CN-121995163-A - Arc fault detection method, device and system

CN121995163ACN 121995163 ACN121995163 ACN 121995163ACN-121995163-A

Abstract

The application provides an arc fault detection method, device and system, which are used for extracting current characteristics, voltage characteristics and load characteristics based on current signals and voltage signals in an input circuit of an electric appliance and operation signals of an internal control module of the electric appliance, inputting the current characteristics, the voltage characteristics and the load characteristics into an arc fault detection model to detect the arc faults, and solving the problems that the arc fault detection is easily interfered by load fluctuation and noise and has poor adaptability to complex load scenes only through current and voltage in the prior art.

Inventors

WANG XUEFENG
XIE QIWEI
LI SHILIN
HE XUEDONG

Assignees

珠海格力电器股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251224

Claims (10)

1. A method of arc fault detection, the method comprising: Collecting current signals and voltage signals in an input circuit of an electric appliance; Acquiring an operation signal of the internal control module of the electrical appliance; extracting a current signature based on the current signal; Extracting a voltage characteristic based on the voltage signal; Extracting load characteristics based on the operation signals of the internal control module of the electric appliance; And (3) inputting the extracted characteristic vector combination consisting of the voltage characteristic, the current characteristic and the load characteristic into a trained arc fault detection model to obtain an arc fault detection result so as to prevent misjudgment of the internal state change of the electric appliance as an arc fault.
2. The method according to claim 1, characterized in that: The operation signals of the internal control module of the electric appliance comprise operation control signals and/or operation state signals; The obtaining the operation signal of the internal control module of the electrical appliance comprises obtaining the operation control signal and/or the operation state signal of the internal control module of the electrical appliance.
3. The method according to claim 2, characterized in that: The load signature includes a load operation signature including a load operation control signature and/or a load operation status signature; the extracting load characteristics based on the operation signal of the internal control module of the electric appliance comprises the following steps: extracting the load operation control characteristics based on the operation control signal of the appliance internal control module, and/or, And extracting the load running state characteristics based on the running state signals of the internal control module of the electric appliance.
4. A method according to claim 3, characterized in that: The load operation control features comprise control signal frequency features, and/or control signal detail features and/or control signal time domain features; Based on the operation control signal of the internal control module of the electrical appliance, extracting the load operation control feature comprises: performing Fast Fourier Transform (FFT) on an operation control signal of the internal control module of the electrical appliance to obtain a frequency characteristic of the control signal; And/or performing wavelet transform (DWT) on the operation control signal of the internal control module of the electrical appliance to obtain the detail characteristics of the control signal; and/or calculating the change rate of the operation control signal of the internal control module of the electrical appliance to obtain the time domain characteristics of the control signal.
5. The method according to any one of claims 1-4, wherein: The internal control module of the electric appliance comprises a control circuit and a control chip; The operation control signal comprises a signal related to control, which is sent to the control circuit by the control chip; the operating state signal comprises a state-related signal acquired from the control chip.
6. The method according to claim 2, characterized in that: The operation control signal of the internal control module of the electric appliance comprises at least one of Pulse Width Modulation (PWM) duty cycle, PWM switching frequency, power Factor Correction (PFC) current instruction and Power Factor Correction (PFC) voltage instruction; and/or the number of the groups of groups, The running state signals of the internal control module of the electric appliance comprise at least one of a relay switch state, a relay protection state, a Power Factor Correction (PFC) hiccup mode, a PFC continuous mode and a PFC discontinuous mode.
7. The method of claim 1, wherein the arc fault detection model is one of a deep learning neural network model of a multi-layer perceptron MLP model, an extreme learning ELM model, a convolutional neural network CNN model comprising at least 2 hidden layers.
8. The method of claim 7, wherein the arc fault detection model is trained as follows: Collecting current signals and voltage signals of different electric appliances in an input circuit in a state without arc faults and in a state with arc faults respectively, and extracting corresponding current characteristics and voltage characteristics; respectively acquiring operation signals of an internal control module of the electric appliance in the arc fault-free state and the arc fault-free state of the different electric appliances, and extracting corresponding load characteristics; Combining the current characteristics, the voltage characteristics and the load characteristics of different electrical appliances in the arc fault-free state and the arc fault-free state to form a characteristic vector combination of a normal sample and an arc sample; and inputting the feature vector combination of the normal sample and the arc sample into an arc fault detection model for training, and obtaining the arc fault detection model after training.
9. An arc fault detection apparatus, comprising: The acquisition unit is used for acquiring current signals and voltage signals in an input circuit of the electrical appliance; The acquisition unit is used for acquiring an operation signal of an internal control module of the electrical appliance; the fault judging unit is used for extracting current characteristics and voltage characteristics based on the current signals and the voltage signals, extracting load characteristics based on operation signals of the internal control module of the electric appliance, and inputting the extracted voltage characteristics, the current characteristics and the load characteristics into a trained arc fault detection model to obtain an arc fault detection result so as to prevent misjudgment of internal state change of the electric appliance as an arc fault.
10. An arc fault detection system comprising a circuit interrupting device and an arc fault detection device as claimed in claim 9, When the arc fault probability output by the arc fault detection device is larger than a preset threshold value, the circuit breaking device cuts off the input power supply of the electric appliance.

