Abnormality detection method and device, washing machine and storage medium

Abstract

The invention relates to an anomaly detection method and device, a washing machine and a storage medium, and the anomaly detection method comprises the steps that first target information of the washing machine is obtained, and the first target information comprises the current water level in the washing machine and the current rotating speed of a motor; determining second target information of the washing machine under the current water level and the current rotating speed; wherein the second target information at least comprises target audio information at a roller in the washing machine, target vibration information of the roller, target audio information of a motor and target vibration information of the motor; and detecting whether the running state of the washing machine is abnormal or not according to the first target information and the second target information.

Description

Technical Field

[0001] The present disclosure relates to the technical field of electrical appliance control, and particularly to an abnormal detection method, apparatus, washing machine and storage medium. Background Art

[0002] With the development of household electrical appliances, the functions of household electrical appliances are becoming more and more, and they are becoming more intelligent. Household electrical appliances include a variety, and the washing machine, as a household electrical appliance with a high usage frequency in daily life, can clean clothes. There are various types and functions of washing machines. Different types of washing machines have different functions and different effects after cleaning clothes.

[0003] The washing machine may have abnormalities during daily use, such as abnormalities in the drum or the motor, which will affect the operation of the washing machine. Summary of the Invention

[0004] The present disclosure provides an abnormal detection method, apparatus, washing machine and storage medium.

[0005] In a first aspect of an embodiment of the present disclosure, an abnormal detection method is provided, including: obtaining first target information of the washing machine, where the first target information includes: the current water level in the washing machine and the current rotation speed of the motor; determining second target information of the washing machine at the current water level and the current rotation speed; where the target information at least includes target audio information at the drum of the washing machine, target vibration information of the drum, target audio information of the motor, and target vibration information of the motor; detecting whether the operating state of the washing machine is abnormal according to the first target information and the second target information.

[0006] In an embodiment, the determining the second target information of the washing machine includes: collecting a first original audio signal at the drum of the washing machine and a second original audio signal of the motor; respectively filtering the first original audio signal and the second original audio signal to obtain a first audio signal corresponding to the first original audio signal and a second audio signal corresponding to the second original audio signal; extracting feature information of the first audio signal to obtain first feature information; extracting feature information of the second audio signal to obtain second feature information; determining the first feature information as the target audio information at the drum, and determining the second feature information as the target audio information of the motor.

[0007] In one embodiment, extracting the feature information of the first audio signal to obtain first feature information includes: extracting a first signal sequence of each rotation period of the roller from the first audio signal; segmenting the first signal sequence to obtain N first target sequences; wherein, each of the first target sequences includes M sampling points; determining a first frequency peak corresponding to each of the first target sequences; determining a second frequency peak of the rotation period according to the first frequency peak; and determining the first feature information according to the second frequency peak.

[0008] In one embodiment, extracting the feature information of the second audio signal to obtain second feature information includes: extracting a second signal sequence of each rotation period of the motor from the second audio signal; segmenting the second signal sequence to obtain N second target sequences; wherein, each of the second target sequences includes M sampling points; determining a third frequency peak corresponding to each of the second target sequences; determining a fourth frequency peak of the rotation period according to the third frequency peak; and determining the second feature information according to the fourth frequency peak.

[0009] In one embodiment, determining the second target information of the washing machine includes: collecting a first original vibration signal at the roller of the washing machine and a second original vibration signal of the motor; respectively filtering the first original vibration signal and the second original vibration signal to obtain a first vibration signal corresponding to the first original vibration signal and a second vibration signal corresponding to the second original vibration signal; extracting the feature information of the first vibration signal to obtain third feature information; extracting the feature information of the second vibration signal to obtain fourth feature information; determining the third feature information as the target vibration information at the roller, and determining the fourth feature information as the target vibration information of the motor.

[0010] In one embodiment, respectively filtering the first original vibration signal and the second original vibration signal to obtain a first vibration signal corresponding to the first original vibration signal and a second vibration signal corresponding to the second original vibration signal includes: determining a maximum value and a minimum value of a first target vibration signal; wherein, the first target vibration signal is the first original vibration signal or the second original vibration signal; determining a plurality of components of the first target vibration signal according to the maximum value, the minimum value and the first target vibration signal; determining a second target vibration signal according to the components determined in the previous k times; wherein, the second target vibration signal is the first vibration signal or the second vibration signal; wherein, when the first target vibration signal is the first original vibration signal, the second target vibration signal is the first vibration signal; and when the first target vibration signal is the second original vibration signal, the second target vibration signal is the second vibration signal.

[0011] In one embodiment, extracting the characteristic information of the first vibration signal to obtain third characteristic information includes: extracting a third signal sequence of each rotation period of the drum from the first vibration signal; segmenting the third signal sequence to obtain F third target sequences; wherein, each of the third target sequences includes H sampling points; determining the first energy of each of the third target sequences, the first total energy of the F third target sequences, and the first proportion of each of the first energies in the first total energy; determining the first energy entropy of each of the rotation periods according to the first energy, the first total energy, and the first proportion; and determining the third characteristic information according to the first energy entropy.

[0012] In one embodiment, extracting the characteristic information of the second vibration signal to obtain fourth characteristic information includes: extracting a fourth signal sequence of each rotation period of the motor from the second vibration signal; segmenting the fourth signal sequence to obtain F fourth target sequences; wherein, each of the fourth target sequences includes H sampling points; determining the second energy of each of the fourth target sequences, the second total energy of the F fourth target sequences, and the second proportion of each of the second energies in the second total energy; determining the second energy entropy of each of the rotation periods according to the second energy, the second total energy, and the second proportion; and determining the fourth characteristic information according to the second energy entropy.

[0013] In one embodiment, determining whether the operating state of the washing machine is abnormal includes: using a target detection model to determine whether the operating state of the washing machine is abnormal according to the current water level, the current rotation speed, the target audio information at the drum, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor.

[0014] In one embodiment, the target detection model is a model obtained by training an initial network model using a training sample set in a machine learning manner; the training sample set includes: a positive sample set, including a plurality of different positive samples, each positive sample including: a water level sample, a positive motor speed sample, sample information of normal sound at the drum in the washing machine, sample information of normal vibration of the drum, sample information of normal sound of the motor, and sample information of normal vibration of the motor; the water levels and / or motor speeds corresponding to different positive samples are the same or different; each of the positive samples has its own first label, which is used to identify that the output of the positive sample corresponding to the initial network model is normal; a negative sample set, including a plurality of negative samples, each negative sample including: the water level sample, a negative motor speed sample, negative sample audio information at the drum in the washing machine, negative sample vibration information of the drum, negative sample audio information of the motor, and negative sample vibration information of the motor; wherein, at least one of the negative sample audio information at the drum in the washing machine, the negative sample vibration information of the drum, the negative sample audio information of the motor, and the negative sample vibration information of the motor is abnormal information; each of the negative samples has its own second label, which is used to identify that the output of the negative sample corresponding to the initial network model is abnormal.

[0015] In a second aspect of the embodiments of the present disclosure, there is provided an anomaly detection device, including: an acquisition module, configured to acquire first target information of a washing machine, where the first target information includes: the current speed and current water level of the motor in the washing machine; a determination module, configured to determine second target information of the washing machine at the current water level and the current speed; where the target information at least includes target audio information at the drum in the washing machine, target vibration information of the drum, target audio information of the motor, and target vibration information of the motor; a detection module, configured to detect whether an abnormal operation state of the washing machine occurs according to the first target information and the second target information.

[0016] In a third aspect of the embodiments of the present disclosure, there is provided a washing machine, including:

[0017] a processor and a memory for storing executable instructions that can run on the processor, wherein: when the processor is used to run the executable instructions, the executable instructions execute the method according to any one of the above embodiments.

[0018] In a fourth aspect of the embodiments of the present disclosure, there is provided a non-transitory computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, the method according to any one of the above embodiments is implemented.

[0019] The technical solutions provided by the embodiments of the present disclosure may include the following beneficial effects:

[0020] The solution provided by the embodiments of the present disclosure obtains the current water level in the washing machine and the current rotational speed of the motor, and combines the target audio information, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor at the current water level and the current rotational speed to detect whether the operating state of the washing machine is abnormal. By determining whether the washing machine is abnormal through multiple types of information, the accuracy of detecting whether the washing machine is abnormal is improved, and dynamic detection and analysis of the working state of the washing machine can be achieved, facilitating the user to obtain whether the laundry state of the washing machine is normal.

[0021] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure.

[0023] Figure 1 is a schematic diagram of an anomaly detection method shown according to an exemplary embodiment;

[0024] Figure 2 is a schematic diagram of determining target information shown according to an exemplary embodiment;

[0025] Figure 3 is a schematic diagram of an original audio signal collected shown according to an exemplary embodiment;

[0026] Figure 4 is a waveform schematic diagram of a first original audio signal or a second original audio signal shown according to an exemplary embodiment;

[0027] Figure 5 is a waveform schematic diagram of a first audio signal or a second audio signal obtained after filtering shown according to an exemplary embodiment;

[0028] Figure 6 is a schematic diagram of obtaining first feature information shown according to an exemplary embodiment;

[0029] Figure 7 is a schematic diagram of obtaining second feature information shown according to an exemplary embodiment;

[0030] Figure 8 is a schematic diagram of a target sequence shown according to an exemplary embodiment;

[0031] Figure 9 is a schematic diagram of a frequency peak shown according to an exemplary embodiment;

[0032] Figure 10 It is a schematic diagram showing the determination of target information of a washing machine according to an exemplary embodiment;

[0033] Figure 11 It is a schematic diagram of an acquired original vibration signal according to an exemplary embodiment;

[0034] Figure 12 It is a waveform schematic diagram of another first original vibration signal or second original vibration signal according to an exemplary embodiment;

[0035] Figure 13 It is a schematic diagram showing the filtering of the original vibration signal according to an exemplary embodiment;

[0036] Figure 14 It is a schematic diagram showing the determination of components of the first original vibration signal or the second original vibration signal according to an exemplary embodiment;

[0037] Figure 15 It is a schematic diagram of components of a first target vibration signal according to an exemplary embodiment;

[0038] Figure 16 It is according to an exemplary embodiment showing a Figure 15 corresponding frequency domain schematic diagram;

[0039] Figure 17(a) is a schematic diagram of a first target vibration signal according to an exemplary embodiment;

[0040] Figure 17(b) is a spectrogram of a first target vibration signal according to an exemplary embodiment;

[0041] Figure 18(a) is a waveform schematic diagram of a first target vibration signal obtained after filtering the first target vibration signal according to an exemplary embodiment;

[0042] Figure 18(b) is a spectrogram of a second target vibration signal obtained after filtering the first target vibration signal according to an exemplary embodiment;

[0043] Figure 19 It is a schematic diagram showing the extraction of third feature information according to an exemplary embodiment;

[0044] Figure 20 It is a waveform schematic diagram of a third target sequence according to an exemplary embodiment;

[0045] Figure 21 It is a schematic diagram showing the determination of fourth feature information according to an exemplary embodiment;

[0046] Figure 22 Schematic diagram of an anomaly detection device shown according to an exemplary embodiment;

[0047] Figure 23 Schematic diagram of the energy entropy of positive and negative samples shown according to an exemplary embodiment;

[0048] Figure 24 Schematic diagram of the waveforms of positive and negative samples shown according to an exemplary embodiment;

[0049] Figure 25 Block diagram of a washing machine shown according to an exemplary embodiment. Detailed implementation manners

[0050] Here, the exemplary embodiments will be described in detail, and examples thereof are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present disclosure. On the contrary, they are merely examples of devices consistent with some aspects of the present disclosure as detailed in the appended claims.

