Intelligent electric kettle automatic fault detection method and system

Abstract

The present application belongs to the field of intelligent detection and feature quantization, and proposes an automatic fault detection method and system for intelligent electric kettle, specifically: first, identify the execution points from the intelligent electric kettle, then form an execution current sequence from each execution point, perform Fourier transform feature extraction from the execution current sequence to form an execution feature group, then calculate the abnormal pointing scalar through the execution feature group, and finally make fault judgment according to the abnormal pointing scalar. The present application collects and analyzes the current signal data of the intelligent electric kettle in real time during the working process. In the case that the intelligent electric kettle has multiple functions, the cumbersome setting process of monitoring and modulating the control components of the intelligent electric kettle is avoided, and the function program where the fault occurs can be automatically located, and then the error code is displayed to realize fault judgment, while avoiding fault error judgment caused by asynchronous state between multiple monitoring methods, greatly improving the specific function fault recognition accuracy of the intelligent electric kettle.

Description

Technical Field

[0001] This invention belongs to the field of intelligent detection and feature quantification, and specifically relates to an automatic fault detection method and system for an intelligent electric kettle. Background Technology

[0002] Smart electric kettles possess diverse functionalities, leading to their wide range of applications. This versatility relies on both the software programs programmed by the manufacturer and various sensing and monitoring methods. The combination of these two aspects greatly enriches the kettle's application scenarios. However, the manufacturer-programmed functions typically have immutable requirements, achieving the desired function through the combined action of sensing and control. Therefore, multi-functional electric kettles place high demands on the monitoring and control sensitivity of their control components, including sensors and controllers. With increasing usage years, these sensors and controllers experience varying degrees of aging, often with slow-moving changes, leading to kettle malfunctions. This is one of the main technical challenges in the design and production of smart electric kettles. Currently, the industry standard is to monitor each sensor and controller. When a monitored device ages or malfunctions, its corresponding function is disabled, preventing users from experiencing functional discrepancies—the problem of the actual function not matching the intended setting. However, this method suffers from several drawbacks. First, the monitoring behavior of each component requires individual modulation of the functional program. Each preset functional program necessitates separate threshold measurements and settings for its various control components through experimentation. Therefore, the setup process for multi-functional electric kettles is inefficient and cumbersome. Second, the integrity of the monitoring function risks being compromised over time because this method relies on multiple circuits monitoring components, which can easily lead to asynchronous states between monitoring methods and incorrect fault determinations. Therefore, a more efficient and intelligent fault detection method and system are urgently needed. Summary of the Invention

[0003] The purpose of this invention is to propose an automatic fault detection method and system for intelligent electric kettles, so as to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0004] To achieve the above objectives, according to one aspect of the present invention, an automatic fault detection method for a smart electric kettle is provided, the method comprising the following steps:

[0005] The system identifies execution points from the smart electric kettle; each execution point forms an execution current sequence; Fourier transform features are extracted from the execution current sequence to form an execution feature group; an anomaly pointer scalar is calculated using the execution feature group; and fault diagnosis is performed based on the anomaly pointer scalar.

[0006] Furthermore, the method for identifying execution points from a smart electric kettle is as follows: the smart electric kettle has several functional programs, and each execution of any functional program is recorded as an execution point of that functional program. Execution points in which the work task is not fully executed and the duration is less than 120 seconds are excluded.

[0007] Furthermore, the method for forming the execution current sequence at each execution point is as follows: for any execution point, the current signal of the control element is collected in real time by the current sensor of the smart electric kettle, and the current signals obtained from the start to the end of the execution point are constructed into a sequence according to the time order, which is denoted as the execution current sequence.

[0008] Furthermore, the method for extracting Fourier transform features from the execution current sequence to form the execution feature group is as follows: the execution current sequence is converted into a frequency domain signal using a fast Fourier transform to obtain spectral data, and the main frequency and spectral amplitude vector are extracted from the spectral data as the execution feature group.

[0009] Furthermore, the method for calculating the anomaly pointing scalar through the execution feature group is as follows: the average value of all components of the spectral amplitude vector is denoted as ave. For any spectral amplitude vector, the maximum value of its components is denoted as Mhx, the minimum value as Mhn, and the average value as ave. The balance divergence Vle is calculated as follows: Vle = ave(Mhx - ave) / (ave - Mhn). If the balance divergence of the execution point is greater than the component corresponding to the main frequency, it is denoted as the first divergence point; otherwise, it is a stable divergence execution point. The first stable divergence execution point in the reverse time direction of any first divergence point is taken as its inverse search execution point.

