Power optimization control method, module and system based on non-intrusive load identification

Abstract

The invention discloses a power optimization control method, module and system based on non-intrusive load identification, and relates to the technical field of electrical monitoring, and the method comprises the steps: obtaining the electrical data of an electrical appliance load in a target circuit; preprocessing the electrical data of the target circuit; extracting multi-dimensional features from the preprocessed electrical data of the target circuit; inputting the multi-dimensional features of the target circuit into a pre-trained load identification model to identify an electric appliance load; and according to the identified multi-dimensional characteristics of the electrical appliance load, independent power optimization control is carried out on the electrical appliance load. According to the invention, accurate energy consumption identification and control of each electric appliance load in a superstore and a house can be realized, and energy waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of electrical monitoring, and in particular, to a power optimization control method, module and system based on non-intrusive load identification. Background Art

[0002] An energy consumption monitoring system is a monitoring system that uses advanced computer control technology, management software and energy-saving system programs to make the equipment (lighting, air conditioning, refrigerator, microwave oven) in a building's mechanical and electrical systems or a building complex operate in an orderly, coordinated and scientific manner, so as to effectively ensure a comfortable working environment in the building, achieve energy conservation, reduce the workload of maintenance management and operating costs.

[0003] Existing energy consumption monitoring systems are mainly based on overall load management, and can only provide total energy consumption data, lacking the ability to accurately identify and optimize control of sub-loads, and thus unable to identify and distinguish the energy consumption of each electrical device; therefore, it is impossible to optimize and control the energy consumption of a single device, resulting in energy waste and reduced equipment operation efficiency.

[0004] In addition, traditional energy consumption monitoring systems cannot monitor and predict the operating status of equipment in real time, and cannot detect and prevent equipment failures in a timely manner, thus affecting the safety and stability of the equipment. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a power optimization control method, module and system based on non-intrusive load identification, which can achieve accurate energy consumption identification and control of electrical loads and reduce energy waste.

[0006] To solve the above technical problem, the present invention provides a power optimization control method based on non-intrusive load identification, including: acquiring electrical data of electrical loads in a target circuit, where the electrical data includes voltage information and current information; preprocessing the electrical data of the target circuit; extracting multi-dimensional features from the preprocessed electrical data of the target circuit, where the multi-dimensional features include active power features, reactive power features and harmonic component features; inputting the multi-dimensional features of the target circuit into a pre-trained load identification model to identify electrical loads; and respectively performing independent power optimization control on the electrical loads according to the identified multi-dimensional features of the electrical loads.

[0007] As an improvement of the above solution, the training steps of the load recognition model include: obtaining the electrical data of electrical loads in the reference circuit; preprocessing the electrical data of the reference circuit; extracting multi-dimensional features from the preprocessed electrical data of the reference circuit; performing labeling processing on the multi-dimensional features of the reference circuit according to the load names of the electrical loads in the reference circuit to form labeled feature data; inputting the labeled feature data into the load recognition model to train the load recognition model.

[0008] As an improvement of the above solution, the step of extracting the active power feature from the preprocessed electrical data includes: calculating the active power according to the formula P = UI cosθ, where P is the active power, U is the voltage information, I is the current information, and cosθ is the power factor; extracting the maximum value, minimum value, mean value, and root mean square of the active power to form the active power feature.

[0009] As an improvement of the above solution, the step of extracting the reactive power feature from the preprocessed electrical data includes: calculating the reactive power according to the formula Q = UI sinθ, where Q is the reactive power, U is the voltage information, I is the current information, and θ is the phase angle between the current information and the voltage information; extracting the maximum value, minimum value, mean value, and root mean square of the reactive power to form the reactive power feature.

[0010] As an improvement of the above solution, the step of extracting the harmonic component feature from the preprocessed electrical data includes: according to the formula i(t) = i a (t) + i b (t), decomposing the preprocessed current information into non-active current and active current, where i(t) is the current information, i a (t) is the non-active current, and i b (t) is the active current; performing Fourier transform on the non-active current to generate the harmonic components of the non-active current; performing dimensionality reduction processing on the harmonic components to extract the main harmonic information; selecting the harmonic amplitude with the largest difference from the main harmonic information as the harmonic component feature.

