Power supply intelligent control management system based on dynamic load identification

Abstract

The invention discloses a power supply intelligent control management system based on dynamic load identification. The system comprises a data acquisition module, a signal processing module, a dynamic load identification module, a control decision module, an execution and feedback module, a man-machine interaction module and a remote communication module. The data acquisition module is used for performing high-frequency sampling on voltage, current, power factor, active power, reactive power and current waveform data in a power supply circuit; and the signal processing module is used for preprocessing the collected data, including filtering, denoising, normalization and feature extraction. The limitation that a traditional power supply control system can only execute static timing or coarse-grained control is broken through, and dynamic identification, intelligent management and remote interaction control of loads are realized by fusing multiple intelligent algorithms and deep customizable strategies; the method has remarkable technical progress and practical application value in the aspects of energy conservation, safety, intelligence and user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of power intelligent control, and in particular to a power supply intelligent control and management system based on dynamic load identification. Background Art

[0002] With the rapid growth of intelligent power consumption, traditional power management systems are struggling to meet the requirements for efficiency, safety, and energy conservation. Existing systems, mostly based on timing control or static identification strategies, are unable to accurately perceive the behavioral characteristics of different load types, especially the dynamically changing current and voltage characteristics, resulting in delayed or inaccurate control strategies.

[0003] While current smart sockets or energy consumption monitoring devices can perform simple load identification, their identification capabilities are weak, their response speed is slow, and they lack intelligent decision-making mechanisms, making them incapable of real-time management and precise control of multiple loads switching at high frequencies. Therefore, a power supply intelligent control and management system based on dynamic load identification is proposed. Summary of the Invention

[0004] (1) Technical problems solved

[0005] In response to the shortcomings of the existing technology, the present invention provides a power supply intelligent control and management system based on dynamic load identification, which solves the problem that although the smart sockets or energy consumption monitoring devices in the existing technology can perform simple load identification, they have weak identification capabilities, slow response speeds, and lack intelligent judgment mechanisms, making it impossible to achieve real-time management and precise control during high-frequency switching of multiple loads.

[0006] (2) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a power supply intelligent control and management system based on dynamic load identification, including a data acquisition module, a signal processing module, a dynamic load identification module, a control decision module, an execution and feedback module, a human-computer interaction module and a remote communication module;

[0008] Data acquisition module: used for high-frequency sampling of voltage, current, power factor, active power, reactive power, and current waveform data in the power supply line;

[0009] Signal processing module: pre-processes the collected data, including filtering, denoising, normalization, and feature extraction;

[0010] Dynamic load identification module: uses fusion machine learning including decision tree, support vector machine, convolutional neural network and feature matching algorithm to identify load type and operating status in real time;

[0011] Control decision module: Based on the identification results, it calls the corresponding control strategy library to decide whether to perform power on / off control, power regulation, peak / valley electricity price response, or safe power off on the target load;

[0012] Execution and feedback module: controls the intelligent circuit breaker or relay to execute corresponding actions, and feeds back the execution results and system status to the central controller;

[0013] Human-computer interaction module: realizes system status visualization, user-defined control strategy, historical data query and abnormal alarm through mobile terminal App or Web platform;

[0014] Remote communication module: supports Ethernet, Wi-Fi, NB-IoT, and ZigBee communication protocols to ensure system remote control and cloud data synchronization.

[0015] As a further preferred embodiment of the present invention, the data acquisition module includes a voltage acquisition unit, which performs real-time sampling of the AC voltage of the power supply line, supports 220V / 380V, and has a sampling accuracy of ±0.5%; a current acquisition unit: uses a Hall sensor or current transformer to collect single-phase / three-phase current data; a power factor detection unit: calculates active power, reactive power and power factor for identifying inductive / resistive loads; a waveform acquisition unit: collects current and voltage waveforms for nonlinear feature analysis; an environmental perception unit: is used to collect temperature, humidity, smoke and other information to provide auxiliary judgment for the safe operation of the load.