[0051] Reference Figure 1 , is a schematic diagram of an anomaly detection method, which can be applied to a washing machine. The method includes:

[0052] S100: Obtain first target information of the washing machine, where the first target information includes: the current water level in the washing machine and the current rotation speed of the motor.

[0053] S200: Determine second target information of the washing machine at the current water level and the current rotation speed; where the second target information at least includes target audio information at the drum in the washing machine, target vibration information of the drum, target audio information of the motor, and target vibration information of the motor.

[0054] S300, Detect whether the operating state of the washing machine is abnormal according to the first target information and the second target information.

[0055] For S100, the washing machine can be any model of washing machine. The structures, sizes, and various parameters of different washing machines may be different, and washing machines having a motor and a drum are within the protection scope of this embodiment.

[0056] The washing machine may also have sensors, and the sensors can detect the current water level and the current rotation speed. For example, a sensor for detecting the water level and a sensor for detecting the rotation speed of the motor. The models and parameters of the sensors are not limited and can be determined according to actual usage requirements.

[0057] Exemplarily, the sensor for detecting the water level can be a water level detector, and the sensor for detecting the motor speed can be a speed sensor.

[0058] The positions of the sensor for detecting the current water level and the sensor for detecting the current speed can be determined according to actual usage requirements. For example, the sensor for detecting the current water level can be located inside the drum, and the sensor for detecting the current speed can be located in the motor.

[0059] Of course, the positions of these sensors can be at other locations as long as they can achieve the function of detecting the corresponding information.

[0060] The washing machine can also have a controller. These sensors are connected to the controller and can interact with the controller for information. During the operation of the washing machine, these sensors collect the corresponding information and transmit the collected information to the controller.

[0061] Exemplarily, the method for obtaining the current water level in the washing machine can also include: determining the current water level according to the water injection volume, the capacity of the drum, and the water absorption of the currently to-be-washed laundry. The water absorption of the currently to-be-washed laundry can be determined according to the material and weight of the currently to-be-washed laundry.

[0062] Exemplarily, the washing machine can have different washing modes. The current speeds of the motors corresponding to different washing modes may be different, and the motor speeds corresponding to each washing mode can be preset, for example, stored in the controller, and the controller can directly obtain the corresponding speed.

[0063] Exemplarily, the current water level and the current speed can also be the water level and speed when there is water and the laundry to be washed in the drum and the laundry in the drum is being washed.

[0064] Exemplarily, the current water level can also be the water level when there is no water and no laundry to be washed in the drum, and the current water level at this time is 0.

[0065] Exemplarily, the current water level can also be the water level when there is water in the drum and no laundry to be washed.

[0066] Exemplarily, the current speed of the motor can be the speed when there is no water and no laundry to be washed in the drum, that is, the idling speed.

[0067] Exemplarily, the current speed of the motor can be the speed when there is water in the drum and no laundry to be washed.

[0068] Exemplarily, the current speed of the motor can be the speed when there is laundry to be dried in the drum.

[0069] Exemplarily, the current speed of the motor can also be 0, that is, the situation where the motor does not rotate.

[0070] The first target information of the washing machine can be determined by the above method, that is, the current water level in the washing machine and the current rotational speed of the motor.

[0071] For S200, during the operation of the washing machine to wash clothes, the drum and the motor are in a rotating state, and the motor drives the drum to rotate. The rotational speed of the motor is the same as that of the drum, the current rotational speed of the motor is the current rotational speed of the drum, and the rotation direction of the motor is the same as that of the drum.

[0072] Due to the characteristics of the hardware, the motor and the drum will vibrate during rotation, there may be a certain amount of deviation, and sounds will be emitted due to friction or other reasons during rotation. During this process, the second target information for detecting whether the washing machine has an abnormality can be determined.

[0073] The second target information may include the target audio information at the drum of the washing machine, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor. The information included in the target information may be normal information or abnormal information. Therefore, it is necessary to obtain this information and determine whether the washing machine has an abnormality based on this information. When this information is normal information, the washing machine is normal; when this information is abnormal information, the washing machine has an abnormality.

[0074] Since the current water level and the current rotational speed of the washing machine have been obtained, the target information at the current water level and the current rotational speed can be determined. Thus, it is possible to determine whether the washing machine has an abnormality during operation by combining the current water level, the current rotational speed of the motor, and the second target information, thereby improving the accuracy of detection.

[0075] The target audio information at the drum may be the sound of the drum rotating, or the sound jointly generated by the sound of the drum rotating, the water in the drum, and the objects to be washed in the drum.

[0076] The target vibration information of the drum may be the vibration information generated by the drum at the current water level and the current rotational speed, the target audio information of the motor is the audio information generated by the motor at the current water level and the current rotational speed, and the target vibration information of the motor may be the vibration information generated by the motor at the current water level and the current rotational speed.

[0077] The target audio information at the drum and the target audio information of the motor can be obtained by a sensor for detecting sound, and the target vibration information of the drum and the target vibration information of the motor can be obtained by a sensor for detecting vibration.

[0078] Exemplarily, the target audio information at the drum and the target audio information of the motor may still be the information obtained by processing the audio information acquired by the sensor for detecting sound. The target vibration information of the drum and the target vibration information of the motor may also be the information obtained by processing the vibration information acquired by the sensor for detecting vibration.

[0079] Exemplarily, the sensor for detecting audio information may be a sound sensor, such as a sound wave sensor, including a microphone, etc. The sensor for detecting vibration may be a displacement sensor or a vibration sensor, etc.

[0080] Exemplarily, the position of the sensor may also be determined according to actual usage requirements. For example, the sensor for detecting the target audio information at the drum may be located on the drum, on the outer sidewall of the drum or at a position close to the outer sidewall of the drum. The sensor for detecting the target vibration information of the drum may be located on the drum, the sensor for detecting the target audio information of the motor may be located on the outer shell of the motor, and the sensor for detecting the target vibration information of the motor may be located on the outer shell of the motor.

[0081] Exemplarily, after the target information is detected by the relevant sensor components, these sensor components may send the detected information to the controller in the washing machine.

[0082] For S300, after determining the first target information and the second target information, it is possible to detect whether the operating state of the washing machine is abnormal according to the first target information and the second target information. It is possible to determine whether the operating state of the washing machine is abnormal through the controller in the washing machine according to the first target information and the second target information, and the determination process and determination method are not limited.

[0083] For example, it is possible to detect whether the operating state of the washing machine is abnormal based on the first target information and the second target information through a trained detection model. The first target information and the second target information may be used as the input of the detection model, and the input of the target model is the detection result.

[0084] Of course, it is also possible to perform detection based on the first target information and the second target information in other ways. For example, it is possible to determine whether the operating state of the washing machine is abnormal according to the first target information, the second target information and the reference information.

[0085] Determining whether the washing machine is abnormal through multiple pieces of information improves the accuracy of detecting whether the washing machine is abnormal, enables dynamic detection and analysis of the working state of the washing machine, and facilitates the user to obtain whether the laundry state of the washing machine is normal.

[0086] In one embodiment, refer to Figure 2 , which is a schematic diagram for determining the second target information. Determining the second target information of the washing machine includes:

[0087] S201, collect the first original audio signal at the drum of the washing machine and the second original audio signal of the motor.

[0088] S202, filter the first original audio signal and the second original audio signal respectively to obtain the first audio signal corresponding to the first original audio signal and the second audio signal corresponding to the second original audio signal.

[0089] S203, extract the feature information of the first audio signal to obtain the first feature information.

[0090] S204, extract the feature information of the second audio signal to obtain the second feature information.

[0091] S205, determine the first feature information as the target audio information at the drum, and determine the second feature information as the target audio information of the motor.

[0092] Exemplarily, S203 can be executed first, or S204 can be executed first.

[0093] The first original audio signal is the audio signal at the drum collected by the sensor, and the second original audio signal is the audio signal emitted by the motor during rotation collected by the sensor. As unprocessed audio signals, the first original audio signal and the second original audio signal may include some other noises and clutter, such as audio signals in the environment, and these noises will affect the detection result and reduce the detection accuracy.

[0094] By filtering the first original audio signal and the second original audio signal, the noises in the first original audio signal and the second original audio signal can be filtered out, and the obtained audio signals are closer to the real audio signal at the drum and the real audio signal emitted by the motor.

[0095] After filtering the first original audio signal, the first audio signal is obtained, and after filtering the second original audio signal, the second audio signal is obtained.

[0096] The filtering method can be determined according to the actual service requirements, and all methods capable of filtering are within the protection scope of this embodiment. For example, various filters, or a filtering method based on the frequency of the audio signal are used to filter the first original audio signal and the second original audio signal. Including high-pass filtering, where high-frequency audio signals pass through and low-frequency noise signals are filtered to obtain the first audio signal and the second audio signal.

[0097] Reference Figure 3 , is a schematic diagram of a collected original audio signal, where the horizontal axis is the sampling point and the vertical axis is the amplitude, that is, the magnitude. Figure 3What is shown may be a first original audio signal representing the sound collected at the drum, or a second original audio signal representing the sound collected from the motor.

[0098] Reference Figure 4 , is a waveform schematic diagram of a first original audio signal or a second original audio signal. Reference Figure 5 , is a waveform schematic diagram of a first audio signal or a second audio signal obtained after filtering. Figure 4 And Figure 5 Are frequency domain diagrams.

[0099] Figure 4 The frequency of the first original audio signal within the box in [[ ]] is relatively low and is filtered out by filtering, Figure 5 The waveform within the box shown in [[ ]] is the waveform of the first audio signal after filtering. Figure 5 The white noise and / or low-frequency clutter in the first audio signal or the second audio signal shown in [[ ]] are reduced compared to Figure 4 Before filtering, thereby improving the signal-to-noise ratio of the first audio signal and the second audio signal. Figure 4 And Figure 5 Are frequency domain diagrams.

