Non-intrusive load identification method and device
By combining a voltage sampling loop with phase difference calculation and data restoration methods, load identification for multiple users is achieved, solving the problems of high cost, large size and complex installation in existing technologies, and improving the ease of installation and identification accuracy of the equipment.
Patent Information
- Application Number
- CN202511973492.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-25
- Publication Date
- 2026-02-17
AI Technical Summary
Existing non-intrusive load identification devices require multiple current and voltage sensors, resulting in high cost, large size, and complex installation, making them difficult to apply in scenarios such as apartment buildings.
A single voltage sampling loop is used to obtain the standard frequency voltage and current sampling data of each user through phase difference calculation. The data is restored and corrected using cubic spline interpolation and phase-locked loop methods to achieve load identification of multiple users.
It reduces the hardware cost and size of the load identification device, simplifies the installation process, and improves identification accuracy and system fault tolerance.
Smart Images
Figure CN121540972A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a non-intrusive load monitoring (NILM) method, belonging to the field of power management technology. Background Technology
[0002] Non-intrusive load identification (NILM) is an advanced power management technology that can identify and monitor the energy consumption of individual electrical devices by analyzing power data on a bus without directly connecting to the circuits of each device. This technology can be widely applied in smart grids, energy management, home automation, and other fields, providing strong support for energy conservation, emission reduction, and optimized electricity use.
[0003] The core of NILM lies in the high-precision acquisition and real-time processing of signals such as current and voltage in user circuits. Then, through algorithmic models, these signals are analyzed to extract characteristic information of different electrical devices, such as power, startup characteristics, and operating cycles, thereby enabling the identification and energy consumption metering of each device. Compared to traditional interventional monitoring methods, NILM eliminates the need to install sensors or meters on every electrical device, significantly reducing costs and implementation complexity.
[0004] Existing load identification devices primarily identify charging loads non-intrusively by collecting voltage and current information at the user's inlet terminal. These devices are typically installed in the user's incoming electrical box. However, electrical boxes were not initially designed with load identification devices in mind, resulting in very limited space during design and selection. For example, in an apartment building with dozens of households, installing a load identification device for each household would be too expensive and insufficient to accommodate all units in the existing meter box. Using a single load identification device installed in the unit's meter box to identify the loads of all users in that unit can alleviate this problem to some extent. However, even with this approach, obtaining accurate current and voltage information for each user requires configuring a current sensor, a voltage sensor, and corresponding signal sampling circuits for each user's circuit to obtain voltage and current synchronization signals. This results in an excessive number of components, leading to high costs and bulky devices. Furthermore, installation requires ensuring that the current and voltage sensors for the same user are installed in the same circuit, making the process overly complex and prone to errors. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a non-intrusive load identification method. Only one voltage sampling circuit is needed to identify the load of multiple users at the same time. The implementation cost and installation volume of the equipment are greatly reduced. Moreover, the current sensors of each user can be arbitrarily connected to the power input line without the need for manual matching of voltage and current circuits, making construction and installation simpler.
[0006] The present invention specifically adopts the following technical solutions to solve the above-mentioned technical problems: A non-intrusive load identification method identifies the load of each user based on standard frequency voltage sampling data and standard frequency current sampling data of each user connected to the same main incoming line terminal. The standard frequency voltage sampling data and standard frequency current sampling data of each user are obtained using the following method: S1. At the same sampling frequency, the voltage signal of any phase of the main incoming line and the current signal of any inlet circuit of each user are sampled synchronously for at least one complete voltage cycle to obtain the voltage sampling data UL of the main incoming line and the current sampling data Ii of each user circuit, where i is the sequence number of any user circuit connected to the main incoming line. S2. Restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency. S3. Obtain the phase difference Δφi between the in-phase in-circuit voltage signal and the sampled total incoming line voltage signal of each user's sampled in-circuit current signal using the following method: S301. Extract the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit, and normalize it to -180°~180°. S302. For each user circuit, calculate the three inherent phase differences between the sampled in-circuit current signal and the sampled total incoming line voltage signal when they are in phase and when they are out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, △θ_native_C_i = △θ_i-120°; S303. For each user circuit, select one of the three inherent phase differences that falls within the preset inherent phase difference range, and subtract Δθ_i from it to obtain the phase difference Δφi between the in-phase in-circuit voltage signal of the user's sampled in-circuit current signal and the sampled total incoming line voltage signal. S4. For each user circuit, the standard frequency voltage sampling data UsL is corrected according to the corresponding phase difference Δφi, thus obtaining the standard frequency voltage sampling data of each user circuit.