[0010] Calculate the cosine similarity between the first scattered point and its inverse search execution point. If the cosine similarity is greater than 1, then... Then, the first scattered point is designated as the second scattered point, and i is set as the index of the second scattered point. The total number of second scattered points within the fault tolerance interval is denoted as Num. The spectral amplitude vectors of the i-th second scattered point and its inverse search execution point are respectively denoted as crt. i and cht i The main frequencies are denoted as crt. i .mhz and cht i .mhz; Remember crt i .ave and cht i The ratio of .ave is the inversion coefficient RA. i , will cht i and CRT iThe cosine similarity is denoted as SE. i ; Calculate the anomaly pointing to the scalar Nstg:

[0011] ;

[0012] CRT i .ax and crt i .av represent the variance and mean of each component in the i-th second scatter point, respectively, and ln() is the logarithmic function with the natural logarithm e as the base.

[0013] The above-mentioned process of calculating the abnormal pointer scalar is based on the screening of the second scattered element point. Therefore, it can effectively quantify the risk of functional deviation induced by the slow aging of individual components in the complex circuit design of a multifunctional smart electric kettle where the tasks of various components overlap. However, since this quantification method has low comparative strength with historical data, it is more likely to lose adaptability to the low-order data of the abnormal pointer scalar value in the subsequent detection process for electric kettles with longer usage time, resulting in incorrect judgment results of fault detection. To solve this problem, the present invention also proposes another better solution.

[0014] Preferably, the method for calculating the anomaly pointer scalar through the execution feature group is as follows: the mode of the main frequencies of all execution points is denoted as the reference main frequency; the spectral amplitude sequence corresponding to any reference main frequency constitutes the frequency amplitude feature vector; the standard deviation of all amplitude values ​​in the frequency amplitude feature vector is the frequency amplitude outlier rate; the weighted average of the frequency amplitude feature vectors of all reference main frequencies is denoted as the reference frequency amplitude feature vector; the sum of the squares of all amplitude values ​​in the reference frequency amplitude feature vector is denoted as the reference spectral energy; if the sum of the squares of all amplitude values ​​in the spectral amplitude sequence corresponding to any execution point is greater than the reference spectral energy, then the execution point is denoted as a singular spectral point; and all execution points between the singular spectral point and the first singular spectral point in the reverse execution time direction are set as an execution point set, denoted as a sub-spectral retrieval set.

[0015] Within the sub-spectrum retrieval set, the number of all non-singular spectral points is denoted as the amplitude lag order (Lago). Based on the amplitude lag order and the spectral amplitude sequence, the amplitude correlation (Framp) is calculated.

[0016] ;

[0017] Where MNum is the number of amplitude values ​​contained in each spectral amplitude sequence, FSmag and ENmag are the spectral amplitude sequences corresponding to the first and last singular spectral points in the sub-spectral retrieval set, respectively, max() and hs() are the maximum value function and harmonic mean function, respectively, and exp() is the exponential function with the natural constant e as the base.

[0018] The calculation principle of frequency amplitude correlation is that the spectral complexity of other execution points can be determined by calculating the frequency amplitude correlation between two specific execution points, without the need to quantize the spectral amplitude sequence of each execution point. This effectively optimizes the monitoring process and reduces the inefficient and cumbersome setup process in fault analysis.

[0019] Within the sub-spectrum retrieval set, the number of inversions of all amplitude values ​​in the spectral amplitude sequence corresponding to any execution point is denoted as the spectral entropy value. The product of the spectral entropy value and the frequency amplitude correlation of any execution point is the spectral complexity Plety of that execution point. The execution point corresponding to the maximum dominant frequency in the sub-spectrum retrieval set is denoted as the high-frequency execution point. The spectral amplitude sequence corresponding to any non-high-frequency execution point constitutes a frequency amplitude feature matrix Chama, and the frequency amplitude feature matrix of the high-frequency execution point is denoted as Hchma. The product of the transpose of the frequency amplitude feature matrix of any execution point and the frequency amplitude feature matrix of the high-frequency execution point is denoted as the heterogeneous coherence matrix Cgrum. The difference between the frequency amplitude feature matrix of any execution point and the frequency amplitude feature matrices of other execution points is obtained to obtain all spectral difference feature matrices of that execution point. The variance of all elements in the spectral difference feature matrix is ​​calculated and denoted as the spectral discrepancy number. The mean of all spectral discrepancies numbers of any execution point is denoted as the spectral discrepancy coefficient. The spectral discrepancy coefficients of all execution points are sorted in ascending order, and the ratio of the index value corresponding to any execution point to the total number of execution points is denoted as the spectral risk coefficient Trais.

