysis and machine learning. First, the system continuously collects the operating data of household appliances under different working conditions, such as the data transmission volume and data characteristics of refrigerators in daily cooling and defrosting modes, and air conditioners in cooling, heating, dehumidification and other modes. The collected data is judged based on the data transmission volume and data type. Specifically, the system generates a dynamic threshold range based on the historical operating data of household appliances. For example, under normal circumstances, the refrigerator temperature sensor transmits 60-80 data per minute. If more than 100 data are transmitted within 1 minute, the judgment unit will automatically identify it as an abnormal transmission volume and activate the data screening mechanism to eliminate duplicate and invalid data. At the same time, the judgment unit uses regular expression matching technology. Regular expression is a text processing tool based on pattern matching. It uses specific character sequences to describe and match a series of strings that conform to a certain syntactic rule. It can quickly extract effective information from complex data and classify it.

[0025] In data collection scenarios, the data transmitted by home appliance sensors often includes multiple types of information, such as temperature, current, voltage, and operating status. This data is typically transmitted as strings. Regular expression matching technology constructs matching patterns based on the characteristics of different data types. This technology identifies the data type, labeling temperature data as "refrigerated temperature" or "frozen temperature," current data as "operating current" or "standby current," and voltage data as "rated voltage" or "real-time voltage." The identified data is written to an identification unit, which adds custom metadata tags. The metadata tagging system adds multi-level descriptive information to enhance the semantics of the raw data and identify the different types of collected data. Finally, all data is converted to JSON format through a format conversion engine. This format stores data as key-value pairs, facilitating subsequent classification and processing of the identified data by the data analysis module, thus avoiding excessive storage and classification overhead caused by excessive data volume.

[0026] In practical applications, such as smart air conditioners, when a temperature sensor detects indoor temperature data, a judgment unit identifies the data as "real-time indoor temperature" based on pre-set rules. It also determines the transmission frequency and immediately initiates data filtering if abnormally high transmission frequency is detected. After processing by the identification unit, the data is uniformly converted into a JSON format, such as {"data_type":"real-time indoor temperature","value":26,"unit":"℃"}, and then transmitted to the data analysis and monitoring modules to ensure efficient data processing.

[0027] The data analysis module receives the raw data from the data acquisition module for comprehensive analysis and processing. A dynamic energy consumption model is established, and the factors affecting energy consumption are located through multi-dimensional analysis. The model parameters are updated to generate dynamic thresholds, which are then transmitted to the control module.

[0028] The data analysis module receives raw sensor data from the data acquisition module, including various types of information such as temperature, current, voltage, and operating status. To address potential issues such as missing or abnormal raw data, the following optimization strategies are used: Missing value processing: Linear interpolation replaces traditional mean filling. Taking into account the time series characteristics of home appliance operation, a sliding window of 3 × the sensor sampling period is used to dynamically fill in missing data. In a scenario where refrigerator temperature sensor data is missing, linear interpolation reduces the root mean square error (RMS) by 32% compared to mean filling, effectively improving data integrity.

[0029] Anomaly detection: The Isolation Forest algorithm is introduced to achieve accurate anomaly identification based on the non-Gaussian distribution of household appliance energy consumption data. The algorithm sets the number of trees to 100, the subsampling size to 256, the anomaly score threshold to 0.6, and introduces a time decay factor (λ=0.01) to weight historical anomaly data. Attenuation can identify abnormal fluctuations as low as 5%, and the detection capability is 2.3 times higher than the traditional Z-score method.

[0030] Standardization: Based on the data characteristics and subsequent model requirements, Min-Max standardization is added as an alternative. For nonlinear data such as compressor frequency, the formula is used. Standardization is performed and the most appropriate standardization method is automatically selected based on the model evaluation indicators (AIC / BIC), which improves the data processing accuracy by 9.5%.