[0011] As an improvement of the above solution, the power optimization control method based on non-intrusive load recognition further includes: obtaining the status data of the electrical loads in the target circuit; judging whether the electrical loads are abnormal according to the multi-dimensional features and status data of the identified electrical loads; when it is judged that the electrical loads are abnormal, generating and sending an abnormal information to the host computer.

[0012] As an improvement of the above solution, judging whether the electrical loads are abnormal according to the multi-dimensional features and status data of the identified electrical loads through a preset safety threshold and / or a status prediction model.

[0013] Accordingly, the present invention further provides a power optimization control module based on non-intrusive load identification, which includes a memory and a processor. The memory stores a computer program. Wherein, when the processor executes the computer program, the steps of the above-mentioned power optimization control method based on non-intrusive load identification are realized.

[0014] Accordingly, the present invention further provides a power optimization control system based on non-intrusive load identification, which includes a current acquisition device, a voltage acquisition device, and the above-mentioned power optimization control module based on non-intrusive load identification. The power optimization control module based on non-intrusive load identification is respectively connected to the current acquisition device and the voltage acquisition device; the current acquisition device is used to collect and send the current signal of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification; the voltage acquisition device is used to collect and send the voltage signal of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification.

[0015] As an improvement of the above solution, the power optimization control system based on non-intrusive load identification further includes a status acquisition device and a host computer connected to the power optimization control module based on non-intrusive load identification; the status acquisition device is used to collect and send the status data of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification.

[0016] Implementing the present invention has the following beneficial effects:

[0017] The present invention adopts a multi-feature extraction technology. According to the steady-state current data and steady-state voltage data after median filtering, the active power feature, reactive power feature, and harmonic component features of each order of non-active current are extracted;

[0018] The present invention adopts a refined load identification algorithm. According to the extracted multi-feature data, a load identification model of machine learning is trained to realize the accurate identification of a single electrical load;

[0019] The present invention adopts a power optimization control strategy. According to the load identification result, a power optimization control strategy is developed to reduce energy consumption and improve the operation efficiency of the equipment;

[0020] Furthermore, the present invention adopts a safety and stability evaluation, real-time reporting and early warning mechanism to monitor the operation status of the equipment in real time. Through the safety and stability evaluation, equipment failures are prevented and reduced to ensure the continuous and stable operation of the system; and a real-time reporting mechanism is established. When the system identifies an abnormality or potential risk, the user can be notified in time to take control measures. Description of the Drawings

[0021] Figure 1It is the flowchart of the first embodiment of the power optimization control method based on non-intrusive load identification of the present invention;

[0022] Figure 2 It is the training flowchart of the load identification model in the present invention;

[0023] Figure 3 It is the flowchart of the second embodiment of the power optimization control method based on non-intrusive load identification of the present invention;

[0024] Figure 4 It is the schematic structural diagram of the first embodiment of the power optimization control system based on non-intrusive load identification of the present invention;

[0025] Figure 5 It is the schematic structural diagram of the second embodiment of the power optimization control system based on non-intrusive load identification of the present invention. Detailed implementation manners

[0026] To make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0027] See Figure 1 , Figure 1 shows the flowchart of the first embodiment of the power optimization control method based on non-intrusive load identification of the present invention, which includes:

[0028] S101, acquiring the electrical data of the electrical loads in the target circuit;

[0029] In practical applications, devices such as current transformers, voltage sensors, and smart meters installed in the target circuit can be used to collect the electrical data of each electrical load (such as common electrical loads like lighting, air conditioners, refrigerators, and microwave ovens) in the target circuit (power distribution system); the electrical data includes voltage information (such as steady-state voltage data) and current information (such as steady-state current data).

[0030] S102, preprocessing the electrical data of the target circuit;

[0031] The preprocessing includes algorithms such as median filtering and wavelet transform, which can effectively reduce the noise and interference in the original electrical data to improve the data quality.

[0032] S103, extracting multi-dimensional features from the preprocessed electrical data of the target circuit;

[0033] Since different electrical loads have different impacts on the circuit, different electrical loads have unique active power characteristics, reactive power characteristics, and harmonic component characteristics of non-active current. In the present invention, the multi-dimensional features include active power characteristics, reactive power characteristics, and harmonic component characteristics of non-active current.