[0020] The anomaly pointer scalar Anosc for the current execution point is calculated based on the anomaly coherence matrix and the spectral risk coefficient.

[0021] Anosc = ln(Plety) × tr[Cgrum]×Trais MFre ∑ k2=1 MNum (APyu k2 ) 2 ;

[0022] Where k2 is the cumulative variable, APyu k2 Let t be the k2th amplitude value of the spectral amplitude sequence corresponding to the execution point, tr[] is the diagonal summation function, which returns the sum of each diagonal element in the call matrix. The call matrix in the formula is a heterogeneous coherence matrix; ln() is the logarithmic function with the natural constant e as the base.

[0023] Beneficial effects: Since the anomaly pointer scalar is analyzed in a comprehensive manner based on the set of current signal sequences from different periods, it can effectively quantify the risk of functional deviation caused by the slow aging of individual components in the complex circuit design of a multifunctional smart electric kettle where the tasks of various components overlap. This allows for the identification of the probability of program functionality collapse in the current state, thereby maximizing the consistency between the functional effect and the original setting.

[0024] Furthermore, the method for fault judgment based on the abnormal pointer scalar is as follows: For any functional program, an integer variable is preset as the mutation tolerance interval errgp. The minimum value among the abnormal pointer scalars of the most recent errgp execution points is recorded as the nearest abnormal pointer scalar. The average value of the abnormal pointer scalars of all execution points from the first execution point to the current time is recorded as the pointer scalar benchmark. When the nearest abnormal pointer scalar is greater than the pointer scalar benchmark, the functional program is judged to have a fault. When the functional program is used again, the fault is judged by displaying the error code of the functional program on the display screen of the smart electric kettle.

[0025] Preferably, all undefined variables in this invention, if not explicitly defined, can be manually set thresholds.

[0026] This invention also provides an automatic fault detection system for a smart electric kettle. The system includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the automatic fault detection method for a smart electric kettle. This system can run on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers. The runnable system may include, but is not limited to, processors, memory, and server clusters. The processor executes the computer program within the following system units:

[0027] An execution point identification unit is used to identify execution points from a smart electric kettle;

[0028] An execution current sequence acquisition unit is used to generate an execution current sequence for each execution point;

[0029] The feature group construction unit is used to extract Fourier transform features from the execution current sequence to form an execution feature group;

[0030] The anomaly pointer scalar calculation unit is used to calculate the anomaly pointer scalar by executing the feature group;

[0031] The fault diagnosis unit is used to diagnose faults based on the abnormal pointer scalar.

[0032] The beneficial effects of this invention are as follows: This invention provides an automatic fault detection method and system for smart electric kettles. By real-time acquisition and analysis of current signal data during the operation of the smart electric kettle, it effectively quantifies the risk of functional deviation caused by the slow aging of individual components in the complex circuit design of a multifunctional smart electric kettle, where tasks overlap among various components. Functional deviation risk refers to the problem of the actual function not matching the originally set effect. Furthermore, the judgment method of this invention, while adapting to the multifunctional characteristics of smart electric kettles, avoids the cumbersome setting process of monitoring and modulating each control component of the smart electric kettle. This reduces the tedious measurement and testing of component monitoring thresholds during product development, especially in scenarios where monitoring failures occur due to the aging of control components over time. It avoids erroneous fault judgments caused by asynchronous states between multiple monitoring methods. This greatly improves the accuracy of identifying specific functional faults in smart electric kettles and significantly reduces the potential for functional loss during use. Attached Figure Description

[0033] The above and other features of the present invention will become more apparent from the detailed description of the embodiments shown in conjunction with the accompanying drawings. In the accompanying drawings, the same reference numerals denote the same or similar elements. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without any creative effort. In the drawings:

[0034] Figure 1 The diagram shows a flowchart of an automatic fault detection method for a smart electric kettle.