[0100] Exemplarily, the first original audio signal and the second original audio signal can be filtered multiple times to obtain the first audio signal and the second audio signal.

[0101] After obtaining the first audio signal and the second audio signal, since there may still be insufficient orthogonality, insufficient smoothness, and unprominent characteristic information in the filtered audio signals, the characteristic information of the first audio signal can be extracted to obtain the first characteristic information; the characteristic information of the second audio signal can be extracted to obtain the second characteristic information. The feature extraction method is not limited, and any method capable of performing feature extraction on audio signals can be used. Including but not limited to: Fourier transform, fast Fourier transform, wavelet transform, and short-time Fourier transform, etc., and feature extraction can also be performed through convolution and other methods. The detailed feature extraction process can be referred to the subsequent embodiments.

[0102] After obtaining the first characteristic information and the second characteristic information, the first characteristic information is determined as the target audio information at the drum, and the second characteristic information is determined as the target audio information of the motor.

[0103] Through S201 to S205, the quality of the obtained target audio signal is higher, and it can better represent the sound at the drum and the sound emitted by the motor itself, thereby facilitating the improvement of the detection accuracy.

[0104] In one embodiment, filtering the first original audio signal to obtain a first audio signal, and filtering the second original audio signal to obtain a second audio signal can be performed through the following formula (1):

[0105] y[i] = α × y[i - 1] + α × (x[i] - x[i - 1]), α ∈ (0, 1) (1)

[0106] Where i represents the number of filtering times, x[i] represents the first original audio signal or the second original audio signal after i times of filtering, α represents the filtering coefficient, and y[i] represents the first audio signal or the second audio signal, i ∈ [0, ∞+).

[0107] In one embodiment, referring to Figure 6 , it is a schematic diagram for obtaining the first feature information. S203, extracting the feature information of the first audio signal to obtain the first feature information, includes:

[0108] S2031, extracting the first signal sequence of each rotation period of the drum from the first audio signal.

[0109] S2032, segmenting the first signal sequence to obtain N first target sequences; where each first target sequence includes M sampling points.

[0110] S2033, determining the first frequency peak corresponding to each first target sequence.

[0111] S2034, determining the second frequency peak of the rotation period according to the first frequency peak.

[0112] S2035, determining the first feature information according to the second frequency peak.

[0113] Since the drum is in a rotating state, the first audio signal includes audio signals corresponding to multiple rotation periods of the drum. After obtaining the first audio signal, the first signal sequence of the drum in each rotation period is extracted from the first audio signal, starting from the first rotation period, and the first signal sequence in each rotation period is extracted.

[0114] When the current water level remains unchanged, the current rotation speed remains unchanged, and the objects to be cleaned in the drum remain unchanged, the audio signals in each period are correlated, which can also be called relevant.

[0115] Here, the first rotation period is taken as an example for illustration, and the processing method for the audio signal in each rotation period is the same.

[0116] The rotation period of the drum can be determined by corresponding sensing devices, which can detect each rotation period of the drum, and these devices can detect every time the drum rotates one week. For example, the first audio signal can be divided according to the duration of each period to obtain the first signal sequence of each rotation period.

[0117] Exemplarily, the first signal sequence can be denoted as y(t).

[0118] Taking one of the first signal sequences as an example, after obtaining the first signal sequence, the first signal sequence is segmented to obtain N first target sequences. The segmentation method can be determined according to actual needs. For example, it can be evenly segmented according to the duration, or the first signal sequence can be segmented based on the Hann window to obtain N first target sequences with equal lengths. N is a positive integer. For example, N is 10.

[0119] Exemplarily, the first target sequence can be represented by y(j).

[0120] Each first target sequence can include M sampling points, so the first signal sequence corresponding to each period includes M*N sampling points. M is a positive integer.

[0121] After obtaining the first target sequence, the first frequency peak corresponding to each first target sequence can be determined. The determination method is not limited. For example, perform Fourier transform on the first target sequence to obtain the peak value of the signal frequency in the first target sequence, that is, the first frequency peak. In this way, the first frequency peaks of each first target sequence can be obtained.

[0122] Exemplarily, the first frequency peak represents the quantity of audio signals at each frequency in the first target sequence.

[0123] After determining the first frequency peaks of the N first target sequences, the second frequency peak of the audio signal corresponding to the rotation period where the first target sequence is located, that is, the frequency peak of the first signal sequence, can be determined according to the first frequency peaks of the N first target sequences. For example, the highest first frequency peak among the first frequency peaks of the N first target sequences can be used as the second frequency peak, or the average peak value of the first frequency peaks of the N first target sequences can be used as the second frequency peak. It can also be to determine the second peak through other methods based on the first frequency peaks of the N first target sequences.

[0124] Exemplarily, the second frequency peak can be the quantity value of each frequency determined according to the first frequency peak.

[0125] In this way, the second frequency peaks of the audio signals corresponding to each rotation period, that is, the frequency peaks of each first signal sequence, can be obtained.

[0126] After determining the second frequency peak, the first characteristic information is determined according to the second frequency peak. Since the first audio signal includes first signal sequences corresponding to multiple rotation periods, the first characteristic information can be determined according to the second frequency peaks corresponding to each rotation period. For example, the maximum frequency peak among the second frequency peaks of each rotation period can be used as the first characteristic information, or the average frequency peak of the second frequency peaks of each rotation period can be used as the first characteristic information.

[0127] In this way, the characteristic information of the first audio information, that is, the first characteristic information, is obtained.

[0128] In one embodiment, refer to Figure 7 , which is a schematic diagram for obtaining the second characteristic information. S204, extract the characteristic information of the second audio signal to obtain the second characteristic information, including:

[0129] S2041, extract the second signal sequences of each rotation period of the motor from the second audio signal.

[0130] S2042, segment the second signal sequences to obtain N second target sequences; where each second target sequence includes M sampling points.

[0131] S2043, determine the third frequency peak corresponding to each second target sequence.

[0132] S2044, determine the fourth frequency peak of the rotation period according to the third frequency peak.

[0133] S2045, determine the second characteristic information according to the fourth frequency peak.

[0134] The process of obtaining the second characteristic information can refer to the process of obtaining the first characteristic information, and the principle is the same.

[0135] Since the motor is in a rotating state, the second audio signal includes audio signals corresponding to multiple rotation periods of the motor. After obtaining the second audio signal, extract the second signal sequences of the motor in each rotation period from the second audio signal, starting from the first rotation period, and extract the second signal sequences in each rotation period.

[0136] When the current water level remains unchanged, the current rotation speed remains unchanged, and the objects to be cleaned in the drum remain unchanged, the audio signals in each period are correlated, which can also be called relevant.

[0137] Here, the first rotation period is taken as an example for illustration, and the processing method for the audio signals in each rotation period is the same.

[0138] The rotation period of the motor can be determined by corresponding sensing devices, which can detect each rotation period of the drum, and these devices can detect every time the motor rotates one week. For example, the second audio signal can be divided according to the duration of each period to obtain the second signal sequence of each rotation period.

[0139] Exemplarily, the second signal sequence can be denoted as y1(t).

[0140] Taking one of the second signal sequences as an example, after obtaining the second signal sequence, the second signal sequence is segmented to obtain N second target sequences. The segmentation method can be determined according to actual needs. For example, it can be evenly segmented according to the duration, or the second signal sequence can be segmented based on the Hann window to obtain N second target sequences with equal lengths. N is a positive integer. For example, N is 10.

[0141] Exemplarily, the second target sequence can be represented by y1(j).

[0142] Each second target sequence can include M sampling points, so the second signal sequence corresponding to each period includes M*N sampling points.

[0143] After obtaining the second target sequence, the third frequency peak corresponding to each second target sequence can be determined. The determination method is not limited. For example, perform Fourier transform on the second target sequence to obtain the peak value of the signal frequency in the second target sequence, that is, the third frequency peak. In this way, the third frequency peak of each second target sequence can be obtained.

[0144] Exemplarily, the third frequency peak represents the number of audio signals at each frequency in the second target sequence.

[0145] After determining the third frequency peaks of the N second target sequences, the fourth frequency peak of the audio signal corresponding to the rotation period where the second target sequence is located, that is, the frequency peak of the second signal sequence, can be determined according to the third frequency peaks of the N second target sequences. For example, the highest third frequency peak among the third frequency peaks of the N second target sequences can be used as the fourth frequency peak, or the average peak value of the third frequency peaks of the N second target sequences can be used as the fourth frequency peak. It can also be determined the fourth peak through other methods related to the third frequency peaks of the N second target sequences.

[0146] Exemplarily, the fourth frequency peak can be the quantity value of each frequency determined according to the third frequency peak.

[0147] In this way, the fourth frequency peak of the audio signal corresponding to each rotation period, that is, the frequency peak of each second signal sequence, can be obtained.

[0148] After determining the fourth frequency peak, the second characteristic information is determined according to the fourth frequency peak. Since the second audio signal includes second signal sequences corresponding to multiple rotation periods, the second characteristic information can be determined according to the fourth frequency peaks corresponding to each rotation period. For example, the maximum frequency peak among the fourth frequency peaks can be used as the second characteristic information, or the average frequency peak of the fourth frequency peaks can be used as the second characteristic information.

[0149] In this way, the characteristic information of the second audio information, that is, the second characteristic information, is obtained.

[0150] Reference Figure 8 , is a schematic diagram of a target sequence, Figure 8 shows the waveform schematic diagrams of 7 target sequences. The horizontal axis represents the sampling points, and the vertical axis represents the amplitude. The 7 target sequences can represent the waveform diagrams of the audio signals corresponding to 7 first target sequences, or can represent the waveform diagrams of the audio signals corresponding to 7 second target sequences. The frequencies and amplitudes of the audio signals in each target sequence may be different.

[0151] Reference Figure 9 , is a schematic diagram of a frequency peak, Figure 9 shows the waveform schematic diagrams of 7 frequency peaks. The horizontal axis represents the frequency, and the vertical axis represents the quantity. The 7 target sequences can represent the quantities of the frequencies where the audio signals corresponding to 7 first target sequences are located, or can represent the quantities of the frequencies where the audio signals corresponding to 7 second target sequences are located. Figure 9 The 7 graphs from top to bottom in Figure 8 correspond to the 7 graphs from top to bottom in Figure 9 respectively. For example, Figure 8 the uppermost graph in Figure 9 is determined according to the uppermost graph in Figure 8 The higher the quantity on the vertical axis in each graph in Figure 9 means that the quantity of the frequency where the audio signal corresponding to the first target sequence or the second target sequence shown in Figure 9 is located is more. For example,

[0152] In one embodiment, reference Figure 10 , is a schematic diagram of determining the second target information of a washing machine. The method includes:

[0153] S206, collecting the first original vibration signal at the drum of the washing machine and the second original vibration signal of the motor.