[0016] As a further preferred embodiment of the present invention, the signal processing module includes a filtering processing unit: using low-pass filtering, wavelet noise reduction and other methods to remove interference signals; a normalization unit: standardizing parameters of different amplitudes to avoid weight imbalance between features; a short-time Fourier transform unit: used to extract time-frequency change features; a feature extraction unit: extracting key features such as surge current, startup time, and waveform spikes; a data compression and encoding unit: used to reduce the amount of transmitted data and speed up processing.

[0017] As a further preferred embodiment of the present invention, the dynamic load identification module includes a load feature modeling unit: constructing multidimensional feature templates of different loads through training samples; a machine learning identification unit: supporting SVM, decision tree, KNN, and CNN algorithm models; a timing matching unit: identifying short-term fluctuating load behavior through DTW; a state identification unit: determining whether the load is in the startup, running, standby, shutdown, or abnormal state; a model update and self-learning unit: supporting online incremental learning and continuously optimizing the recognition accuracy; and an error correction unit: using historical data to perform secondary judgment on the abnormal recognition results.

[0018] As a further preferred embodiment of the present invention, the control decision module includes a policy library management unit: presetting control templates for different devices, such as the energy-saving operation mode of air conditioners; a user policy customization unit: allowing users to set rules; a rule engine unit: generating control commands based on if-else rules or Boolean logic conditions; a behavior analysis unit: analyzing user usage habits and recommending automated energy-saving solutions; and an abnormal response decision unit: triggering a safe power-off command when a power abnormality is detected.

[0019] As a further preferred embodiment of the present invention, the execution and feedback module includes an actuator control unit: controlling relays, circuit breakers, and dimmer terminal devices; an execution confirmation unit: confirming whether the command is successfully executed through current sampling or status signals; a local buffer execution unit: in case of network interruption, executing operations according to the last model or strategy; a control timeout and retry unit: handling abnormal execution failures or response delays; and a feedback upload unit: uploading the execution status to the main control system or cloud platform in real time.

[0020] As a further preferred embodiment of the present invention, the human-computer interaction module includes a mobile App interface unit: supporting iOS and Android platforms to realize remote control and status viewing; a Web platform unit: suitable for desktop browsers, providing an advanced graphical interface; a data visualization unit: displaying power curves, load identification results, and historical energy consumption information; a strategy setting unit: users can customize control rules, time period control, and load priority; an abnormal alarm unit: supporting alarm mechanisms via SMS, WeChat, and push notifications.

[0021] (3) Beneficial effects

[0022] The present invention provides a power supply intelligent control and management system based on dynamic load identification. It has the following beneficial effects:

[0023] Compared with the prior art, the present invention has the following advantages:

[0024] High real-time recognition accuracy: Combining deep learning and time series analysis methods to achieve accurate recognition of complex load states;

[0025] Fast response speed: The system has millisecond-level dynamic response capability, suitable for high-frequency load switching scenarios;

[0026] Significant energy consumption optimization: Power supply strategy is adjusted according to real-time load status, effectively reducing redundant power consumption;

[0027] Strong user-friendliness: provides an intuitive human-computer interaction interface and supports customized control logic;

[0028] Strong adaptability: suitable for various power consumption environments such as homes, office buildings, industrial plants, etc., with good scalability;

[0029] High safety: It can automatically cut off power when abnormal load behavior is detected, such as short circuit, overload, and abnormal temperature rise, reducing the risk of accidents.