[0035] Figure 2 The diagram shows the structure of an automatic fault detection system for an intelligent electric kettle. Detailed Implementation

[0036] The following will provide a clear and complete description of the concept, specific structure, and technical effects of the present invention in conjunction with the embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of the present invention. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0037] like Figure 1 The diagram shown is a flowchart of an automatic fault detection method for a smart electric kettle. The following section will discuss this method in conjunction with... Figure 1 This invention describes an automatic fault detection method for a smart electric kettle according to an embodiment of the present invention, the method comprising the following steps:

[0038] The system identifies execution points from the smart electric kettle; each execution point forms an execution current sequence; Fourier transform features are extracted from the execution current sequence to form an execution feature group; an anomaly pointer scalar is calculated using the execution feature group; and fault diagnosis is performed based on the anomaly pointer scalar.

[0039] Furthermore, the method for identifying execution points from a smart electric kettle is as follows: the smart electric kettle has several functional programs, and each execution of any functional program is recorded as an execution point of that functional program. Execution points in which the work task is not fully executed and the duration is less than 120 seconds are excluded.

[0040] A smart electric kettle refers to a multi-functional electric kettle that includes pre-set programs. These programs include multi-level temperature control for heating, porridge making, soup making, dessert preparation, and tea brewing. Each execution of a program is recorded as an execution point, meaning each execution point corresponds to one operation of a specific program within the kettle.

[0041] Furthermore, the method for forming the execution current sequence at each execution point is as follows: for any execution point, the current signal of the control element is collected in real time by the current sensor of the smart electric kettle, and the current signals obtained from the start to the end of the execution point are constructed into a sequence according to the time order, which is denoted as the execution current sequence.

[0042] The current signal is the current value obtained from the power line of the electric kettle, thus covering the current consumption information of all internal control and actuator components.

[0043] Furthermore, the method for extracting Fourier transform features from the execution current sequence to form the execution feature group is as follows: the execution current sequence is converted into a frequency domain signal using a fast Fourier transform to obtain spectral data, and the main frequency and spectral amplitude vector are extracted from the spectral data as the execution feature group.

[0044] The spectral data contains multiple frequency components and their corresponding amplitude and phase information; the spectral amplitude vector is a sequence of amplitude values ​​arranged in ascending order of frequency components.

[0045] Furthermore, the method for calculating the anomaly pointing scalar through the execution feature group is as follows: the average value of all components of the spectral amplitude vector is denoted as ave. For any spectral amplitude vector, the maximum value of its components is denoted as Mhx, the minimum value as Mhn, and the average value as ave. The balance divergence Vle is calculated as follows: Vle = ave(Mhx - ave) / (ave - Mhn). If the balance divergence of the execution point is greater than the component corresponding to the main frequency, it is denoted as the first divergence point; otherwise, it is a stable divergence execution point. The first stable divergence execution point in the reverse time direction of any first divergence point is taken as its inverse search execution point.

[0046] Calculate the cosine similarity between the first scattered point and its inverse search execution point. If the cosine similarity is greater than 1, then... Then, the first scattered point is designated as the second scattered point, and i is set as the index of the second scattered point. The total number of second scattered points within the fault tolerance interval is denoted as Num. The spectral amplitude vectors of the i-th second scattered point and its inverse search execution point are respectively denoted as crt. i and cht i The main frequencies are denoted as crt. i .mhz and cht i .mhz; Remember crt i .ave and cht i The ratio of .ave is the inversion coefficient RA. i , will cht i and CRT i The cosine similarity is denoted as SE. i ; Calculate the anomaly pointing to the scalar Nstg:

[0047] ;

[0048] CRT i .ax and crt i .av represents the variance and mean of each component in the i-th second scatter point, ln() is the logarithmic function with the natural logarithm e as the base, and crt i .ave and cht i .ave represents the average value of all components of the spectral amplitude vector of the i-th second scatter point and its inverse search execution point.

[0049] A fault tolerance interval is formed by 30-50 execution points in reverse time from the current execution point; if the number of execution points in the fault tolerance interval is less than 30, subsequent calculations cannot be performed.