[0154] In S207, filter the first original vibration signal and the second original vibration signal respectively to obtain a first vibration signal corresponding to the first original vibration signal and a second vibration signal corresponding to the second original vibration signal.

[0155] In S208, extract the characteristic information of the first vibration signal to obtain the third characteristic information.

[0156] In S209, extract the characteristic information of the second vibration signal to obtain the fourth characteristic information.

[0157] In S210, determine the third characteristic information as the target vibration information of the drum and determine the fourth characteristic information as the target vibration information of the motor.

[0158] There is no execution sequence relationship between these five steps of S206 to S210 and these five steps of S201 to S205. S206 to S210 can be executed first, or S201 to S205 can be executed first.

[0159] This embodiment is an example of processing the vibration signal of the collected drum and the vibration signal of the collected motor.

[0160] Exemplarily, S208 can be executed first, or S209 can be executed first.

[0161] The first original vibration signal is the vibration signal of the drum collected by the sensor, and the second original vibration signal is the vibration signal of the motor during rotation collected by the sensor. As unprocessed vibration signals, the first original vibration sound signal and the second original vibration signal may include some other noise components.

[0162] Exemplarily, due to the deviation of the washing machine assembly process, placement method and internal counterweight, the first original vibration signal and the second original vibration signal of the drum and the motor during actual operation contain noise components. These noises will affect the detection result and reduce the detection accuracy.

[0163] Exemplarily, the first original vibration signal and the second original vibration signal contain noise components, which may be caused by the dissipation of energy.

[0164] By filtering the first original vibration signal and the second original vibration signal, the noise components in the first original vibration signal and the second original vibration signal can be filtered out, and the obtained vibration signals are closer to the real vibration signal of the drum and the real vibration signal of the motor.

[0165] After filtering the first original vibration signal, a first vibration signal is obtained, and after filtering the second original vibration signal, a second vibration signal is obtained.

[0166] The filtering method can be determined according to actual service requirements, and any method capable of filtering vibration signals is within the protection scope of this embodiment. For example, various filters capable of filtering vibration signals.

[0167] Reference Figure 11 , which is a schematic diagram of an acquired original vibration signal. The horizontal axis is the sampling point, and the vertical axis is the amplitude, that is, the magnitude. Figure 11 What is shown may be the first original vibration signal representing the acquired roller vibration signal, or the second original vibration signal representing the acquired motor vibration signal.

[0168] Reference Figure 12 , which is a waveform schematic diagram of another first original vibration signal or second original vibration signal. The horizontal axis represents the sampling point, and the vertical axis represents the magnitude.

[0169] Taking the first original vibration signal as an example, Figure 12 The amplitudes of the first original vibration signal within the three circles in are relatively large and may be noise components, which can be filtered out through filtering.

[0170] Exemplarily, the first original vibration signal and the second original vibration signal can be filtered multiple times to obtain the first vibration signal and the second vibration signal.

[0171] After obtaining the first vibration signal and the second vibration signal, since there may still be situations such as a large amount of data, mixed signals, and no obvious features in the filtered vibration signals, the characteristic information of the first vibration signal can be extracted to obtain the third characteristic information; the characteristic information of the second vibration signal can be extracted to obtain the fourth characteristic information. The characteristic extraction method is not limited, and any method capable of extracting the characteristics of vibration signals can be used. Including but not limited to: characteristic extraction methods based on energy, which can be methods related to energy characteristic extraction such as short-time energy, energy entropy, short-time autocorrelation coefficient, short-time power spectral density, etc. The detailed characteristic extraction process can refer to the subsequent embodiments.

[0172] After obtaining the third characteristic information and the fourth characteristic information, the third characteristic information is determined as the target vibration information of the roller, and the fourth characteristic information is determined as the target vibration information of the motor.

[0173] Through S206 to S210, the obtained target vibration signal has higher quality and can better represent the vibration of the roller and the vibration of the motor itself, thereby facilitating the improvement of the detection accuracy.

[0174] In one embodiment, reference Figure 13, which is a schematic diagram of filtering the original vibration signal. S207, filter the first original vibration signal and the second original vibration signal respectively to obtain the first vibration signal corresponding to the first original vibration signal and the second vibration signal corresponding to the second original vibration signal, including:

[0175] S2071, determine the maximum value and the minimum value of the first target vibration signal; wherein, the first target vibration signal is the first original vibration signal or the second original vibration signal.

[0176] S2072, determine multiple components of the first target vibration signal according to the maximum value, the minimum value and the first target vibration signal.

[0177] S2073, determine the second target vibration signal according to the components determined in the previous k times; wherein, the second target vibration signal is the first vibration signal or the second vibration signal.

[0178] Wherein, when the first target vibration signal is the first original vibration signal, the second target vibration signal is the first vibration signal; when the first target vibration signal is the second original vibration signal, the second target vibration signal is the second vibration signal. For the sake of convenience of description, here the first original vibration signal or the second original vibration signal is denoted as the first target vibration signal, that is, the first target vibration signal can represent the first original vibration signal or the second original vibration signal.

[0179] After obtaining the first original vibration signal, the maximum value and the minimum value of the amplitude in the first original vibration signal can be determined. After obtaining the second original vibration signal, the maximum value and the minimum value of the amplitude in the second original vibration signal can be determined. The methods capable of determining the maximum value and the minimum value of the amplitude in the first original vibration signal and the second original vibration signal are all within the protection scope of this embodiment.

[0180] Both the maximum value and the minimum value of the amplitude in the first original vibration signal can be multiple, and both the maximum value and the minimum value of the amplitude in the second original vibration signal can be multiple.

[0181] For example, the maximum value and the minimum value of the amplitude in the first original vibration signal can be determined based on the derivative and the gradient, and the maximum value and the minimum value of the amplitude in the second original vibration signal can be determined based on the derivative and the gradient.

[0182] After determining the maximum value and the minimum value of the first target vibration signal, determine multiple components of the first target vibration signal according to the maximum value, the minimum value and the first target vibration signal. The component here can be the intrinsic mode function IMF.

[0183] S2072, the process of determining the component of the first target vibration signal can be based on the intrinsic mode decomposition process, including:

[0184] Spline fitting is used for each maximum value of the first target vibration signal to obtain a maximum value signal. Spline fitting is used for each minimum value of the first target vibration signal to obtain a minimum value signal. A mean signal is determined based on the maximum value signal and the minimum value signal. The component of the first target vibration signal is determined based on the first target vibration signal and the mean signal.

[0185] Exemplarily, to determine the mean signal based on the maximum value signal and the minimum value signal, the average signal of the maximum value signal and the minimum value signal can be determined as the mean signal.

[0186] Exemplarily, the maximum value signal corresponds to a maximum envelope, and the minimum value signal corresponds to a minimum envelope. To determine the mean envelope based on the maximum envelope and the minimum envelope, the average envelope of the maximum envelope and the minimum envelope can be determined as the mean envelope. The signal formed by the mean envelope is the mean signal.

[0187] Exemplarily, the mean value in the mean signal can also be determined by the following formula (2):

[0188]

[0189] where envmaxp represents the maximum value and envminp represents the minimum value.

[0190] After determining the mean signal, a third vibration signal is determined based on the first target vibration signal and the mean signal. The difference signal obtained by subtracting the mean signal from the first target vibration signal can be used as the third vibration signal.

[0191] If the third vibration signal meets the following preset conditions, then this third vibration signal is used as the component determined for the first time.

[0192] The preset conditions include: within the entire time range, the number of local extreme points and the number of zero-crossing points must be equal or differ by at most one. And, at any moment, the average value of the upper envelope formed by local maximum points and the lower envelope formed by local minimum points is zero, that is, the upper envelope and the lower envelope are locally symmetric with respect to the time axis.

[0193] If the third vibration signal does not meet the preset conditions, then the above-mentioned intrinsic mode decomposition is performed on the third vibration signal, and the above process is repeated until a vibration signal that meets the above preset conditions is obtained. The vibration signal that meets the above preset conditions is used as the component determined for the first time.

[0194] After determining the component determined for the first time, the signal obtained by subtracting the component determined for the first time from the first target vibration signal can be used as the decomposition object, and the above-mentioned intrinsic mode decomposition is performed to determine the component determined for the second time.

[0195] By repeating this process, the first K determined components can be obtained. The first determined component can be denoted as IMF1, the second determined component can be denoted as IMF2... and the k-th determined component can be denoted as IMFk.

[0196] After obtaining the first k determined components, the second target vibration signal can be determined based on the first k determined components. For example, the first k determined components can be superimposed, and the superimposed signal can be used as the second target vibration signal.

[0197] When the first target vibration signal is the first original vibration signal, the second target vibration signal is the first original vibration signal. When the first target vibration signal is the second original vibration signal, the second target vibration signal is the second vibration signal. That is, the first original vibration signal and the second original vibration signal can be filtered through the above filtering method.

[0198] Reference Figure 14 is a schematic diagram for determining the components of the first original vibration signal or the second original vibration signal.

[0199] Figure 14 The input signal x(t) in [[ ]] is the first target vibration signal, which can be the first original vibration signal or the second original vibration signal. Determine all the maximum and minimum points in x(t), and use cubic spline interpolation to fit the maximum envelope line envmaxp and the minimum envelope line envminp. The mean envelope line envop determined based on the maximum envelope line envmaxp and the minimum envelope line envminp can be determined by the above formula (2). After determining the mean envelope line, determine whether the difference signal m(t) obtained by subtracting the mean envelope line envop from x(t) satisfies the above preset conditions. The preset conditions can include: IMF = m(t); r(t) = x(t) - m(t), and this adjustment can also be the IMF condition.

[0200] If m(t) satisfies the preset conditions and r(t) is a monotonic function or less than the preset value, then r(t) is determined as the component. If m(t) does not satisfy the preset conditions, then replace x(t) with m(t) and re-execute the Figure 14 process shown. If m(t) satisfies the preset conditions and r(t) is not a monotonic function or not less than the preset value, then replace x(t) with r(t) and re-execute the Figure 14 process shown.

[0201] Reference Figure 15 is a schematic diagram of the components of the first target vibration signal, which can be a schematic diagram of the components of the first original vibration signal or a schematic diagram of the components of the second original vibration signal. Figure 15 shows a schematic diagram of the components obtained in the first 6 times, that is, a waveform signal schematic diagram of IMF1 to IMF6.Figure 16 As the frequency domain schematic diagram corresponding to Figure 15 , it shows the distribution of frequencies in each component signal.

[0202] Referring to FIG. 17(a), it is a schematic diagram of a first target vibration signal, that is, a schematic diagram of an unfiltered signal waveform. The horizontal axis is the sampling point, and the vertical axis is the amplitude. Referring to FIG. 17(b), it is a spectrogram of the first target vibration signal, that is, a spectrogram of the unfiltered signal. The horizontal axis is the frequency, and the vertical axis is the amplitude, representing the quantity of each frequency.