Description

Arc fault detection method, device and system Technical Field The present application relates to the field of fault detection technologies, and in particular, to an arc fault detection method, apparatus, and system. Background With the continuous improvement of the electrification degree of the modern society, the application of electrical equipment in production and life is more and more extensive, and the hidden danger of electrical fire is also increasingly prominent. The arc fault is one of main reasons for causing an electrical fire, and the rapid and accurate detection and positioning of the arc fault are of great significance for preventing the fire, protecting equipment and ensuring the reliability of power supply. Traditional arc fault detection methods rely mainly on abnormal changes in current and/or voltage, such as current harmonic analysis, voltage jump detection. However, the inventors have found that the current, voltage of the series circuit is affected by the impedance characteristics of the load, and that certain normal operating conditions of complex loads (e.g. variable frequency motors, light emitting diode LED dimmed lamps) can generate signals that are highly similar to an arc. For example, inductive loads such as high-frequency harmonic waves (200-500 Hz) and current abrupt changes when a variable frequency motor is started overlap with an arc harmonic mode, so that the inductive loads are prone to be misjudged as faults, nonlinear capacitive loads such as 50-100kHz high-frequency pulses, voltage fluctuations and PWM noise of an LED dimming lamp (PWM control without PFC function) are also prone to be misidentified as arc high-frequency components, and resistive loads such as 10-100kHz switching noise and current ripple of a switching power supply with a Power Factor Correction (PFC) function are also prone to be classified as arc signals. Therefore, when the signal wave and noise interference are faced to the complex load scene, the traditional method is difficult to distinguish the fault arc from the normal electric signal and noise interference, and the power supply is cut off due to easy misjudgment, so that unnecessary power failure loss is caused. Disclosure of Invention In view of the above, the present application provides an arc fault detection method, apparatus and system, which extracts a current characteristic and a voltage characteristic based on a current signal and a voltage signal in an input circuit of an electrical apparatus, and extracts a load characteristic through an operation signal of an internal control module of the electrical apparatus, and inputs the load characteristic to an arc fault detection model to perform arc fault detection, so as to prevent misjudging an internal state change of the electrical apparatus as an arc fault. The problem that arc fault detection is easily affected by load fluctuation and noise interference only through current and voltage in the prior art is solved. According to an aspect of the present application, there is provided an arc fault detection method including: Collecting current signals and voltage signals in an input circuit of an electric appliance; Acquiring an operation signal of the internal control module of the electrical appliance; extracting a current signature based on the current signal; Extracting a voltage characteristic based on the voltage signal; Extracting load characteristics based on the operation signals of the internal control module of the electric appliance; And (3) inputting the extracted characteristic vector combination consisting of the voltage characteristic, the current characteristic and the load characteristic into a trained arc fault detection model to obtain an arc fault detection result so as to prevent misjudgment of the internal state change of the electric appliance as an arc fault. Alternatively, the process may be carried out in a single-stage, The operation signals of the internal control module of the electric appliance comprise operation control signals and/or operation state signals; The obtaining the operation signal of the internal control module of the electrical appliance comprises obtaining the operation control signal and/or the operation state signal of the internal control module of the electrical appliance. Alternatively, the process may be carried out in a single-stage, The load signature includes a load operation signature including a load operation control signature and/or a load operation status signature; the extracting load characteristics based on the operation signal of the internal control module of the electric appliance comprises the following steps: extracting the load operation control characteristics based on the operation control signal of the appliance internal control module, and/or, And extracting the load running state characteristics based on the running state signals of the internal control module of the electric appliance. Alternatively, the process