[0050] Preferably, the method for calculating the anomaly pointer scalar through the execution feature set is as follows: any execution point corresponds to an execution current sequence feature set consisting of the main frequency MFre and the spectral amplitude sequence APyu. The mode of the main frequencies of all execution points is denoted as the reference main frequency. The spectral amplitude sequence corresponding to any reference main frequency constitutes the frequency amplitude feature vector. The standard deviation of all amplitude values ​​in the frequency amplitude feature vector is the frequency amplitude outlier rate. The weighted average of the frequency amplitude feature vectors of all reference main frequencies is denoted as the reference frequency amplitude feature vector, where the weight of each frequency amplitude feature vector is the reciprocal of the frequency amplitude outlier rate. The cumulative sum of the squares of all amplitude values ​​in the reference frequency amplitude feature vector is denoted as the reference spectral energy. If the cumulative sum of the squares of all amplitude values ​​in the spectral amplitude sequence corresponding to any execution point is greater than the reference spectral energy, then the execution point is denoted as a singular spectral point. All execution points between this singular spectral point and the first singular spectral point in the reverse execution time direction form an execution point set, denoted as a sub-spectral retrieval set.

[0051] Within the sub-spectrum retrieval set, the number of all non-singular spectral points is denoted as the amplitude lag order (Lago). Based on the amplitude lag order and the spectral amplitude sequence, the amplitude correlation (Framp) is calculated.

[0052] ;

[0053] Where MNum is the number of amplitude values ​​contained in each spectral amplitude sequence, FSmag and ENmag are the spectral amplitude sequences corresponding to the first and last singular spectral points in the sub-spectral retrieval set, respectively, max() and hs() are the maximum value function and harmonic mean function, respectively, and exp() is the exponential function with the natural constant e as the base.

[0054] Within the sub-spectrum retrieval set, the number of inversions of all amplitude values ​​in the spectral amplitude sequence corresponding to any execution point is denoted as the spectral entropy value. The product of the spectral entropy value and the frequency amplitude correlation of any execution point is the spectral complexity Plety of that execution point. The execution point corresponding to the maximum dominant frequency within the sub-spectrum retrieval set is denoted as the high-frequency execution point. The spectral amplitude sequence corresponding to any non-high-frequency execution point constitutes a frequency amplitude feature matrix Chama, and the frequency amplitude feature matrix of the high-frequency execution point is denoted as Hchma. Both Hchma and Chama are matrices with one row and MNum columns, and their mathematical expressions are Chama=[MGti_1, MGti_2,…, MGti_MNum] and Hchma=[HGti_1, HGti_2,…, HGti_MNum], respectively. The product of the transpose of the frequency amplitude feature matrix of any execution point and the frequency amplitude feature matrix of the high-frequency execution point is denoted as the heterocoordinate matrix Cgrum, i.e., Cgrum=Chama. T ×Hchma; Subtract the frequency amplitude feature matrix of any execution point from the frequency amplitude feature matrices of other execution points to obtain the spectrum difference feature matrix of all execution points. Calculate the variance of all elements in the spectrum difference feature matrix, denoted as the spectrum discrepancy number. Let the mean of all spectrum discrepancies numbers of any execution point be the spectrum discrepancy coefficient. Sort the spectrum discrepancy coefficients of all execution points in ascending order. Let the ratio of the index value of any execution point to the total number of execution points be the spectrum risk coefficient Trais.

[0055] The anomaly pointer scalar Anosc for the current execution point is calculated based on the anomaly coherence matrix and the spectral risk coefficient.

[0056] Anosc = ln(Plety) × tr[Cgrum]×Trais MFre ∑ k2=1 MNum (APyu k2 ) 2 ;

[0057] Where k2 is the cumulative variable, APyu k2Let t be the k2th amplitude value of the spectral amplitude sequence corresponding to the execution point, tr[] is the diagonal summation function, which returns the sum of each diagonal element in the call matrix. The call matrix in the formula is a heterogeneous coherence matrix; ln() is the logarithmic function with the natural constant e as the base.

[0058] Furthermore, the method for fault diagnosis based on the anomaly pointer scalar is as follows: for any functional program, a preset integer variable is used as the mutation tolerance interval errgp. Its value range is errgp∈[5,20] The minimum value among the exception pointer scalars of the most recent errgp execution points is recorded as the nearest exception pointer scalar. The average value of the exception pointer scalars of all execution points from the first execution point to the current time is recorded as the pointer scalar reference. When the nearest exception pointer scalar is greater than the pointer scalar reference, it is determined that the function program has a fault. The next time the function program is enabled, the fault is judged by displaying the error code of the function program on the display screen of the smart electric kettle.