[0203] Referring to FIG. 18(a), it is a schematic diagram of the waveform of the first target vibration signal obtained after filtering the first target vibration signal. The horizontal axis is the sampling point, and the vertical axis is the amplitude. Referring to FIG. 18(b), it is a spectrogram of the second target vibration signal obtained after filtering the first target vibration signal. The horizontal axis is the frequency, and the vertical axis is the amplitude, representing the quantity of each frequency.

[0204] In one embodiment, referring to Figure 19 , it is a schematic diagram of extracting third feature information. S208, extracting the feature information of the first vibration signal to obtain third feature information includes:

[0205] S2081, extracting the third signal sequence of each rotation period of the drum from the first vibration signal.

[0206] S2082, segmenting the third signal sequence to obtain F third target sequences; wherein, each third target sequence includes H sampling points.

[0207] S2083, determining the first energy of each third target sequence, the first total energy of the F third target sequences, and the first proportion of each first energy in the first total energy.

[0208] S2084, determining the first energy entropy of each rotation period according to the first energy, the first total energy, and the first proportion.

[0209] S2085, determining the third feature information according to the first energy entropy.

[0210] Since the drum is in a rotating state, the first vibration signal includes vibration signals corresponding to multiple rotation periods of the drum. After obtaining the first vibration signal, the third signal sequence of the drum in each rotation period is extracted from the first vibration signal, starting from the first rotation period, and the third signal sequence in each rotation period is extracted. One rotation of the drum corresponds to one third signal sequence.

[0211] Under the condition that the current water level remains unchanged, the current rotation speed remains unchanged, and the objects to be cleaned in the drum remain unchanged, the vibration signals in each period are correlated, which can also be called correlation.

[0212] Here, the first rotation period is taken as an example for illustration, and the processing method for the vibration signals within each rotation period is the same.

[0213] The rotation period of the drum can be determined by corresponding sensing devices, which can detect each rotation period of the drum, and these devices can detect every time the drum rotates one week. For example, the first vibration signal can be divided according to the duration of each period to obtain the third signal sequence of each rotation period.

[0214] Exemplarily, the third signal sequence can be denoted as z(t).

[0215] Taking one of the third signal sequences as an example, after obtaining the third signal sequence, the third signal sequence is segmented to obtain F third target sequences. The segmentation method can be determined according to actual requirements. For example, it can be evenly segmented according to the duration, or the third signal sequence can be segmented based on a Hamming window to obtain F third target sequences with equal lengths. F is a positive integer. For example, F is 16.

[0216] Exemplarily, the third target sequence can be represented by z(j). j represents the number of third target sequences.

[0217] Each first target sequence may include H sampling points, then the third signal sequence corresponding to each period includes H*F sampling points, that is, the third signal sequence for one rotation of the drum includes H*F sampling points. H is a positive integer.

[0218] Reference Figure 20 shows a waveform schematic diagram of a third target sequence. Figure 20 shows a waveform schematic diagram of 16 consecutive third target sequences when F is equal to 16. Figure 20 One third target sequence in can also be represented as one frame. Figure 20 In, the abscissa of the waveform schematic diagram of each third target sequence is frequency, and the ordinate is amplitude.

[0219] After obtaining the third target sequence, determine the first energy of each third target sequence, the first total energy of the F third target sequences, and the first proportion of each first energy in the first total energy.

[0220] Exemplarily, the first energy, the first total energy, and the first proportion can be determined by the following formula (3),

[0221]

[0222] Among them, E j represents the first energy, E represents the first total energy, and p j represents the first proportion.

[0223] Of course, the first energy, the first total energy, and the first ratio can also be determined by other means.

[0224] After determining the first energy, the first total energy, and the first ratio, the first energy entropy of each rotation period is determined according to the first energy, the first total energy, and the first ratio.

[0225] The first energy entropy can be determined by the following formula (4):

[0226]

[0227] Of course, the first energy entropy can also be determined by other means.

[0228] The obtained first energy entropy is the first energy entropy of one rotation period. By the above method, the first energy entropy of the third signal sequence corresponding to each rotation period can be determined.

[0229] After determining the first energy entropy of the third signal sequence corresponding to each rotation period, the third characteristic information can be determined according to the first energy entropy of the third signal sequence corresponding to each rotation period. For example, the matrix of the first energy entropy of the third signal sequence corresponding to each rotation period can be used as the third characteristic information. The maximum value of the first energy entropy of the third signal sequence corresponding to each rotation period can also be used as the third characteristic information, or the average value of the first energy entropy of the third signal sequence corresponding to each rotation period can be used as the third characteristic information.

[0230] Of course, the third characteristic information can also be determined by other means according to the first energy entropy.

[0231] In another embodiment, referring to Figure 21 , which is a schematic diagram for determining the fourth characteristic information, S209, extracting the characteristic information of the second vibration signal to obtain the fourth characteristic information includes:

[0232] S2091, extracting the fourth signal sequence of each rotation period of the motor from the second vibration signal.

[0233] S2092, segmenting the fourth signal sequence to obtain F fourth target sequences; where the fourth target sequence includes H sampling points.

[0234] S2093, determining the second energy of each fourth target sequence, the second total energy of the F fourth target sequences, and the second ratio of each second energy in the second total energy.

[0235] S2094, determining the second energy entropy of each rotation period according to the second energy, the second total energy, and the second ratio.

[0236] S2095, determine the fourth characteristic information according to the second energy entropy.

[0237] Since the motor is in a rotating state, the second vibration signal includes vibration signals corresponding to multiple rotation cycles of the motor. After obtaining the second vibration signal, extract the fourth signal sequence within each rotation cycle of the motor from the second vibration signal, starting from the first rotation cycle, and extract the fourth signal sequence within each rotation cycle. One rotation of the motor corresponds to one fourth signal sequence.

[0238] When the current water level remains unchanged, the current rotational speed remains unchanged, and the objects to be cleaned in the drum remain unchanged, the vibration signals within each cycle are correlated, which can also be called relevant.

[0239] Here, the first rotation cycle is taken as an example for illustration, and the processing method for the vibration signals within each rotation cycle is the same.

[0240] The rotation cycle of the motor can be determined by corresponding sensing devices, which can detect each rotation cycle of the motor, and these devices can detect every time the motor rotates one week. For example, the second vibration signal can be divided according to the duration of each cycle to obtain the fourth signal sequence of each rotation cycle.

[0241] Exemplarily, the fourth signal sequence can be denoted as s(t).

[0242] Taking one of the fourth signal sequences as an example, after obtaining the fourth signal sequence, divide the fourth signal sequence to obtain F third target sequences. The division method can be determined according to actual needs. For example, it can be evenly divided according to the duration, or the fourth signal sequence can be divided based on a Hamming window to obtain F fourth target sequences with equal lengths. F is a positive integer. For example, F is 16.

[0243] Exemplarily, the fourth target sequence can be represented by s(j). j represents the number of fourth target sequences. Of course, in order to distinguish the number of third target sequences, w or other letters can also be used to replace j.

[0244] Each fourth target sequence can include H sampling points, so the fourth signal sequence corresponding to each cycle includes H*F sampling points, that is, the fourth signal sequence for one rotation of the drum includes H*F sampling points. H is a positive integer.

[0245] After obtaining the fourth target sequence, determine the first energy of each fourth target sequence, the first total energy of the F fourth target sequences, and the first proportion of each first energy in the first total energy.

[0246] Exemplarily, the second energy, the second total energy, and the second proportion can be determined by the image formula (5).

[0247]

[0248] Among them, E j represents the second energy, E represents the second total energy, and p j represents the second proportion.

[0249] Of course, the second energy, the second total energy, and the second proportion can also be determined by other means.

[0250] After determining the second energy, the second total energy, and the second proportion, determine the second energy entropy of each rotation period according to the second energy, the second total energy, and the second proportion.

[0251] The second energy entropy can be determined by the following formula (6):

[0252]

[0253] Of course, the second energy entropy can also be determined by other means.

[0254] Exemplarily, j in formula (5) and formula (6) can be replaced by w or other letters.

[0255] The obtained second energy entropy is the second energy entropy of one rotation period. The second energy entropy of the fourth signal sequence corresponding to each rotation period can be determined by the above method.

[0256] After determining the second energy entropy of the fourth signal sequence corresponding to each rotation period, the fourth characteristic information can be determined according to the second energy entropy of the fourth signal sequence corresponding to each rotation period. For example, the matrix of the second energy entropy of the fourth signal sequence corresponding to each rotation period can be used as the fourth characteristic information. The maximum value of the second energy entropy of the fourth signal sequence corresponding to each rotation period can also be used as the fourth characteristic information, or the average value of the second energy entropy of the fourth signal sequence corresponding to each rotation period can be used as the fourth characteristic information.

[0257] Of course, the fourth characteristic information can also be determined by other means according to the second energy entropy.

[0258] In another embodiment, S300, detect whether the operating state of the washing machine is abnormal, including:

[0259] Use the target detection model to determine whether the operating state of the washing machine is abnormal according to the current water level, the current rotation speed, the target audio information at the drum, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor.

[0260] This embodiment shows a method for determining whether the operating state of a washing machine is abnormal after determining the current water level, current rotation speed, target audio information at the drum, target vibration information of the drum, target audio information of the motor, and target vibration information of the motor. The detection of whether the operating state of the washing machine is abnormal is achieved through a network model.

[0261] The target detection model is a model obtained by training an initial network model using a training sample set through machine learning.

[0262] The training sample set includes: a positive sample set and a negative sample set.

[0263] The positive sample set includes multiple different positive samples. Each positive sample includes: a water level sample, a positive motor rotation speed sample, and sample information of normal sound at the drum in the washing machine, normal vibration sample information of the drum, normal sound sample information of the motor, and normal vibration sample information of the motor under the water level sample and the positive motor rotation speed sample; each positive sample has its own first label, which is used to identify that the output of the positive sample corresponding to the initial network model is normal.

[0264] The water level body of the water level sample of the body has different rotation speeds for different positive motor rotation speed samples. The positive samples include sample information of normal sound at the drum in the washing machine, normal vibration sample information of the drum, normal sound sample information of the motor, and normal vibration sample information of the motor under various water level samples and / or motor rotation speed samples. The water level information of the water level sample can be in the form of a water level grade, that is, the current water level corresponds to its own water level grade, and each water level grade corresponds to a water level range. If the current water level is in which water level range, it corresponds to the corresponding water level grade. The water level grades can include high water level, medium water level, low water level, and zero water level.

[0265] Similarly, the rotation speed information of the motor rotation speed sample can also correspond to a rotation speed grade. Each rotation speed grade corresponds to a rotation speed range. If the current rotation speed is in which rotation speed range, it corresponds to the corresponding rotation speed grade. The rotation speed grades can include high rotation speed, medium rotation speed, low rotation speed, and zero rotation speed.