[0007] Preferably, cubic spline interpolation is used to restore the voltage sampling data UL and current sampling data Ii to standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency.
[0008] In one embodiment, the inherent phase difference ranges from -70° to 50°.
[0009] Preferably, the phase difference between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit is extracted using the following method: the fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi are calculated using the FFT method, and the phase of the fundamental components UsL_0 and Isi_0 is obtained using the phase-locked loop method, and the difference between the two is calculated, thus obtaining the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit.
[0010] Preferably, correcting the standard frequency voltage sampling data UsL according to the corresponding phase difference Δφi means: applying a phase delay Δφi to the standard frequency voltage sampling data UsL to obtain a new standard frequency voltage sampling data after phase delay.
[0011] Based on the same inventive concept, the following technical solutions can also be obtained: A non-intrusive load identification device includes a load identification unit for non-intrusive load identification of each user based on standard frequency voltage sampling data and standard frequency current sampling data of each user connected to the same main incoming line terminal; it also includes a data acquisition unit for acquiring the standard frequency voltage sampling data and standard frequency current sampling data of each user through the following methods: S1. At the same sampling frequency, the voltage signal of any phase of the main incoming line and the current signal of any inlet circuit of each user are sampled synchronously for at least one complete voltage cycle to obtain the voltage sampling data UL of the main incoming line and the current sampling data Ii of each user circuit, where i is the sequence number of any user circuit connected to the main incoming line. S2. Restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency. S3. Obtain the phase difference Δφi between the in-phase in-circuit voltage signal and the sampled total incoming line voltage signal of each user's sampled in-circuit current signal using the following method: S301. Extract the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit, and normalize it to -180°~180°. S302. For each user circuit, calculate the three inherent phase differences between the sampled in-circuit current signal and the sampled total incoming line voltage signal when they are in phase and when they are out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, △θ_native_C_i = △θ_i-120°; S303. For each user circuit, select one of the three inherent phase differences that falls within the preset inherent phase difference range, and subtract Δθ_i from it to obtain the phase difference Δφi between the in-phase in-circuit voltage signal of the user's sampled in-circuit current signal and the sampled total incoming line voltage signal. S4. For each user circuit, the standard frequency voltage sampling data UsL is corrected according to the corresponding phase difference Δφi, thus obtaining the standard frequency voltage sampling data of each user circuit.
[0012] Preferably, cubic spline interpolation is used to restore the voltage sampling data UL and current sampling data Ii to standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency.
[0013] In one embodiment, the inherent phase difference ranges from -70° to 50°.
[0014] Preferably, the phase difference between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit is extracted using the following method: the fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi are calculated using the FFT method, and the phase of the fundamental components UsL_0 and Isi_0 is obtained using the phase-locked loop method, and the difference between the two is calculated, thus obtaining the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit.
[0015] Preferably, correcting the standard frequency voltage sampling data UsL according to the corresponding phase difference Δφi means: applying a phase delay Δφi to the standard frequency voltage sampling data UsL to obtain a new standard frequency voltage sampling data after phase delay.
[0016] Compared with the prior art, the technical solution of the present invention has the following beneficial effects: This invention requires only one voltage sampling circuit to accurately extract the voltage and current information needed for identifying multiple user loads, effectively reducing the cost of the hardware circuit of the load identification device, reducing the size of the load identification device, and facilitating practical applications. In this invention, the power input line and current sensor on the user side can be connected arbitrarily, eliminating the need for manual matching of voltage and current loops as in traditional applications. This reduces the difficulty of installation and construction, and improves the fault tolerance of the system. This invention uses standard frequency voltage and current sampling data to reduce the impact of frequency drift in actual power grids and improve identification accuracy. Attached Figure Description
[0017] Figure 1 This is a schematic diagram of a specific application example of the non-intrusive load identification device of the present invention. Detailed Implementation
[0018] To address the shortcomings of existing technologies, the present invention integrates multi-channel identification functions into a single identification device, utilizing a single voltage sampling loop to acquire voltage data from multiple users, thereby reducing the cost of the hardware circuitry of the load identification device and decreasing its size.