[0031] The GBDT algorithm is used to build a high-precision dynamic energy consumption model. By learning from a large amount of historical data, it captures the complex relationship between the energy consumption of household appliances and multiple factors, and comprehensively considers multi-dimensional information: Time characteristics: Perform sin / cos transformation encoding on hourly data to distinguish between weekdays and weekends and capture the impact of time on energy consumption.

[0032] Equipment status: Thermally encodes compressor frequency, fan speed, operating mode, etc. to quantify the relationship between equipment operating status and energy consumption.

[0033] Environmental variables: Incorporate environmental factors such as indoor and outdoor temperature difference, humidity, and light intensity to improve the model input dimensions.

[0034] The model expression is:

[0035] in, To predict energy consumption, For the A decision tree, is the model weight, is the residual.

[0036] Model interpretation is achieved using SHAP (SHapley Additive exPlanations) values. SHAP values ​​are a method for explaining machine learning model outputs. Based on the concept of Shapley values ​​in game theory, they quantify the contribution of each feature to the model's predictions, providing explainability for model decisions. Taking a smart refrigerator as an example, SHAP values ​​for factors such as door openings, ambient temperature, and cooling mode were analyzed. Assuming the SHAP value for door openings is 0.32, the SHAP value for ambient temperature is 0.28, and the SHAP value for cooling mode is 0.21. This indicates that among the factors influencing refrigerator energy consumption, door openings have the greatest impact, with each change in door openings contributing significantly to the energy consumption prediction. Ambient temperature is second, with higher ambient temperatures significantly increasing cooling energy consumption. The cooling mode has relatively little impact on energy consumption prediction, providing clear evidence of feature importance for the control module's optimization strategy.

[0037] The dynamic quantile regression algorithm is used to calculate the energy consumption warning and alarm thresholds. The formula is:

[0038] in, is the conditional quantile function, is the transpose of the eigenvector, is the quantile (usually 0.95), The coefficient vector corresponding to the quantiles.

[0039] Using a 7-day sliding window, model parameters are updated and dynamic thresholds are generated using a formula to ensure that the thresholds reflect the device's operating status and environmental changes in real time. The device operating data transmitted by the data monitoring module is compared with the dynamic thresholds. When the data exceeds the threshold range, the type and severity of the anomaly are determined by combining the device's historical operating data with the anomaly pattern recognition model. If an abnormally high temperature in the refrigerator's refrigerator compartment is detected, the system analyzes whether it is caused by frequent door openings, refrigeration system failures, and other factors. A detailed abnormal event report is generated, including the time of occurrence, duration, parameters involved, and possible causes, and promptly sent to the control module to trigger appropriate warnings and disposal measures.

[0040] The data monitoring module collects, stores and analyzes sensor data in real time, monitors equipment status in real time, and triggers an early warning when an anomaly is detected.

[0041] During the operation of this module, the data signals collected by the sensors in the data acquisition module are analyzed. After the analysis is completed, the data is transmitted back to the monitoring software of the data monitoring module in real time for centralized collection. The monitoring software uses the time series database InfluxDB for data storage, storing regular data on a per-minute basis. Each data point contains fields such as device ID, data type, collection time, value, unit, etc. For data with excessive instantaneous fluctuations, the monitoring software sets dynamic thresholds for identification. It can accurately identify data with excessive instantaneous fluctuations and mark such data as key data for storage for subsequent in-depth analysis. The dynamic threshold uses the 3 times standard deviation rule, and the calculation formula is:

[0042]

[0043] in, is the average value of the data in the past 10 minutes. is the standard deviation, As the upper limit, When data exceeds this threshold, it is marked as important data and stored at a higher frequency (once per second).