[0059] The most recent refers to the one with the smallest time distance from the current moment.

[0060] An embodiment of the present invention provides an automatic fault detection system for an intelligent electric kettle, such as... Figure 2 The diagram shows a structural diagram of an automatic fault detection system for an intelligent electric kettle according to the present invention. This embodiment of the automatic fault detection system for an intelligent electric kettle includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps in the above embodiment of the automatic fault detection method for an intelligent electric kettle.

[0061] The system includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program in units of the following system:

[0062] An execution point identification unit is used to identify execution points from a smart electric kettle;

[0063] An execution current sequence acquisition unit is used to generate an execution current sequence for each execution point;

[0064] The feature group construction unit is used to extract Fourier transform features from the execution current sequence to form an execution feature group;

[0065] The anomaly pointer scalar calculation unit is used to calculate the anomaly pointer scalar by executing the feature group;

[0066] The fault diagnosis unit is used to diagnose faults based on the abnormal pointer scalar.

[0067] The aforementioned automatic fault detection system for an intelligent electric kettle can run on computing devices such as desktop computers, laptops, handheld computers, and cloud servers. The system that can run on this system may include, but is not limited to, processors and memory. Those skilled in the art will understand that the example described is merely an illustration of an automatic fault detection system for an intelligent electric kettle and does not constitute a limitation on such a system. It may include more or fewer components, or a combination of certain components, or different components. For example, the aforementioned automatic fault detection system for an intelligent electric kettle may also include input / output devices, network access devices, buses, etc.

[0068] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center of the intelligent electric kettle automatic fault detection system, connecting various parts of the system via various interfaces and lines.

[0069] The memory can be used to store the computer program and / or modules. The processor implements various functions of the intelligent electric kettle automatic fault detection system by running or executing the computer program and / or modules stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0070] Although the invention has been described in considerable detail and particularly with regard to several of the described embodiments, it is not intended to limit itself to any of these details or embodiments or any particular embodiment, thereby effectively covering the intended scope of the invention. Furthermore, the invention has been described above with respect to embodiments foreseeable by the inventors in order to provide a useful description, and non-substantial modifications to the invention that have not yet been foreseen may still represent equivalent modifications.

Claims

1. A method for automatic fault detection in a smart electric kettle, characterized in that, The method includes the following steps: identifying execution points from the smart electric kettle; forming an execution current sequence for each execution point; extracting execution feature groups from the execution current sequences using Fourier transform; calculating an anomaly pointer scalar using the execution feature groups; and determining faults based on the anomaly pointer scalar. The execution feature group is constructed from the current sequence, including the main frequency and the spectral amplitude vector. The method for calculating the anomaly pointer scalar through the execution feature group is as follows: the average value of all components of the spectral amplitude vector is denoted as ave. For any spectral amplitude vector, its maximum value is denoted as Mhx, its minimum value as Mhn, and its average value as ave. The balance divergence Vle is calculated as follows: Vle = ave(Mhx - ave) / (ave - Mhn). If the balance divergence of the execution point is greater than the component corresponding to the main frequency, it is denoted as the first divergence point; otherwise, it is a stable divergence execution point. The first stable divergence execution point in the reverse time direction of any first divergence point is taken as its inverse search execution point. Calculate the cosine similarity between the first scattered point and its inverse search execution point. If the cosine similarity is greater than 1, then... Then, the first scattered point is designated as the second scattered point, and i is set as the index of the second scattered point. The total number of second scattered points within the fault tolerance interval is denoted as Num. The spectral amplitude vectors of the i-th second scattered point and its inverse search execution point are respectively denoted as crt. i and cht i The main frequencies are denoted as crt. i .mhz and cht i .mhz; Remember crt i .ave and cht i The ratio of .ave is the inversion coefficient RA. i , will cht i and CRT i The cosine similarity is denoted as SE. i ; Calculate the anomaly pointing to the scalar Nstg: ； CRT i .ax and crt i .av represents the variance and mean of each component in the i-th second scatter point, respectively.

2. The automatic fault detection method for an intelligent electric kettle according to claim 1, characterized in that, The method for identifying execution points from a smart electric kettle is as follows: A smart electric kettle has several functional programs. Each execution of any functional program is recorded as an execution point of that functional program. Execution points in which the work task is not fully executed and the duration is less than 120 seconds are excluded.