[0019] The present invention specifically adopts the following technical solutions to solve the above-mentioned technical problems: A non-intrusive load identification method identifies the load of each user based on standard frequency voltage sampling data and standard frequency current sampling data of each user connected to the same main incoming line terminal. The standard frequency voltage sampling data and standard frequency current sampling data of each user are obtained using the following method: S1. At the same sampling frequency, the voltage signal of any phase of the main incoming line and the current signal of any inlet circuit of each user are sampled synchronously for at least one complete voltage cycle to obtain the voltage sampling data UL of the main incoming line and the current sampling data Ii of each user circuit, where i is the sequence number of any user circuit connected to the main incoming line. S2. Restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency. S3. Obtain the phase difference Δφi between the in-phase in-circuit voltage signal and the sampled total incoming line voltage signal of each user's sampled in-circuit current signal using the following method: S301. Extract the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit, and normalize it to -180°~180°. S302. For each user circuit, calculate the three inherent phase differences between the sampled in-circuit current signal and the sampled total incoming line voltage signal when they are in phase and when they are out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, △θ_native_C_i = △θ_i-120°; S303. For each user circuit, select one of the three inherent phase differences that falls within the preset inherent phase difference range, and subtract Δθ_i from it to obtain the phase difference Δφi between the in-phase in-circuit voltage signal of the user's sampled in-circuit current signal and the sampled total incoming line voltage signal. S4. For each user circuit, the standard frequency voltage sampling data UsL is corrected according to the corresponding phase difference Δφi, thus obtaining the standard frequency voltage sampling data of each user circuit.
[0020] The correction of the standard frequency voltage sampling data UsL based on the corresponding phase difference Δφi means applying a phase delay Δφi to the standard frequency voltage sampling data UsL to obtain a new standard frequency voltage sampling data with a phase delay.
[0021] Load identification models are typically established under the standard frequency of the power grid (50Hz or 60Hz). However, the actual voltage and current frequencies often fluctuate and deviate from the labeled frequency. In order to better match the load identification model and reduce identification errors, this invention needs to transform the voltage and current sampling data from the actual frequency to the standard frequency. This transformation process can be implemented using various existing technologies. This invention preferably uses cubic spline interpolation to restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi.
[0022] The phase difference between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit can be obtained using various existing technologies; the present invention preferably uses the following method to extract it: the fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi are calculated using the FFT method, and the phase of the fundamental components UsL_0 and Isi_0 is obtained using the phase-locked loop method and the difference between the two is calculated, thus obtaining the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit.
[0023] The inherent phase difference range is related to the distribution characteristics of electrical appliances in different types of power consumption scenarios. Taking residential power consumption scenarios as an example, the electrical appliances used in residential power consumption are usually mainly inductive and resistive loads, with almost no purely capacitive loads. Capacitive loads are essentially electronic devices with a "diode rectification + capacitor filtering" topology. The capacitive performance is more obvious under light loads and weakens under heavy loads. Therefore, the inherent phase difference range can be obtained in advance by statistically analyzing the distribution characteristics of electrical appliances in the same type of power consumption scenario, or by statistically analyzing the actual phase difference data between the sampled inlet circuit current signal and the sampled main inlet voltage signal when they are in phase in the same type of power consumption scenario. In residential power consumption scenarios, the inherent phase difference range is [-70°, 50°].
[0024] To facilitate public understanding, the technical solution of the present invention will be described in detail below through a specific embodiment and in conjunction with the accompanying drawings: A specific application example of the non-intrusive load identification device of the present invention is as follows: Figure 1 As shown, in this example, users 1 through 4 are connected to the same main incoming line terminal. Figure 1 As shown, the load identification device uses a voltage sensor to draw power from phase A of the main incoming line to obtain the voltage signal UL. At the same time, it uses current sensors to measure the current of the six incoming lines of users 1 to 4, with corresponding current signals I1, I2, I3, I4, I5 and I6, respectively. Among them, user 3 is connected to phase C, user 2 is connected to phase B, user 1 is connected to phase A, and user 4 is a three-phase incoming line, with phases A, B and C simultaneously inside the house.