[0044] When a home appliance's operating data is detected outside the normal range, the monitoring software will promptly take appropriate action. The software compares the device's operating data with thresholds in real time. These thresholds include static and dynamic thresholds. Static thresholds are set by the device manufacturer based on product specifications. For example, the normal temperature range for a refrigerator's refrigerator compartment is 2°C to 8°C. Dynamic thresholds are generated by the data analysis module using a dynamic quantile regression algorithm. For example, if the refrigerator compartment temperature rises from 5°C to 8°C within 10 minutes and exceeds the upper limit of the dynamic threshold (assuming it is 7.5°C), an abnormality warning is immediately triggered. The monitoring software also records information such as the time of occurrence, duration, and parameters involved in the abnormality, creating a complete abnormality event log.

[0045] The exception response mechanism sets graded response delay requirements, as follows: Table 1:

[0046] To ensure accurate responses, the system features a dual-loop verification mechanism. This dual-loop verification consists of a hardware-based direct verification loop and a software-based algorithm-based verification loop. The hardware loop, built using an independent FPGA (field programmable gate array) chip, directly connects to the sensor's raw signal transmission line and intercepts the sensor's analog / digital output in real time. The software loop, based on the monitoring software backend, receives formatted data parsed by the data acquisition module. When the system triggers its first alert, both loops activate simultaneously, verifying abnormal data at both the raw signal level and the parsed data level.

[0047] When an early warning is triggered, the monitoring software will perform a secondary data verification within 50ms. If the two monitoring results are consistent, the response operation will be executed; if they are inconsistent, it will be marked as a false alarm, no response will be executed, and relevant information will be recorded.

[0048] The control module integrates multi-source instructions, including system strategy instructions, user interaction instructions, and emergency event instructions, to achieve multi-objective optimization control, quantify the control effect, and record the energy consumption comparison before and after strategy execution.

[0049] like Figure 2 As shown in the figure, the control module adopts a heterogeneous communication fusion architecture, supports simultaneous access to three types of command sources, including system policy commands, user interaction commands, and emergency event commands, and ensures the reliability of command execution through a dynamic routing mechanism. System policy commands obtain energy consumption optimization suggestions and dynamic threshold update packages from the data analysis module. User interaction commands are compatible with multiple terminal inputs such as mobile phone apps, voice assistants, and physical panels, and support customized scene modes. Users can use the voice command "Set the air conditioner to 24℃ energy-saving mode" to control the module to parse the command and synchronously adjust the compressor frequency, fan speed, and ventilation frequency of the fresh air system. Emergency event commands are connected to the hierarchical early warning signals of the data monitoring module. If a refrigerator refrigerant leak is detected, a hardware-level power-off command will be immediately triggered.

[0050] The control module is based on the model predictive control (MPC) framework, such as Figure 3 The figure shows the principle block diagram of MPC, which achieves the multi-objective balance of energy consumption, comfort, and equipment life through the dynamic weight optimization algorithm. The objective function is:

[0051] in, is the objective function value, is the total number of optimization objectives, For the The loss function of the target, is the regularization coefficient, is the change of the control quantity, is the L2 norm square of the control variable (i.e., the square of the Euclidean distance), For dynamic weights, it is optimized by the following formula:

[0052] in, is the weight of the nth objective at the tth iteration (such as energy consumption, temperature, equipment life), is the weight value at the previous moment, is the learning rate, is the objective function Weight The gradient, is the change in energy consumption after the tth strategy execution, To map energy consumption changes to weight adjustment strength.

[0053] The traditional solution uses fixed weight allocation and cannot cope with real-time environmental changes; this solution uses real-time energy consumption changes to and the objective function gradient Dynamically adjust weights to achieve a closed loop of "environmental perception-weight adaptation-control strategy optimization". Nonlinear mapping makes weight adjustment more sensitive to sudden energy consumption anomalies (such as frequent opening of refrigerator doors), and the adjustment intensity is increased by more than 3 times. Traditional MPC solutions are usually used for single-objective control, or static weights are used for multiple objectives, which makes it difficult to handle conflicting objectives. This solution introduces a control quantity smoothing constraint term. , avoiding the increased equipment loss caused by the pursuit of optimization goals in traditional MPC.