[0030] The system of the present invention 1. integrates multiple algorithms such as decision trees, support vector machines (SVM), convolutional neural networks (CNN) and feature matching, and can perform real-time and accurate identification of various electrical loads; it can identify load types (such as inductive / resistive) and states (startup, operation, standby, abnormality), and adapt to complex and dynamic power consumption environments; it uses timing matching (such as DTW) and self-learning mechanisms to greatly improve the stability and generalization ability of identification. 2. It improves the safety and stability of the operation of electrical equipment. The control decision module supports power anomaly monitoring, safe power off, overload protection and other functions; the real-time feedback mechanism ensures the reliability of control execution, and automatically retries when the control fails, effectively preventing potential safety hazards; the environmental perception unit (such as temperature, humidity, smoke) assists in judging the safe operation status of the load, improving the stability of the overall power system. 3. Supports intelligent energy consumption management and policy optimization. The system can independently formulate or recommend optimal control strategies based on load characteristics and user behavior, achieving energy-saving operation. Users can configure intelligent policies such as time period control and load prioritization through the app / web interface, configuring them on demand. A peak-valley electricity price response mechanism effectively reduces electricity costs and improves system economics. 4. Enables visualization and remote intelligent control. The system provides a real-time data visualization interface, including power trend charts, load operation logs, and policy execution records. Support for multiple communication methods such as NB-IoT, Wi-Fi, and ZigBee ensures stable remote connectivity in scenarios such as homes and industrial facilities. Users can remotely monitor system status, adjust policies, and receive abnormality alerts, improving ease of use and proactive intervention capabilities. 5. Flexible architecture adapts to various application scenarios. The modular design allows for flexible deployment in home distribution boxes, commercial buildings, factory distribution systems, and other applications. It supports local edge processing and cloud-based collaboration, enabling both offline autonomy and centralized management through the cloud platform. It is compatible with single-phase and three-phase power grids, supporting 220V / 380V power supply environments, and has excellent versatility and scalability. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a schematic diagram of the system principle framework of the present invention. DETAILED DESCRIPTION

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0033] See also Figure 1 The embodiment of the present invention provides a technical solution: a power supply intelligent control and management system based on dynamic load identification, including a data acquisition module, a signal processing module, a dynamic load identification module, a control decision module, an execution and feedback module, a human-computer interaction module and a remote communication module.

[0034] Data acquisition module: used for high-frequency sampling of voltage, current, power factor, active power, reactive power, and current waveform data in the power supply line;

[0035] Signal processing module: pre-processes the collected data, including filtering, denoising, normalization, and feature extraction;

[0036] Dynamic load identification module: uses fusion machine learning including decision tree, support vector machine, convolutional neural network and feature matching algorithm to identify load type and operating status in real time;

[0037] Control decision module: Based on the identification results, it calls the corresponding control strategy library to decide whether to perform power on / off control, power regulation, peak / valley electricity price response, or safe power off on the target load;

[0038] Execution and feedback module: controls the intelligent circuit breaker or relay to execute corresponding actions, and feeds back the execution results and system status to the central controller;

[0039] Human-computer interaction module: realizes system status visualization, user-defined control strategy, historical data query and abnormal alarm through mobile terminal App or Web platform;

[0040] Remote communication module: supports Ethernet, Wi-Fi, NB-IoT, and ZigBee communication protocols to ensure system remote control and cloud data synchronization.

[0041] The data acquisition module includes a voltage acquisition unit, which samples the AC voltage of the power supply line in real time, supports 220V / 380V, and has a sampling accuracy of ±0.5%; a current acquisition unit: uses a Hall sensor or current transformer to collect single-phase / three-phase current data; a power factor detection unit: calculates active and reactive power and power factor for identifying inductive / resistive loads; a waveform acquisition unit: collects current and voltage waveforms for nonlinear feature analysis; and an environmental perception unit: collects temperature, humidity, smoke and other information to provide auxiliary judgment for safe load operation.

[0042] The signal processing module includes a filtering processing unit: which uses low-pass filtering, wavelet noise reduction and other methods to remove interference signals; a normalization unit: which standardizes parameters of different amplitudes to avoid weight imbalance between features; a short-time Fourier transform unit: which is used to extract time-frequency change features; a feature extraction unit: which extracts key features such as inrush current, start-up time, and waveform spikes; and a data compression and encoding unit: which is used to reduce the amount of transmitted data and speed up processing.

[0043] The dynamic load identification module includes a load feature modeling unit: it constructs multi-dimensional feature templates of different loads through training samples; a machine learning identification unit: it supports SVM, decision tree, KNN, and CNN algorithm models; a timing matching unit: it identifies short-term fluctuating load behavior through DTW; a state identification unit: it determines whether the load is in the startup, running, standby, shutdown, or abnormal state; a model update and self-learning unit: it supports online incremental learning and continuously optimizes the recognition accuracy; and an error correction unit: it uses historical data to make secondary judgments on abnormal recognition results.