[0025] The non-intrusive load identification device of the present invention includes a data acquisition unit and a load identification unit; the data acquisition unit is used to acquire standard frequency voltage sampling data and standard frequency current sampling data of each user; the load identification unit is used to perform non-intrusive load identification for each user based on the standard frequency voltage sampling data and standard frequency current sampling data of each user.
[0026] The load identification unit needs to acquire the voltage and current synchronization signals of each user to identify the load for each user. Taking user 1 as an example, since the voltage is connected to phase A and the user's load is connected to phase C, directly using UL and I1 as synchronization signals cannot accurately identify the user's load. Therefore, it is necessary to use a data acquisition unit to acquire the standard frequency voltage sampling data and standard frequency current sampling data of each user based on the voltage signal UL and the current signals I1 to I6. In this embodiment, the working process of the data acquisition unit specifically includes the following steps: S1. At the same sampling frequency, the voltage signal of any phase of the main incoming line and the current signal of any inlet circuit of each user are sampled synchronously for at least one complete voltage cycle to obtain the voltage sampling data UL of the main incoming line and the current sampling data Ii of each user circuit, where i is the sequence number of any user circuit connected to the main incoming line. In this embodiment, data of length 20ms are sampled for UL and Ii (i=1,2,3,4) at a frequency of fs = 10kHz to obtain discrete data sequences: UL(n) = { UL(1), UL(2), ..., UL(n)}, Ii(n) = { Ii (1), Ii(2), ..., Ii (n)}, n=1,2, ...,200.
[0027] S2. Restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency. In this embodiment, cubic spline interpolation is used to restore the actual sampled data to standard frequency sampled data at the standard frequency. Taking the restoration of data UL(n) to standard frequency voltage UsL(n) data as an example, the specific process is as follows: ① Assume the actual voltage UL frequency is f 1 = 51Hz, the actual voltage satisfies: u i =311sin(2π f 1 t i The first 10 sampling points are shown in Table 1: Table 1
[0028] ② Define the target time axis at 50Hz: t k ' = k × T0 / (n - 1), k = 0,1,2,..., n-1, n = f s / f 0 = 200, so we can calculate: t k = 0.0001005025 k ( k = 0, 1, 2, ..., 199); the theoretical 50Hz values corresponding to the first 10 target points are shown in Table 2: Table 2
[0029] ③ The actual sampled values are restored to the standard frequency values using cubic spline interpolation, with the target point k=1 ( t Taking 1'=0.1005025ms as an example, the specific steps are as follows: Step 1: Solve for the second derivative Mn (natural spline boundary: M0=0, M199=0) Three bending moment equations under uniform sampling: M n +4M n+1 + M n+2 = 6 / t s 2 ( x n+2 – 2 x n+1 + x n ) Let n=0, and substitute it into x 0=0, x 1 = 9.9520 x 2 = 19.9009, so the equation is: 4M1 + M2 = -186000; Let n=1, and substitute it into... x 1 = 9.9520 x 2 = 19.9009, x 3 = 29.8405, from which we can find: M1 = -372000, M2 = -372000.
[0030] Step 2: Calculate the spline segment coefficients (interval [...]) t 1, t 2]) Spline segment expression: S1( t )= a 1+ b 1( t - t 1)+ c 1( t - t 1) 2 + d 1( t - t 1) 3 From step two: a 1= x 1 = 9.9520 (passing through the origin) t 1); b 1=( x 2- x 1) / ts – t s / 6 (2M1+M2)=99507.6; c 1 = M1 / 2 = -186000; d 1 = (M1 - M2) / 6 t s =0 (M1=M2, cubic terms are invalid).