[0034] The extraction methods of active power characteristics, reactive power characteristics, and harmonic component characteristics will be described separately as follows:

[0035] I. Active power characteristics

[0036] The present invention mainly aims at load identification of common electrical appliances such as lighting, air conditioners, refrigerators, and microwave ovens, and uses the calculation method of single-phase AC circuits when calculating power.

[0037] Correspondingly, the steps for extracting active power characteristics from the preprocessed electrical data include:

[0038] (1) Calculate the active power according to the following formula:

[0039] P = UI cosθ

[0040] Where,

[0041] P is the active power;

[0042] U is the voltage information;

[0043] I is the current information;

[0044] cosθ is the power factor.

[0045] (2) Extract the maximum value, minimum value, mean value, and root mean square of the active power to form the active power characteristics.

[0046] II. Reactive power characteristics

[0047] The reactive power characteristics are helpful for describing the variation of reactive power in different time periods.

[0048] Correspondingly, the steps for extracting reactive power characteristics from the preprocessed electrical data include:

[0049] (1) Calculate the reactive power according to the following formula:

[0050] Q = UI sinθ

[0051] Where,

[0052] Q is the reactive power;

[0053] U is the voltage information;

[0054] I is the current information;

[0055] θ is the phase angle between the current information and the voltage information.

[0056] (2) Extract the maximum value, minimum value, mean value, and root mean square of the reactive power to form the reactive power characteristics.

[0057] III. Harmonic Component Characteristics

[0058] Correspondingly, the steps for extracting the harmonic component characteristics of the non-active current from the preprocessed electrical data include:

[0059] (1) Decompose the preprocessed current information into non-active current and active current according to the following formula:

[0060] i(t) = i a (t) + i b (t)

[0061] where:

[0062] i(t) is the current information;

[0063] i a (t) is the non-active current;

[0064] i b (t) is the active current.

[0065] (2) Perform Fourier transform on the non-active current to generate the harmonic components of the non-active current;

[0066] (3) Perform dimensionality reduction on the harmonic components to extract the main harmonic information;

[0067] Furthermore, Kernel Principal Component Analysis (KPCA) can be used to perform dimensionality reduction on the harmonic component characteristics of each order of the non-active current to extract the main harmonic information.

[0068] (4) Select the harmonic amplitude with the largest difference from the main harmonic information as the harmonic component characteristic.

[0069] Select the harmonic amplitude with the largest difference as the maximum difference harmonic characteristic, that is, the harmonic component characteristic.

[0070] Therefore, the present invention can adopt multi-feature extraction technology to extract the active power characteristic, reactive power characteristic, and harmonic component characteristics of each order of the non-active current from the steady-state current data and steady-state voltage data.

[0071] S104. Input the multi-dimensional characteristics of the target circuit into a pre-trained load recognition model to identify the electrical load;

[0072] It should be noted that the load recognition model can be a machine learning model (such as SVM, XGBoost classification model, fuzzy clustering model, random forest, neural network);

[0073] The present invention uses the trained machine learning model to learn and classify the extracted multi-dimensional characteristics to identify and distinguish different electrical loads.

[0074] S105, respectively perform independent power optimization control on the electrical loads according to the identified multi-dimensional characteristics of the electrical loads.

[0075] According to the load identification result, start the preset power optimization control strategy (such as demand response, load balancing, etc.) to optimize the control of different electrical loads, adjust the operating parameters of the electrical equipment, so as to reduce energy consumption and improve the operating efficiency of the equipment and the energy utilization efficiency.

[0076] Specifically, the optimization control strategy includes adjusting the standby power supply according to the working state of the electrical load, dynamically adjusting the output of current and voltage according to the actual power consumption demand, etc.

[0077] Therefore, through the power optimization control method based on non-intrusive load identification of the present invention, it is possible to combine non-intrusive load identification technology and machine learning algorithms to achieve accurate energy consumption identification and control of each electrical equipment in large shopping malls and residences, reduce energy waste; at the same time, through precise control of the energy consumption of each electrical equipment, energy conservation and emission reduction are achieved, so as to optimize the overall power consumption efficiency, improve the power consumption efficiency, reduce energy waste, and improve the safety and stability of the equipment.