[0266] The negative sample set includes multiple negative samples. Each negative sample includes: a water level sample, a negative motor rotation speed sample, and negative sample audio information at the drum in the washing machine, negative sample vibration information of the drum, negative sample audio information of the motor, and negative sample vibration information of the motor under the water level sample and the negative motor rotation speed sample. At least one of the negative sample audio information at the drum in the washing machine, negative sample vibration information of the drum, negative sample audio information of the motor, and negative sample vibration information of the motor is abnormal information; each negative sample has its own second label, which is used to identify that the output of the negative sample corresponding to the initial network model is abnormal.

[0267] The water level body of water level samples is different, and the rotational speeds of negative samples with different motor speeds are different. The negative samples include sample information of abnormal sounds at the drum in the washing machine, sample information of abnormal vibrations of the drum, sample information of abnormal sounds of the motor, and / or sample information of abnormal vibrations of the motor under various water level samples and / or negative samples of motor speeds. The water level information of the water level samples can be in the form of water level grades, that is, the current water level corresponds to its respective water level grade, and each water level grade corresponds to a water level range. If the current water level is within a certain water level range, it corresponds to the corresponding water level grade. The water level grades can include high water level, medium water level, low water level, and zero water level.

[0268] Similarly, the rotational speed information of the motor speed samples can also correspond to rotational speed grades, and each rotational speed grade corresponds to a rotational speed range. If the current rotational speed is within a certain rotational speed range, it corresponds to the corresponding rotational speed grade. The rotational speed grades can include high rotational speed, medium rotational speed, low rotational speed, and zero rotational speed.

[0269] The negative samples include various abnormal samples, which can be samples with abnormal vibration signals and samples with abnormal audio signals.

[0270] Exemplarily, the vibration signals and audio signals in the positive samples and negative samples are also signals processed by the methods corresponding to the embodiments of S200 and the methods corresponding to the embodiments of various methods included in S200.

[0271] In one embodiment, referring to Figure 22 , it is a schematic diagram of an abnormal detection device, and the device includes:

[0272] An acquisition module 1, configured to acquire first target information of the washing machine, where the first target information includes: the current rotational speed and the current water level of the motor in the washing machine;

[0273] A determination module 2, configured to determine second target information of the washing machine at the current water level and the current rotational speed; where the second target information at least includes target audio information at the drum in the washing machine, target vibration information of the drum, target audio information of the motor, and target vibration information of the motor;

[0274] A detection module 3, configured to detect whether the operating state of the washing machine is abnormal according to the first target information and the second target information.

[0275] In one embodiment, the determination module 2 includes:

[0276] A first acquisition sub-module, configured to acquire a first original audio signal at the drum in the washing machine and a second original audio signal of the motor;

[0277] The first filtering sub-module is configured to filter the first original audio signal and the second original audio signal respectively to obtain a first audio signal corresponding to the first original audio signal and a second audio signal corresponding to the second original audio signal;

[0278] The first extraction sub-module is configured to extract the feature information of the first audio signal to obtain first feature information;

[0279] The second extraction sub-module is configured to extract the feature information of the second audio signal to obtain second feature information;

[0280] The first target signal determination sub-module is configured to determine the first feature information as the target audio information at the drum and the second feature information as the target audio information of the motor.

[0281] In one embodiment, the first extraction sub-module includes:

[0282] The first extraction unit is configured to extract a first signal sequence of each rotation period of the drum from the first audio signal;

[0283] The first segmentation unit is configured to segment the first signal sequence to obtain N first target sequences; wherein, each first target sequence includes M sampling points;

[0284] The first frequency peak determination unit is configured to determine a first frequency peak corresponding to each of the first target sequences;

[0285] The second frequency peak determination unit is configured to determine a second frequency peak of the rotation period according to the first frequency peak;

[0286] The first feature information determination unit is configured to determine the first feature information according to the second frequency peak.

[0287] In one embodiment, the second extraction sub-module includes:

[0288] The second extraction unit is configured to extract a second signal sequence of each rotation period of the motor from the second audio signal;

[0289] The second segmentation unit is configured to segment the second signal sequence to obtain N second target sequences; wherein, each second target sequence includes M sampling points;

[0290] The third frequency peak determination unit is configured to determine a third frequency peak corresponding to each of the second target sequences;

[0291] The fourth frequency peak determination unit is configured to determine a fourth frequency peak of the rotation period according to the third frequency peak;

[0292] The second feature information determining unit is configured to determine the second feature information according to the fourth frequency peak.

[0293] In one embodiment, the determining module 2 includes:

[0294] The second acquisition sub-module is configured to acquire a first original vibration signal at a drum in the washing machine and a second original vibration signal of a motor;

[0295] The second filtering sub-module is configured to filter the first original vibration signal and the second original vibration signal respectively to obtain a first vibration signal corresponding to the first original vibration signal and a second vibration signal corresponding to the second original vibration signal;

[0296] The third extraction sub-module is configured to extract feature information of the first vibration signal to obtain third feature information;

[0297] The fourth extraction sub-module is configured to extract feature information of the second vibration signal to obtain fourth feature information;

[0298] The second target signal determining sub-module is configured to determine the third feature information as the target vibration information of the drum and the fourth feature information as the target vibration information of the motor.

[0299] In one embodiment, the second filtering sub-module includes:

[0300] The extreme value determining unit is configured to determine a maximum value and a minimum value of a first target vibration signal; wherein, the first target vibration signal is the first original vibration signal or the second original vibration signal;

[0301] The component determining unit is configured to determine a plurality of components of the first target vibration signal according to the maximum value, the minimum value and the first target vibration signal;

[0302] The second target vibration signal determining unit is configured to determine a second target vibration signal according to the components determined in the previous k times; wherein, the second target vibration signal is the first vibration signal or the second vibration signal;

[0303] Wherein, when the first target vibration signal is the first original vibration signal, the second target vibration signal is the first vibration signal; when the first target vibration signal is the second original vibration signal, the second target vibration signal is the second vibration signal.

[0304] In one embodiment, the third extraction sub-module includes:

[0305] A third extraction unit, configured to extract a third signal sequence of each rotation period of the drum from the first vibration signal;

[0306] A third segmentation unit, configured to segment the third signal sequence to obtain F third target sequences; wherein, each third target sequence includes H sampling points;

[0307] A third determination unit, configured to determine a first energy of each third target sequence, a first total energy of the F third target sequences, and a first proportion of each first energy in the first total energy;

[0308] A first energy entropy determination unit, configured to determine a first energy entropy of each rotation period according to the first energy, the first total energy, and the first proportion;

[0309] A third feature information determination unit, configured to determine the third feature information according to the first energy entropy.

[0310] In one embodiment, a fourth extraction sub-module includes:

[0311] A fourth extraction unit, configured to extract a fourth signal sequence of each rotation period of the motor from the second vibration signal;

[0312] A fourth segmentation unit, configured to segment the fourth signal sequence to obtain F fourth target sequences; wherein, each fourth target sequence includes H sampling points;

[0313] A fourth determination unit, configured to determine a second energy of each fourth target sequence, a second total energy of the F fourth target sequences, and a second proportion of each second energy in the second total energy;

[0314] A second energy entropy determination unit, configured to determine a second energy entropy of each rotation period according to the second energy, the second total energy, and the second proportion;

[0315] A fourth feature information determination unit, configured to determine the fourth feature information according to the second energy entropy.

[0316] In one embodiment, the detection module is configured to:

[0317] Use a target detection model to determine whether the operating state of the washing machine is abnormal according to the target audio information at the drum, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor.

[0318] In one embodiment, the target detection model is a model obtained by training an initial network model using a training sample set by means of machine learning;

[0319] The training sample set includes:

[0320] A positive sample set, including a plurality of different positive samples. Each positive sample includes: a water level sample, a positive motor speed sample, and sample information of normal sound at the drum of the washing machine, sample information of normal vibration of the drum, sample information of normal sound of the motor, and sample information of normal vibration of the motor under the water level sample and the positive motor speed sample; each of the positive samples has its own first label, which is used to identify that the output of the positive sample corresponding to the initial network model is normal;

[0321] A negative sample set, including a plurality of negative samples. Each negative sample includes: the water level sample, a negative motor speed sample, and negative sample audio information at the drum of the washing machine, negative sample vibration information of the drum, negative sample audio information of the motor, and negative sample vibration information of the motor under the water level sample and the negative motor speed sample; wherein, at least one of the negative sample audio information at the drum of the washing machine, the negative sample vibration information of the drum, the negative sample audio information of the motor, and the negative sample vibration information of the motor is abnormal information; each of the negative samples has its own second label, which is used to identify that the output of the negative sample corresponding to the initial network model is abnormal.

[0322] In another embodiment, a washing machine is further provided, including:

[0323] A processor and a memory for storing executable instructions that can run on the processor, wherein:

[0324] When the processor is used to run the executable instructions, the executable instructions execute the method described in any of the above embodiments.

[0325] In another embodiment, a non - temporary computer - readable storage medium is further provided. Computer - executable instructions are stored in the computer - readable storage medium, and when the computer - executable instructions are executed by a processor, the method described in any of the above embodiments is implemented.

[0326] In another embodiment, an implementation manner of an anomaly detection method is further provided.

[0327] When detecting whether the washing machine is abnormal, the back electromotive force of the washing machine motor is used to inspect and detect the attitude of the drum of the motor - fault washing machine and the fault state of the washing machine, and the fault result can be reported to the cloud, which can greatly assist the manufacturer in analyzing faults, collecting information, and summarizing problems.

[0328] The motor back electromotive force is easily affected by factors such as the number of motor poles and motor structure, and it is easy to break through the detection board during the commissioning stage. Simple motor noise detection is easily affected by different water levels and different laundry weights in the drum, and does not meet the technical requirements for actual use.

[0329] The washing machine fault detection scheme based on motor sound does not take into account the pressure effect of water level and clothing weight on motor rotation, nor does it take into account the change of motor sound when the washing machine is tilted, which makes the inspection method not completely effective.

[0330] A washing machine fault detection solution based on vibration amplitude (jitter, displacement size). Due to the different structures of drum washing machines and pulsator washing machines, the friction effect of water flow force and lifting ribs is different, and the distribution of clothes is also different, the vibration amplitude of the barrel will deviate.

[0331] Taking a drum washing machine as an example, a sound wave detector and a displacement detector are installed on the drum shell to detect the sound and vibration information of the drum. At the same time, the same pair of sensors are also installed at the end of the BLDC brushless DC motor, which can be used to detect the sound and vibration information of the brushless DC motor.