[0031] Step 3: Substitute the target time point to calculate the interpolation result. target point and t Time difference of 1: △ t = 5.025×10 -7 s, Substitute into the spline expression: S1( t 1') = 9.9520 + 99507.6×5.025×10 -7 – 186000×(5.025×10 -7 ) 2 ≈10.0019 The interpolation for other target time points can be calculated using the method described above, thus obtaining the standard frequency voltage sequence. The calculation results for the first 10 points are shown in Table 3. Table 3
[0032] By following the above process, you can obtain the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi.
[0033] S3. Obtain the phase difference Δφi between the in-phase in-circuit voltage signal and the sampled total incoming line voltage signal of each user's sampled in-circuit current signal using the following method: S301. Extract the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit, and normalize it to -180°~180°. This embodiment uses the following method to extract the phase difference between UsL and Isi: the fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi are calculated using the FFT method, and the phase of the fundamental components UsL_0 and Isi_0 is obtained using the phase-locked loop method and the difference between them is calculated, thus obtaining the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit; S302. For each user circuit, calculate the three inherent phase differences between the sampled in-circuit current signal and the sampled total incoming line voltage signal when they are in phase and when they are out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, △θ_native_C_i = △θ_i-120°; Taking user 1 as an example, the Δθ_1 calculated in the previous step is approximately 94°; calculate the three inherent phase differences of user 1 respectively: Δθ_native_A_1 = 94°, Δθ_native_B_1 = 214° = -146° (normalized), Δθ_native_C_1 = -26°; S303. For each user circuit, select one of the three inherent phase differences that falls within the preset inherent phase difference range, and subtract Δθ_i from it to obtain the phase difference Δφi between the in-phase in-circuit voltage signal of the user's sampled in-circuit current signal and the sampled total incoming line voltage signal. This embodiment is a residential electricity consumption scenario, so the preset inherent phase difference range is [-70°, 50°]. Taking user 1 as an example, only Δθ_native_C_1 of its three inherent phase differences is within the range of [-70°, 50°]. Therefore, it can be determined that user 1's current sensor is connected in the C-phase circuit, and thus: Δφ1 = Δθ_native_C_n - Δθ_1 = -120° (or 240°). S4. For each user circuit, the standard frequency voltage sampling data UsL is corrected according to the corresponding phase difference Δφi, thus obtaining the standard frequency voltage sampling data of each user circuit. The correction specifically refers to: applying a phase delay Δφi to the standard frequency voltage sampling data UsL to obtain a new standard frequency voltage sampling data after phase delay; Taking User 1 as an example, since △φ1=-120° (or 240°), the discrete value sequence of UsL is delayed by 240° (i.e., delayed by 13.33ms) to obtain the discrete value sequence of C-phase voltage of User 1; since the sampling rate of this embodiment is 10kHz, then: Us1 [k]= UsL[k+133], and Us1 and Is1 are the standard frequency synchronous voltage and current data of User 1.
[0034] After obtaining the synchronous standard frequency voltage sampling data and standard frequency current sampling data of users 1 to 4 through the above process, these data can be input into the load identification unit to identify the load of each user. For example, the load identification results can be used to monitor whether a user is charging an electric bicycle indoors. The load identification unit can adopt various existing load identification models according to actual needs, which is not an innovation of this invention and will not be elaborated here for the sake of space.
Claims
1. A non-intrusive load identification method, according to the standard frequency voltage sampling data and the standard frequency current sampling data of each user accessing the same total line end, respectively, each user is identified by non-intrusive load identification; characterized in that, The standard frequency voltage sampling data and the standard frequency current sampling data of each user are obtained by using the following method: S1. Sampling the voltage signal of any phase of the total incoming line and the current signal of any household loop of each user synchronously for at least one complete voltage period at the same sampling frequency to obtain the voltage sampling data UL of the total incoming line and the current sampling data Ii of each household loop, wherein i is the serial number of any household loop connected to the total incoming line; S2. Restoring the voltage sampling data UL and the current sampling data Ii to the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi; S3. Obtaining the phase difference △φi between the in-phase household loop voltage signal of the sampled household loop current signal and the sampled total incoming line voltage signal of each user by using the following method: S301. Extracting the phase difference △θ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each household loop, and normalizing it to -180°-180°; S302. For each household loop, calculating three inherent phase differences between the sampled household loop current signal and the sampled total incoming line voltage signal when they are in phase and out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, and △θ_native_C_i = △θ_i-120°; S303. For each household loop, selecting one from the three inherent phase differences that falls within a preset inherent phase difference range, and subtracting △θ_i from it to obtain the phase difference △φi between the in-phase household loop voltage signal of the sampled household loop current signal and the sampled total incoming line voltage signal of the user; S4. For each household loop, modifying the standard frequency voltage sampling data UsL according to the corresponding phase difference △φi to obtain the standard frequency voltage sampling data of each household loop.