[0078] See Figure 2 , Figure 2 shows the training flow chart of the load identification model in the present invention, which includes:

[0079] S201, obtain the electrical data of the electrical loads in the reference circuit;

[0080] S202, preprocess the electrical data of the reference circuit;

[0081] S203, extract multi-dimensional characteristics from the preprocessed electrical data of the reference circuit;

[0082] S204, perform tagging processing on the multi-dimensional characteristics of the reference circuit according to the load names of the electrical loads in the reference circuit to form tagged feature data;

[0083] Specifically, according to the load names of the electrical loads, tags can be made for the extracted active power characteristics, reactive power characteristics, and harmonic components of each order of the non-active current, and saved in files in the.npy and.json formats.

[0084] By performing tagging processing on the multi-dimensional characteristics of the reference circuit, the corresponding relationship between the electrical loads and the multi-dimensional characteristics can be clearly shown.

[0085] S205, input the tagged feature data into the load identification model to train the load identification model.

[0086] The present invention divides the tagged feature data according to a certain ratio, and then inputs it into a load recognition model (such as an SVM, XGBoost classification model, fuzzy clustering model) for training and optimization.

[0087] Taking the Support Vector Machine (SVM) model as an example, its training and optimization process includes feature dataset division, SVM model initialization, load recognition model training, model evaluation, and parameter optimization:

[0088] I. Feature dataset division process

[0089] The tagged feature data is divided into a training dataset and a test dataset according to a ratio of 8:2 or 7:3.

[0090] II. SVM model initialization process

[0091] Construct an SVM model, set the penalty coefficient C, set the memory size required for training, the number of iterations, and select a suitable kernel function, such as the Gaussian kernel (RBF) or polynomial kernel function (Poly).

[0092] III. Load recognition model training process

[0093] Use the training set data to train the SVM model.

[0094] During the training process, find the optimal hyperplane by solving the optimization problem, which can maximize the separation of data points of different classes.

[0095] IV. Model evaluation and parameter optimization process

[0096] Use the test set data to evaluate the performance of the SVM model.

[0097] Calculate the statistical number of accurately recognized samples and the total number of recognized samples to determine the recognition accuracy rate of the model, and use the five-fold cross-validation method to evaluate the stability of the model.

[0098] According to the results of model evaluation, adjust the SVM parameters.

[0099] Adjust the penalty coefficient, kernel function coefficient, etc. through methods such as cross-validation and grid search to improve the recognition accuracy rate of the model.

[0100] After determining the optimal SVM parameters, retrain the model using all the training data to obtain the final load recognition model.

[0101] See Figure 3 , Figure 3 which shows the flowchart of the second embodiment of the power optimization control method based on non-intrusive load recognition of the present invention, and it includes:

[0102] S301, Obtain the electrical data of the electrical loads in the target circuit;

[0103] S302, Preprocess the electrical data of the target circuit;

[0104] S303, Extract multi-dimensional features from the preprocessed electrical data of the target circuit;

[0105] S304, Input the multi-dimensional features of the target circuit into a pre-trained load identification model to identify the electrical loads;

[0106] S305, According to the multi-dimensional features of the identified electrical loads, perform independent power optimization control on the electrical loads respectively.

[0107] S306, Obtain the status data of the electrical loads in the target circuit;

[0108] Use sensors and smart meters to collect key status data such as air pressure, temperature, and humidity of the electrical loads in real time.

[0109] S307, Judge whether the electrical loads are abnormal according to the multi-dimensional features and status data of the identified electrical loads;

[0110] It is possible to judge whether the electrical loads are abnormal according to the multi-dimensional features and status data of the identified electrical loads through a preset safety threshold and / or a status prediction model, so as to evaluate and classify potential risks.

[0111] Specifically:

[0112] I. Safety threshold model

[0113] Safety thresholds can be preset for the status parameters of some monitored electrical loads. When the monitored status data exceeds the safety threshold, it will be identified as a potential risk.

[0114] II. Status prediction model

[0115] A status prediction model based on machine learning algorithms can be pre-deployed for some monitored electrical loads to identify abnormal patterns or trends.

[0116] S308, When it is judged that the electrical loads are abnormal, generate and send abnormal information to the host computer.