[0332] This embodiment will fully take into account the dynamic characteristics of the washing machine, and record and calibrate the rotation sound and vibration displacement of the washing machine motor in the waterless state (original state), low water state, medium water state and high water state. In actual operation, the washing machine main control board will match the washing machine motor data obtained by real-time measurement with the standard data calibrated in advance, and analyze and feedback on abnormal situations.

[0333] This embodiment not only analyzes the sound and vibration of the tail motor, but also records the sound characteristics of the water fluid of the drum itself and the displacement of the front end of the drum.

[0334] For example, if a pair of shoes is stuck in a washing machine, the shoes cannot rotate and roll with the water flow. Although the motor rotates normally, the sound of the water flow and the fluid displacement are definitely different from the characteristic data of the natural clothes following the mainstream state. In addition, due to the greater resistance of water, the operating parameters of the motor are definitely different from the normal water resistance state.

[0335] In general, a washing machine is (cause) the rotation of the motor (structure) driving (result) the rotation of the water flow (performance). The current methods only focus on the "cause", but do not combine the presentation of the "cause" and the presentation of the "result" for real-time analysis. This patent aims to optimize this problem and improve the dynamic performance analysis capabilities of the washing machine.

[0336] 1. Data collection and integration part:

[0337] The following takes the data collection under different washing machine water levels and rotation speeds as an example:

[0338] 1) On the premise of low water level - low rotation speed of the washing machine, use the sound sensor and vibration sensor at the bottom of the washing machine to collect the physical information data in this state respectively. 2) Save the first original vibration signal of the D motor - LW - LV, which represents low rotation speed - low water level - the bottom motor of the washing machine. 3) Save the first original audio signal of the B motor - LW - LV, which represents low rotation speed - low water level - the bottom motor of the washing machine. 4) Save the second original vibration signal of the D drum - LW - LV, which represents low rotation speed - low water level - the top drum of the washing machine. 5) Save the second original audio signal of the B drum - LW - LV, which represents low rotation speed - low water level - the top drum of the washing machine.

[0339] The data collection process is carried out by using the ADC conversion circuit - sensor - single - chip microcomputer - embedded board, and it can also be other methods. Audio signals and vibration signals at different water levels and rotation speeds can be collected.

[0340] 6) The acquisition results completed by the data sensor should include data combinations of four water levels and four rotation speeds. Repeat steps 2) - 5), and it is necessary to systematically collect data under different water levels and rotation speeds, including:

[0341] ① The sound of the washing machine motor - noise B 电机 ② The sound of the washing machine drum (water fluid) - noise B 滚筒

[0342] ③ The vibration of the washing machine motor - displacement D 电机 ④ The vibration of the washing machine drum (subject to friction) - displacement D 滚筒 .

[0343] 7) The above data are all collected during the factory calibration stage of the washing machine, and during the collection process, it is necessary to select appropriate "water level parameters" and "motor rotation speed" gradient distributions according to the washing machine capacity and motor characteristics.

[0344] 8) More than 10,000 original single - data need to be collected in a single group, and in necessary cases, more than one group is collected. When performing specific data analysis and modeling, a part of these data should be used, and the remaining unselected data are used as test samples to verify the performance of the model.

[0345] In one embodiment, the first original audio signal collected is as follows:

[0346]

[0347] It can be converted into the following form:

[0348]

[0349] The first original vibration signal collected is as follows:

[0350]

[0351] It can be converted into the following form:

[0352]

[0353] The second original audio signal collected is as follows:

[0354]

[0355] It can be converted into the following form:

[0356]

[0357] The second original vibration signal collected is as follows:

[0358]

[0359] It can be converted into the following form:

[0360]

[0361] The information collected in this embodiment can be used as a training sample.

[0362] II. Filtering of sound signals:

[0363] Since noise and clutter are inevitably accompanied during the rotation of the motor and the water bucket, the original sound data cannot be directly used and needs to be filtered and feature-extracted.

[0364] 9) Perform primary filtering on the sound signal to eliminate the influence of white noise and low-frequency clutter.

[0365] 10) The "water fluid sound" at the high water level - high rotation speed of the drum can be filtered by formula (1).

[0366] 11) Continuously repeat step 10) using the iterative method. Eventually, the sound signal result with low-frequency noise and noise eliminated can be obtained. And the filtered sound signals, including the drum sound signal B 滚筒 and the motor sound signal B 电机 , after high-pass filtering, B 滚筒-滤波 and B 电机-滤波 can be obtained respectively.

[0367] III. Filtering of vibration signals:

[0368] The vibration displacement results during operation contain noise components, which are largely caused by the dissipation of energy. This patent uses the EMD (Empirical Mode) filtering process method to extract the valid data results of D 滚筒 and D 电机 . Taking the "drum displacement" at high water level and high rotational speed of the drum, D 电机-HW-HV as an example.

[0369] 12) Using the method of derivative gradient, calculate all the maximum values of the signal D 电机-HW-HV , save them as envmaxp, and the minimum values, save them as envminp. Then update the value of D 电机-HW-HV using the value of envop.

[0370] Reference can be made to formula (2), D 电机-HW-HV-IMF is equal to D 电机-HW-HV minus envop.

[0371] 13) Check whether the updated D 电机-HW-HV signal meets the conditions. Within the entire data range, the number of zero-crossing points and local extreme points differ by at most 1, and at any time, the average value of the local minimum envelope and the local maximum envelope is approximately 0 (i.e., up and down symmetry).

[0372] If the conditions are met, then the current D 电机-HW-HV-IMF is considered an IMF component (Intrinsic Mode Function) of the original data D 电机-HW-HV . If the conditions are not met, then continue to repeat step 12) for the IMF decomposition of D 电机-HW-HV-IMF data.

[0373] 14) Repeat step 12) and step 13). The original data D 电机-HW-HV , through continuous decomposition, can obtain D 电机-HW-HV-IMF-1 , D 电机-HW-HV-IMF-2 , D 电机-HW-HV-IMF-3 , D 电机-HW-HV-IMF-4 , D 电机-HW-HV-IMF-5 , etc.

[0374] Add up the top five IMF1 - IMF5 components mentioned above. Finally, the synthetic, filtered, and feature - centralized vibration signal result D 电机-HW-HV-IMF-滤波 can be obtained.

[0375] Repeat the above steps 12 - 14). Similarly, finally, the signals D 滚筒-滤波 and D 电机-滤波 can be obtained.

[0376] So far, the filtering and extraction of the original data of the drum sound signal, drum vibration signal, motor sound signal, and motor vibration signal have all been completed.

[0377] IV. Feature extraction of vibration displacement signals:

[0378] The vibration matrix signal D after the above EMI filtering 滚筒-滤波 and D 电机-滤波 still have problems such as a large amount of data, mixed signals, and no obvious features. Therefore, on the basis of filtering, the subsequent process still needs to continue to extract the vibration displacement features. Taking D 电机-滤波-滤波 as an example. Here, methods related to energy feature extraction such as short-time energy, energy entropy, short-time autocorrelation coefficient, and short-time power spectral density can be used, as long as they are technical methods for energy.

[0379] 17) Extract the vibration displacement signal sequence of the first revolution of the motor from the signal D 电机-滤波-滤波 and record it as x(t).

[0380] (The original data D 电机-滤波-滤波 may contain data results of multiple revolutions of the motor. Considering that under the premise of constant load, constant speed, and constant water level, the data results of each revolution should be correlated. Here, only the data of the first revolution is taken as an example)

[0381] 18) Using the Hamming window as the segmentation method, divide the signal x(t) into 16 sequences of equal length {x(1)x(2)…x(i - 1)x(i)}, and each sequence contains N points. (This also means that the signal sequence x(t) of the first revolution of the motor contains a total of 16×N points).

[0382] ) Taking any segment signal sequence x(i) of the first period x(t) as an example, calculate the energy Ei of its sequence frame, the total energy E, and the energy proportion Pi of each frame respectively. Finally, obtain the short-time energy entropy feature EN1 of x(t).

[0383] Referring to the above method, the vibration displacement signal sequence of the second revolution of the motor can also be extracted from the signal D 电机-滤波-滤波 and recorded as x2(t), and the corresponding short-time energy entropy feature EN2 is obtained.

[0384] 21) Repeatedly perform the above steps to obtain the entropy energy feature results EN1, EN2…ENn of multiple independent periods of D 电机-滤波-滤波 .

[0385] They can be saved as EN 电机-HW-HV＝[EN1、EN2…ENn] .

[0386] 22) The above D after filtering 滚筒-滤波 and D 电机-滤波 can obtain EN 滚筒 and EN 电机This data will be used for further data modeling and analysis.

[0387] The information is as follows:

[0388]

[0389]

[0390]

[0391] 5. Feature extraction of sound signals:

[0392] The above sound signal matrix B after high-pass filtering 滚筒-滤波 and B 电机-滤波 There are still problems such as insufficient orthogonality, insufficient smoothness, and lack of prominent feature information. Unlike vibration displacement signals, which start from the energy perspective, sound signals should start from the frequency perspective. (Here, DFT Fourier transform, FFT fast Fourier transform, WT wavelet transform, STFT short-time Fourier transform, etc. can be used, as long as they are feature extraction methods for frequency). This patent takes FFT transform as an example.

[0393] 23) From signal B 电机-HW-HV-滤波 The vibration displacement signal sequence of the first rotation of the motor is extracted and recorded as y(t). 电机-HW-HV-滤波 The data results of multiple cycles of motor rotation can be included. Considering that the load, speed and water level remain unchanged, the data results of each week should be relevant. Here we only take the data of the first week as an example.

[0394] 24) Using the Henning window as the segmentation method, the signal y(t) is divided into 10 sequences of equal length {y(1)y(2)…y(j-1)y(j)}, each of which contains M points. This also means that the signal sequence y(t) of the first rotation of the motor contains a total of 10×M points.

[0395] 25) Take any signal sequence y(j) of the first period y(t) as an example, perform Fourier transform on it and obtain the location of the frequency peak.

[0396] 26) Repeat the above steps to get B 电机-HW-HV-滤波 The FFT main frequency (the frequency corresponding to the highest peak) results of multiple independent cycles are Fre1, Fre2, Fre3...Fren. They can be saved as Fre 电机-HW-HV =[Fre1, Fre2, Fre3...Fren].

[0397] 27) The above filtered B 滚筒-滤波 and B 电机-滤波Fre can be obtained after corresponding processing 滚筒 and Fre 电机 . This data will be used for the next data modeling and analysis.

[0398] The information is as follows:

[0399]

[0400]

[0401]

[0402]

[0403] 28) The motor sound data, motor displacement data, and drum (water fluid) sound data corresponding to different rotational speeds and different water levels (clothing load weights) collected above are used as a data set. Based on different washing machine abnormal conditions as the classification basis, through neural networks such as SVM, CNN, and DSC for training, and using the minimum gradient descent algorithm, the characteristic recognition and analysis model of the motor sound displacement and drum sound displacement regarding different washing machine faults under the conditions of this washing machine model can be obtained. The input data of this model includes EN 滚筒 、EN 电机 、Fre 滚筒 and Fre 电机 , and the output is a multi-classification model containing all common washing machine motor fault types, and the sub-item with the highest probability is the judgment result of the neural network.