[0044] The control decision module includes a policy library management unit: which presets control templates for different devices, such as the energy-saving operation mode of air conditioners; a user policy customization unit: which allows users to set rules; a rule engine unit: which generates control commands based on if-else rules or Boolean logic conditions; a behavior analysis unit: which analyzes user usage habits and recommends automated energy-saving solutions; and an abnormal response decision unit: which triggers a safe power-off command when a power abnormality is detected.

[0045] The execution and feedback module includes an actuator control unit: controls relays, circuit breakers, and dimmer terminal devices; an execution confirmation unit: confirms whether the command is successfully executed through current sampling or status signals; a local buffer execution unit: in the event of network interruption, performs operations according to the last model or strategy; a control timeout and retry unit: handles execution failures or response delay exceptions; and a feedback upload unit: uploads the execution status to the main control system or cloud platform in real time.

[0046] The human-computer interaction module includes a mobile App interface unit: supports iOS and Android platforms to achieve remote control and status viewing; a Web platform unit: suitable for desktop browsers, providing an advanced graphical interface; a data visualization unit: displays power curves, load identification results, and historical energy consumption information; a strategy setting unit: users can customize control rules, time period control, and load priority; an abnormal alarm unit: supports alarm mechanisms such as SMS, WeChat, and push notifications.

[0047] Example 1: Scenario of multi-load identification and control for home users

[0048] This system is embedded in the smart distribution box connected to the user's home, and the household circuit includes multiple electrical appliances such as air conditioners, electric kettles, refrigerators, microwave ovens, etc.

[0049] Step 1: Data Collection

[0050] The system samples each branch through the current transformer and voltage sampling module, and the sampling frequency is set to 10kHz;

[0051] Obtain 20 parameters in total, including current waveform, voltage waveform, effective power, reactive power, etc. for each cycle.

[0052] Step 2: Signal preprocessing

[0053] Perform third-order wavelet noise reduction on the waveform;

[0054] Extract features such as startup surge, current duration, waveform period, etc. and generate feature vectors;

[0055] Step 3: Dynamic load identification

[0056] The system loads the pre-trained CNN model to perform convolution processing on the waveform data and extract convolution features;

[0057] At the same time, it is input into the support vector machine (SVM) classifier for final recognition;

[0058] Identify labels such as "air conditioner: running, cooling mode" and "microwave oven: heating at high power";

[0059] Step 4: Control strategy formulation and implementation

[0060] According to the policy library settings, the air conditioner automatically switches to energy-saving mode after 23:00 at night, and the microwave oven automatically shuts off if it runs continuously for more than 20 minutes;

[0061] The control command executes the relay on-off operation through the local actuator;

[0062] The control result is fed back for confirmation and will automatically retry twice if it fails.

[0063] Step 5: Data upload and user interaction

[0064] All operation records and load status are uploaded to the cloud;

[0065] Users can view power consumption charts and load identification results in real time through the mobile app;

[0066] The user-set "kitchen equipment automatically powers off after 22:00" policy takes effect on the policy management page.

[0067] Example 2: Motor load identification and protection in industrial scenarios

[0068] In small factories, this system is used to monitor the operation of loads such as motors and inverters.

[0069] Detect the initial startup surge of the motor and identify the operating mode;

[0070] If an abnormally prolonged startup time or reduced power factor is detected, the system will determine it as "bearing sticking";

[0071] The safe power-off mechanism is activated, and an alarm is sent to the user and maintenance recommendations are pushed.

[0072] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention. Therefore, from all perspectives, the embodiments should be regarded as illustrative and non-restrictive. The scope of the present invention is defined by the appended claims, not the foregoing description, and it is intended that all variations that come within the meaning and range of equivalents of the claims be included within the present invention. Any reference signs in the claims should not be construed as limiting the claim to which they relate.