[0117] When abnormal or risky situations are judged, trigger the alarm mechanism; among them, some risks will report the abnormal information (alarm information and risk assessment results) to the host computer (such as a central monitoring system or a cloud platform) in real time through a wireless communication module (such as 4G, Wi-Fi, LoRa, etc.); some other risks will notify relevant personnel through audible and visual alarms, text messages, emails or APP push notifications to achieve the remote notification function.

[0118] Furthermore, all abnormal events and alarm information can also be recorded, including time, device ID, abnormal parameters, risk level, etc.

[0119] Therefore, different from Figure 1 the first embodiment shown, in this embodiment, by real-time monitoring of the status data of the electrical load, the safety and stability of the electrical load are evaluated, so as to give early warnings of potential risks, and effectively prevent and reduce equipment failures.

[0120] Correspondingly, the present invention also discloses a power optimization control module based on non-intrusive load identification, including a memory and a processor, the memory stores a computer program, wherein when the processor executes the computer program, the steps of the above-mentioned power optimization control method based on non-intrusive load identification are implemented.

[0121] Referring to Figure 4 , Figure 4 shows a first embodiment of the power optimization control system 100 based on non-intrusive load identification of the present invention, which includes a current acquisition device 1, a voltage acquisition device 2 and a power optimization control module 3 based on non-intrusive load identification. The power optimization control module 3 based on non-intrusive load identification is respectively connected to the current acquisition device 1 and the voltage acquisition device 2. Specifically:

[0122] The current acquisition device 1 is used to collect and send the current signal of the electrical load in the circuit to the power optimization control module 3 based on non-intrusive load identification; among them, the current acquisition device 1 is preferably a current transformer, but not limited thereto, and can be selected according to the actual situation;

[0123] The voltage acquisition device 2 is used to collect and send the voltage signal of the electrical load in the circuit to the power optimization control module 3 based on non-intrusive load identification; among them, the voltage acquisition device 2 is preferably a voltage sensor, but not limited thereto, and can be selected according to the actual situation;

[0124] The power optimization control module 3 based on non-intrusive load identification includes a memory and a processor, the memory stores a computer program, wherein when the processor executes the computer program, the steps of the above-mentioned power optimization control method based on non-intrusive load identification are implemented.

[0125] Generally, the power optimization control module 3 based on non-intrusive load identification is installed inside the distribution box, the electrical load is installed outside the distribution box, and the electrical load is connected to the power optimization control module 3 based on non-intrusive load identification through a wire.

[0126] Referring to Figure 5 , Figure 5Shows the second embodiment of the power optimization control system 100 based on non-intrusive load identification according to the present invention. Different from the Figure 4 second embodiment shown, in this embodiment, the power optimization control system 100 based on non-intrusive load identification further includes a state acquisition device 4 and a host computer 5 connected to the power optimization control module 3 based on non-intrusive load identification.

[0127] The state acquisition device 4 is used to collect and send the state data of the electrical load in the circuit to the power optimization control module 3 based on non-intrusive load identification. Among them, the state acquisition device 4 is preferably a sensor, an intelligent meter, etc., but is not limited thereto, as long as it can collect key state data such as the air pressure, temperature, and humidity of the electrical load in real time.

[0128] When the power optimization control module 3 based on non-intrusive load identification determines that the electrical load is abnormal according to the multi-dimensional characteristics and state data of the identified electrical load, it generates and sends an abnormal message to the host computer 5.

[0129] In summary, through the present invention, the power consumption efficiency of large shopping malls and residences can be effectively improved, the accurate identification and optimization control of sub-divided loads are realized, the refinement level of energy management is improved, and energy waste is reduced; at the same time, the safety and stability of the equipment are enhanced, and equipment failures are reduced. Specifically:

[0130] The present invention adopts a multi-feature extraction technology to extract the active power feature, reactive power feature, and each harmonic component feature of the non-active current according to the steady-state current data and steady-state voltage data after median filtering.

[0131] The present invention adopts a refined load identification algorithm to train a load identification model of machine learning according to the extracted multi-feature data to achieve accurate identification of a single electrical load.

[0132] The present invention adopts a power optimization control strategy to develop a power optimization control strategy according to the load identification result to reduce energy consumption and improve the operation efficiency of the equipment.