3. The automatic fault detection method for an intelligent electric kettle according to claim 1, characterized in that, The method for forming the execution current sequence at each execution point is as follows: For any execution point, the current signal of the control element is collected in real time by the current sensor of the smart electric kettle, and the current signals obtained from the start to the end of the execution point are constructed into a sequence according to the time order, which is denoted as the execution current sequence.

4. The automatic fault detection method for an intelligent electric kettle according to claim 1, characterized in that, The method for extracting Fourier transform features from the execution current sequence to form the execution feature group is as follows: use Fast Fourier Transform to convert the execution current sequence into a frequency domain signal to obtain spectral data, and extract the main frequency and spectral amplitude vector from the spectral data as the execution feature group.

5. The automatic fault detection method for an intelligent electric kettle according to claim 1, characterized in that, The method of calculating the anomaly pointer scalar by executing feature groups can be replaced by: Let the mode of the dominant frequencies of all execution points be the reference dominant frequency; let the spectral amplitude sequence corresponding to any reference dominant frequency constitute the frequency amplitude feature vector; let the standard deviation of all amplitude values ​​within the frequency amplitude feature vector be the frequency amplitude outlier rate; let the weighted average of the frequency amplitude feature vectors of all reference dominant frequencies be the reference frequency amplitude feature vector; let the sum of the squares of all amplitude values ​​within the reference frequency amplitude feature vector be the reference spectral energy; if the sum of the squares of all amplitude values ​​within the spectral amplitude sequence corresponding to any execution point is greater than the reference spectral energy, then the execution point is defined as a singular spectral point; let all execution points between this singular spectral point and the first singular spectral point in the reverse execution time direction form an execution point set, denoted as a sub-spectral retrieval set; within the sub-spectral retrieval set, let the number of all non-singular spectral points be the frequency amplitude lag order; and calculate the frequency amplitude correlation based on the frequency amplitude lag order and the spectral amplitude sequence. Within the sub-spectrum retrieval set, the number of inversions of all amplitude values ​​in the spectral amplitude sequence corresponding to any execution point is denoted as the spectral entropy value. The product of the spectral entropy value and the frequency amplitude correlation of any execution point is the spectral complexity of that execution point. The execution point corresponding to the maximum dominant frequency in the sub-spectrum retrieval set is denoted as the high-frequency execution point. The spectral amplitude sequence corresponding to any non-high-frequency execution point constitutes a frequency amplitude feature matrix. The frequency amplitude feature matrix of the high-frequency execution point is denoted as . The product of the transpose of the frequency amplitude feature matrix of any execution point and the frequency amplitude feature matrix of the high-frequency execution point is denoted as the heterogeneous coherence matrix. The difference between the frequency amplitude feature matrix of any execution point and the frequency amplitude feature matrices of other execution points is obtained to obtain all spectral difference feature matrices of that execution point. The variance of all elements in the spectral difference feature matrix is ​​calculated and denoted as the spectral discrepancy number. The mean of all spectral discrepancies numbers of any execution point is denoted as the spectral discrepancy coefficient. The spectral discrepancy coefficients of all execution points are sorted in ascending order. The ratio of the index value corresponding to any execution point to the total number of execution points is denoted as the spectral risk coefficient. The anomaly pointing scalar of the current execution point is calculated based on the heterogeneous coherence matrix and the spectral risk coefficient.

6. The automatic fault detection method for an intelligent electric kettle according to claim 1, characterized in that, The method for fault diagnosis based on abnormal pointer scalars is as follows: For any functional program, a preset integer variable is used as the mutation tolerance interval errgp; the minimum value among the abnormal pointer scalars of the most recent errgp execution points is recorded as the nearest abnormal pointer scalar; the average value of the abnormal pointer scalars of all execution points from the first execution point to the current time is recorded as the pointer scalar benchmark; when the nearest abnormal pointer scalar is greater than the pointer scalar benchmark, the functional program is determined to be faulty; the next time the functional program is activated, the fault is diagnosed by displaying the error code of the functional program on the display screen of the smart kettle.

7. An automatic fault detection system for an intelligent electric kettle, characterized in that, The intelligent electric kettle automatic fault detection system includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the intelligent electric kettle automatic fault detection method according to any one of claims 1-6. The intelligent electric kettle automatic fault detection system runs on computing devices such as desktop computers, laptops, handheld computers, and cloud data centers.

Images