[0073] In addition, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A power supply intelligent control and management system based on dynamic load identification, characterized by: It includes data acquisition module, signal processing module, dynamic load identification module, control decision module, execution and feedback module, human-computer interaction module and remote communication module; Data acquisition module: used for high-frequency sampling of voltage, current, power factor, active power, reactive power, and current waveform data in the power supply line; Signal processing module: pre-processes the collected data, including filtering, denoising, normalization, and feature extraction; Dynamic load identification module: uses fusion machine learning including decision tree, support vector machine, convolutional neural network and feature matching algorithm to identify load type and operating status in real time; Control decision module: Based on the identification results, it calls the corresponding control strategy library to decide whether to perform power on / off control, power regulation, peak / valley electricity price response, or safe power off on the target load; Execution and feedback module: controls the intelligent circuit breaker or relay to execute corresponding actions, and feeds back the execution results and system status to the central controller; Human-computer interaction module: realizes system status visualization, user-defined control strategy, historical data query and abnormal alarm through mobile terminal App or Web platform; Remote communication module: supports Ethernet, Wi-Fi, NB-IoT, and ZigBee communication protocols to ensure system remote control and cloud data synchronization.

2. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The data acquisition module includes a voltage acquisition unit that samples the AC voltage of the power supply line in real time, supports 220V / 380V, and has a sampling accuracy of ±0.5%; a current acquisition unit: uses a Hall sensor or current transformer to collect single-phase / three-phase current data; a power factor detection unit: calculates active power, reactive power and power factor for identifying inductive / resistive loads; a waveform acquisition unit: collects current and voltage waveforms for nonlinear feature analysis; and an environmental perception unit: collects information such as temperature, humidity, and smoke to provide auxiliary judgment for safe load operation.

3. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The signal processing module includes a filtering unit: using low-pass filtering, wavelet noise reduction and other methods to remove interference signals; a normalization unit: standardizing parameters of different amplitudes to avoid weight imbalance between features; a short-time Fourier transform unit: used to extract time-frequency change features; feature Extraction unit: extracts key features such as inrush current, startup time, and waveform spikes; data compression and encoding unit: used to reduce the amount of transmitted data and speed up processing.

4. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The dynamic load identification module includes a load feature modeling unit: constructing multi-dimensional feature templates of different loads through training samples; Machine learning recognition unit: supports SVM, decision tree, KNN, and CNN algorithm models; timing matching unit: identifies short-term fluctuating load behavior through DTW; state recognition unit: determines whether the load is in the startup, running, standby, shutdown, or abnormal state; Model update and self-learning unit: supports online incremental learning and continuously optimizes recognition accuracy; error correction unit: uses historical data to make secondary judgments on abnormal recognition results.

5. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The control decision module includes a policy library management unit: which presets control templates for different devices, such as the energy-saving operation mode of air conditioners; a user policy customization unit: which allows users to set rules; a rule engine unit: which generates control commands based on if-else rules or Boolean logic conditions; a behavior analysis unit: which analyzes user usage habits and recommends automated energy-saving solutions; and an abnormal response decision unit: which triggers a safe power-off command when a power abnormality is detected.

6. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The execution and feedback module includes an actuator control unit: controls relays, circuit breakers, and dimmer terminal devices; an execution confirmation unit: confirms whether the command is successfully executed through current sampling or status signals; a local buffer execution unit: when the network is interrupted, the operation is executed according to the last model or strategy; a control timeout and retry unit: handles execution failures or response delay exceptions; and a feedback upload unit: uploads the execution status to the main control system or cloud platform in real time.

7. The power intelligent control and management system based on dynamic load identification according to claim 1, characterized in that: The human-computer interaction module includes a mobile App interface unit: supports iOS and Android platforms to achieve remote control and status viewing; a Web platform unit: suitable for desktop browsers, providing an advanced graphical interface; a data visualization unit: displays power curves, load identification results, and historical energy consumption information; a strategy setting unit: users can customize control rules, time period control, and load priority; an abnormal alarm unit: supports alarm mechanisms such as SMS, WeChat, and push notifications.

Images