[0133] The present invention adopts a safety and stability evaluation, real-time reporting, and early warning mechanism to monitor the operation state of the equipment in real time, prevent and reduce equipment failures through safety and stability evaluation, and ensure the continuous and stable operation of the system; and establish a real-time reporting mechanism, when the system identifies an abnormality or potential risk, it can timely notify the user to take control measures.

[0134] The above is the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements are also regarded as the protection scope of the present invention.

Claims

1. A power optimization control method based on non-intrusive load identification, characterized in that Including: Obtain electrical data of electrical loads in the target circuit, where the electrical data includes voltage information and current information; Preprocess the electrical data of the target circuit; Extract multi-dimensional features from the preprocessed electrical data of the target circuit, where the multi-dimensional features include active power features, reactive power features, and harmonic component features; Input the multi-dimensional features of the target circuit into a pre-trained load identification model to identify electrical loads; According to the multi-dimensional features of the identified electrical loads, perform independent power optimization control on the electrical loads respectively.

2. The power optimization control method based on non-intrusive load identification according to claim 1, characterized in that The training steps of the load identification model include: Obtain electrical data of electrical loads in the reference circuit; Preprocess the electrical data of the reference circuit; Extract multi-dimensional features from the preprocessed electrical data of the reference circuit; Perform labeling processing on the multi-dimensional features of the reference circuit according to the load names of the electrical loads in the reference circuit to form labeled feature data; Input the labeled feature data into the load identification model to train the load identification model.

3. The power optimization control method based on non-intrusive load identification according to claim 1 or 2, characterized in that, The steps of extracting active power features from the preprocessed electrical data include: Calculate the active power according to the formula P = UI cosθ, where P is the active power, U is the voltage information, I is the current information, and cosθ is the power factor; Extract the maximum value, minimum value, mean value, and root mean square of the active power to form active power features.

4. The power optimization control method based on non-intrusive load identification according to claim 1 or 2, characterized in that, The steps of extracting reactive power features from the preprocessed electrical data include: Calculate the reactive power according to the formula Q = UI sinθ, where Q is the reactive power, U is the voltage information, I is the current information, and θ is the phase angle between the current information and the voltage information; Extract the maximum value, minimum value, mean value, and root mean square of the reactive power to form reactive power features.

5. The power optimization control method based on non-intrusive load identification according to claim 1 or 2, characterized in that, The steps of extracting harmonic component features from the preprocessed electrical data include: According to the formula i(t) = i a (t) + i b (t), the preprocessed current information is decomposed into inactive current and active current, where i(t) is the current information, i a (t) is the inactive current, and i b (t) is the active current; Perform Fourier transform on the non-active current to generate harmonic components of the non-active current; Perform dimensionality reduction processing on the harmonic components to extract main harmonic information; Select the harmonic amplitude with the largest difference from the main harmonic information as the harmonic component feature.

6. The power optimization control method based on non-intrusive load identification according to claim 1, characterized in that It also includes: Obtain the status data of electrical loads in the target circuit; Judge whether the electrical load is abnormal according to the multi-dimensional features and status data of the identified electrical load; When it is judged that the electrical load is abnormal, generate and send an abnormal information to the host computer.

7. The power optimization control method based on non-intrusive load identification according to claim 6, wherein Judge whether the electrical load is abnormal according to the multi-dimensional features and status data of the identified electrical load through a preset safety threshold and / or a status prediction model.

8. A power optimization control module based on non-intrusive load identification, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the power optimization control method based on non-intrusive load identification according to any one of claims 1 to 7.

9. A power optimization control system based on non-intrusive load identification, characterized in that, Including a current acquisition device, a voltage acquisition device, and the power optimization control module based on non-intrusive load identification according to claim 8, where the power optimization control module based on non-intrusive load identification is respectively connected to the current acquisition device and the voltage acquisition device; The current acquisition device is used to acquire and send the current signal of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification; The voltage acquisition device is used to acquire and send the voltage signal of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification.

10. The power optimization control system based on non-intrusive load identification according to claim 9, characterized in that It further includes a status acquisition device and a host computer connected to the power optimization control module based on non-intrusive load identification; The status acquisition device is used to acquire and send the status data of the electrical load in the circuit to the power optimization control module based on non-intrusive load identification.

Images