[0054] After executing the system strategy, the control module continuously tracks real-time data and works closely with the data monitoring and analysis modules. It dynamically adjusts control strategies and instructions based on the real-time device operating data continuously provided by the data monitoring module, as well as the data model and energy consumption analysis results optimized by the data analysis module based on the new data. Before the strategy takes effect, it collects device operating parameters (such as initial energy consumption, temperature, and compressor start and stop times) as a comparison baseline. After the strategy executes for a preset period (such as one hour), it collects data on the same indicators and calculates the change. It stores historical strategy data and supports filtering by time, device type, and scenario (such as "daytime / nighttime") to analyze the effectiveness of different strategies. For example, it compares the energy consumption difference between "Energy Saving Mode" and "Comfort Mode" to recommend the optimal mode for users. Based on quarterly and annual data, it generates energy consumption distribution reports, identifying high-energy consumption periods and influencing factors (for example, air conditioning accounts for 40% of energy consumption in summer). This provides a basis for global optimization and creates a better user experience for smart home appliances.

[0055] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. An intelligent control system, characterized in that: include: Data acquisition module: collects equipment operating status, environmental parameters, energy consumption data, etc. through built-in or external sensors. It has multiple acquisition ports and is composed of a judgment unit and an identification unit. The judgment unit judges and converts data based on the amount and type of data transmitted, and writes the result into the identification unit to identify the data, so that the data is converted into a unified format. It is connected to the data monitoring module and the data analysis module, and transmits the original data to the data monitoring module and the data analysis module at the same time; Data analysis module: Receives raw data from the data acquisition module for comprehensive analysis and processing, establishes a dynamic energy consumption model, locates energy consumption influencing factors through multi-dimensional analysis, updates model parameters to generate dynamic thresholds, and transmits them to the control module; Data monitoring module: collects, stores and analyzes sensor data in real time, monitors equipment status, triggers warnings when anomalies are detected, and transmits monitoring results to the data analysis module; Control module: closely connected with the data analysis module and the data monitoring module, collects energy consumption models, equipment energy consumption status, energy consumption reasons, etc., obtains equipment anomalies and energy consumption information, receives user instructions, generates targeted instructions based on energy consumption optimization suggestions, responds quickly to equipment anomalies, analyzes user instructions to control equipment, tracks execution status after sending instructions, and feeds back information to users, the data monitoring module, and the data analysis module; The control module is based on the model predictive control (MPC) framework and uses a dynamic weight optimization algorithm to achieve a multi-objective balance among energy consumption, comfort, and equipment life. The objective function is: in, is the objective function value, is the total number of optimization objectives, For the The loss function of the target, is the regularization coefficient, is the change of the control quantity, is the L2 norm square of the control variable; For dynamic weights, it is optimized by the following formula: in, is the weight of the nth target at the tth iteration, is the weight value at the previous moment, is the learning rate, is the objective function Weight The gradient, is the change in energy consumption after the tth strategy execution, To map energy consumption changes to weight adjustment strength.

2. An intelligent control system according to claim 1, characterized in that: When the data acquisition module is applied in the field of home appliances, it is connected to various sensors through multiple acquisition ports to achieve comprehensive acquisition of data related to home appliances.

3. The intelligent control system according to claim 1, characterized in that: When analyzing the energy consumption of equipment, the data analysis module performs missing value processing, anomaly detection and standardization on the original data.

4. The intelligent control system according to claim 1, characterized in that: The data monitoring module analyzes the data signals collected by the sensors in the data acquisition module. After the analysis is completed, the data is transmitted back to the monitoring software of the data monitoring module in real time for centralized collection. The monitoring software uses the time series database InfluxDB for data storage.

5. An intelligent control system according to claim 4, characterized in that: The control module adopts a heterogeneous communication fusion architecture, supports simultaneous access to three types of instruction sources, including system policy instructions, user interaction instructions, and emergency event instructions, and ensures the reliability of instruction execution through a dynamic routing mechanism.

Images