2. The non-invasive load identification method of claim 1, wherein, The voltage sampling data UL and the current sampling data Ii are restored to the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi by using the cubic spline interpolation method.
3. The non-invasive load identification method of claim 1, wherein, The inherent phase difference range is [-70°, 50°].
4. The non-invasive load identification method of claim 1, wherein, The phase difference between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each household loop is extracted by using the following method: calculating the fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi by using the FFT method, and obtaining the phase of the fundamental components UsL_0 and Isi_0 by using the phase-locked loop method and calculating the difference between them to obtain the phase difference △θ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each household loop.
5. The non-invasive load identification method of claim 1, wherein, Modifying the standard frequency voltage sampling data UsL according to the corresponding phase difference △φi means applying a phase delay △φi to the standard frequency voltage sampling data UsL to obtain new standard frequency voltage sampling data after phase delay.
6. A non-intrusive load identification device comprising a load identification unit for performing non-intrusive load identification on each user according to standard frequency voltage sampling data and standard frequency current sampling data of each user accessing the same total line end; characterized in that, It also includes a data acquisition unit for acquiring standard frequency voltage sampling data and standard frequency current sampling data for each user through the following methods: S1. At the same sampling frequency, the voltage signal of any phase of the main incoming line and the current signal of any inlet circuit of each user are sampled synchronously for at least one complete voltage cycle to obtain the voltage sampling data UL of the main incoming line and the current sampling data Ii of each user circuit, where i is the sequence number of any user circuit connected to the main incoming line. S2. Restore the voltage sampling data UL and current sampling data Ii to the standard frequency voltage sampling data UsL and standard frequency current sampling data Isi at the standard frequency. S3. Obtain the phase difference Δφi between the in-phase in-circuit voltage signal and the sampled total incoming line voltage signal of each user's sampled in-circuit current signal using the following method: S301. Extract the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit, and normalize it to -180°~180°. S302. For each user circuit, calculate the three inherent phase differences between the sampled in-circuit current signal and the sampled total incoming line voltage signal when they are in phase and when they are out of phase: △θ_native_A_i = △θ_i, △θ_native_B_i = △θ_i + 120°, △θ_native_C_i = △θ_i-120°; S303. For each user circuit, select one of the three inherent phase differences that falls within the preset inherent phase difference range, and subtract Δθ_i from it to obtain the phase difference Δφi between the in-phase in-circuit voltage signal of the user's sampled in-circuit current signal and the sampled total incoming line voltage signal. S4. For each user circuit, the standard frequency voltage sampling data UsL is corrected according to the corresponding phase difference Δφi, thus obtaining the standard frequency voltage sampling data of each user circuit.
7. The non-invasive load discrimination device of claim 5, wherein, The voltage sampling data UL and the current sampling data Ii were restored to the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi using the cubic spline interpolation method.
8. The non-invasive load discrimination device of claim 5, wherein, The inherent phase difference range is [-70°, 50°].
9. The non-invasive load discrimination device of claim 5, wherein, The phase difference between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit is extracted using the following method: The fundamental component UsL_0 of UsL and the fundamental component Isi_0 of Isi are calculated using the FFT method, and the phase of the fundamental components UsL_0 and Isi_0 is obtained using the phase-locked loop method, and the difference between the two is calculated. Thus, the phase difference Δθ_i between the standard frequency voltage sampling data UsL and the standard frequency current sampling data Isi of each user circuit is obtained.
10. The non-invasive load discrimination device of claim 5, wherein, Correcting the standard frequency voltage sampling data UsL according to the corresponding phase difference Δφi means applying a phase delay Δφi to the standard frequency voltage sampling data UsL to obtain a new standard frequency voltage sampling data with phase delay.