[0404] 29) The above EN 滚筒 、EN 电机 、Fre 滚筒 and Fre 电机 were all tested under the premise of normal and good motors. In fact, in order to detect, identify, and verify data anomalies under different fault conditions, for the entanglement of clothes in the drum (electromagnetic abnormal sound), drum abnormality (mechanical accompanying abnormal sound), and motor abnormality (motor abnormal sound), etc., the corresponding data needs to be collected according to the above steps.

[0405] Refer to Figure 23 , which is a schematic diagram of the energy entropy of positive and negative samples. Refer to Figure 24 , which is a waveform schematic diagram of positive and negative samples.

[0406] 30) Download the washing machine program containing the anomaly detection model into the main board.

[0407] 31) The user puts the clothes into the drum, selects a certain washing program, and sets the water level.

[0408] 32) The washing machine door is closed, and the main board of the washing machine continuously obtains the current water level value W and the motor speed value V of the washing machine through the water level sensor and the motor speed sensor.

[0409] 33) The washing machine continuously reads D 滚筒 、D 电机 、B 滚筒 and B 电机 values.

[0410] 34) After filtering, the values of D 滚筒-滤波 、D 电机-滤波 、B 滚筒-滤波 and B 电机-滤波 are obtained.

[0411] 35) For the extraction of corresponding features, EN 滚筒 、EN 电机 、Fre 滚筒 and Fre 电机 are obtained.

[0412] 36) According to the current rotation speed, water level, and four-dimensional data parameters, substitute them into the classification model to obtain the current state of the washing machine.

[0413] 37) According to the inspection result of the current operating state of the washing machine, the washing machine will record the information of the current failure and upload it in a timely manner, enhancing the user experience while realizing automation.

[0414] In one embodiment, the filtering and feature extraction processing of the sound signal need to be carried out around frequency-based algorithms, but are not limited to those mentioned above. The filtering and feature extraction processing of the vibration signal need to be carried out around energy rate-based algorithms, but are not limited to those mentioned above. If a serious anomaly appears in the detection result of the neural network, such as an obvious deviation from the preset range in the detection result value, switch back to the manual mode and remind the user that there may be a fatal mechanical damage.

[0415] The neural network is only used to represent the "algorithm black box" in the process from data input to output. Since algorithm patents are not transparent. Therefore, in actual application, on the basis of ensuring the basic logic of the above physical parameters, the scope of the analysis algorithm is not restricted.

[0416] Combining coupling factors such as "motor sound + water flow sound" and "motor vibration + drum resonance" to calibrate and detect the state of the washing machine increases the applicability of this method. Based on the essential logic that the rotation of the motor (cause) of the washing machine drives the rotation of the water flow (result), combining the presentation of the "cause" and the "result" for real-time analysis optimizes the solution idea for this problem and improves the dynamic performance analysis ability of the washing machine.

[0417] Through the combined action of the sound sensor and displacement sensor at the top, this solution detects the drum water fluid condition of the washing machine by combining acoustic information and displacement information. This solution fully combines the characteristics of water fluid and the characteristics of the motor, and evaluates the operating state of the washing machine from two aspects of structure and performance. The analysis model established in advance in this solution contains sound waveform and displacement waveform information at different water levels and different speeds, and can be relatively easily migrated on different platforms, and is suitable for different models of current washing machines.

[0418] It should be noted that "first" and "second" in the embodiments of the present disclosure are only for convenience of expression and distinction, and have no other specific meanings.

[0419] Figure 25 is a block diagram of a washing machine shown according to an exemplary embodiment. Refer to Figure 25 , the washing machine may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, a sensor component 814, and a communication component 816.

[0420] The processing component 802 generally controls the overall operation of the washing machine, such as operations associated with display, data communication, camera operation, and recording operation. The processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the above methods. In addition, the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components. For example, the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.

[0421] The memory 804 is configured to store various types of data to support the operation of the washing machine. Examples of these data include instructions for any application or method for operating on the washing machine, messages, pictures, videos, etc. The memory 804 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0422] The power component 806 provides power for various components of the washing machine. The power component 806 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the washing machine.

[0423] The multimedia component 808 includes a screen that provides an output interface between the washing machine and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of the touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component 808 includes a camera. When the washing machine is in an operation mode, such as a shooting mode or a video mode, the camera can receive external multimedia data. The camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

[0424] The audio component 810 is configured to output and / or input audio signals. For example, the audio component 810 includes a microphone (MIC) that is configured to receive external audio signals when the washing machine is in operation modes, such as a recording mode and a voice recognition mode. The received audio signals can be further stored in the memory 804 or transmitted via the communication component 816. In some embodiments, the audio component 810 further includes a speaker for outputting audio signals.

[0425] The I / O interface 812 provides an interface between the processing component 802 and a peripheral interface module, which can be a keyboard, a click wheel, buttons, etc. These buttons can include but are not limited to: a home button, a volume button, a start button, and a lock button.

[0426] The sensor component 814 includes one or more sensors for providing a status assessment of various aspects of the washing machine. For example, the sensor component 814 can detect the open / closed state of the washing machine, the relative positioning of components, such as the display and keypad of the washing machine, the sensor component 814 can also detect a change in the position of the washing machine or a component of the washing machine, the presence or absence of user contact with the washing machine, the orientation or acceleration / deceleration of the washing machine, and the temperature change of the washing machine. The sensor component 814 can include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor component 814 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 814 can further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0427] The communication component 816 is configured to facilitate communication between the washing machine and other devices in a wired or wireless manner. The washing machine can access a communication standard-based wireless network, such as Wi-Fi, 4G, or 5G, or a combination thereof. In an exemplary embodiment, the communication component 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 816 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra-Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0428] In an exemplary embodiment, the washing machine can be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above-described method.

[0429] Those skilled in the art will readily conceive of other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. The present disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common general knowledge or conventional technical means in the art not disclosed herein. The specification and examples are only to be considered as exemplary, and the true scope and spirit of the present disclosure are pointed out by the following claims.

[0430] It should be understood that the present disclosure is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present disclosure is only limited by the appended claims.

Claims

1. An anomaly detection method, characterized in that, it includes: Obtain the first target information of the washing machine, where the first target information includes: the current water level in the washing machine and the current rotation speed of the motor; Determine the second target information of the washing machine at the current water level and the current rotation speed; where the second target information at least includes the target audio information at the drum of the washing machine, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor; Detect whether the operating state of the washing machine is abnormal according to the first target information and the second target information.

2. The method according to claim 1, characterized in that, The determination of the second target information of the washing machine includes: Collect the first original audio signal at the drum of the washing machine and the second original audio signal of the motor; Filter the first original audio signal and the second original audio signal respectively to obtain the first audio signal corresponding to the first original audio signal and the second audio signal corresponding to the second original audio signal; Extract the characteristic information of the first audio signal to obtain the first characteristic information; Extract the characteristic information of the second audio signal to obtain the second characteristic information; Determine the first characteristic information as the target audio information at the drum, and determine the second characteristic information as the target audio information of the motor.

3. The method according to claim 2, characterized in that, The extraction of the characteristic information of the first audio signal to obtain the first characteristic information includes: Extract the first signal sequence of each rotation period of the drum from the first audio signal; Segment the first signal sequence to obtain N first target sequences; where each first target sequence includes M sampling points; Determine the first frequency peak corresponding to each first target sequence; Determine the second frequency peak of the rotation period according to the first frequency peak; Determine the first characteristic information according to the second frequency peak.

4. The method according to claim 2, characterized in that, The extraction of the characteristic information of the second audio signal to obtain the second characteristic information includes: Extract the second signal sequence of each rotation period of the motor from the second audio signal; Segment the second signal sequence to obtain N second target sequences; where each second target sequence includes M sampling points; Determine the third frequency peak corresponding to each second target sequence; Determine the fourth frequency peak of the rotation period according to the third frequency peak; Determine the second characteristic information according to the fourth frequency peak.

5. The method according to claim 1, characterized in that, The determination of the second target information of the washing machine includes: Collect the first original vibration signal at the drum of the washing machine and the second original vibration signal of the motor; Filter the first original vibration signal and the second original vibration signal respectively to obtain the first vibration signal corresponding to the first original vibration signal and the second vibration signal corresponding to the second original vibration signal; Extract the characteristic information of the first vibration signal to obtain the third characteristic information; Extract the characteristic information of the second vibration signal to obtain the fourth characteristic information; Determine the third characteristic information as the target vibration information of the drum, and determine the fourth characteristic information as the target vibration information of the motor.

6. The method according to claim 5, wherein, The filtering the first original vibration signal and the second original vibration signal respectively to obtain a first vibration signal corresponding to the first original vibration signal and a second vibration signal corresponding to the second original vibration signal includes: Determine the maximum value and the minimum value of the first target vibration signal; wherein, the first target vibration signal is the first original vibration signal or the second original vibration signal; Determine a plurality of components of the first target vibration signal according to the maximum value, the minimum value and the first target vibration signal; Determine a second target vibration signal according to the components determined in the previous k times; wherein, the second target vibration signal is the first vibration signal or the second vibration signal; Wherein, when the first target vibration signal is the first original vibration signal, the second target vibration signal is the first vibration signal; when the first target vibration signal is the second original vibration signal, the second target vibration signal is the second vibration signal.

7. The method according to claim 5, wherein, The extracting the characteristic information of the first vibration signal to obtain the third characteristic information includes: Extract a third signal sequence of each rotation period of the drum from the first vibration signal; Segment the third signal sequence to obtain F third target sequences; wherein, each third target sequence includes H sampling points; Determine the first energy of each of the third target sequences, the first total energy of the F third target sequences, and the first proportion of each of the first energies in the first total energy; Determine the first energy entropy of each rotation period according to the first energy, the first total energy and the first proportion; Determine the third characteristic information according to the first energy entropy.

8. The method according to claim 5, wherein, The extracting the characteristic information of the second vibration signal to obtain the fourth characteristic information includes: Extract a fourth signal sequence of each rotation period of the motor from the second vibration signal; Segment the fourth signal sequence to obtain F fourth target sequences; wherein, each fourth target sequence includes H sampling points; Determine the second energy of each of the fourth target sequences, the second total energy of the F fourth target sequences, and the second proportion of each of the second energies in the second total energy; Determine the second energy entropy of each rotation period according to the second energy, the second total energy and the second proportion; Determine the fourth characteristic information according to the second energy entropy.

9. The method according to claim 1, wherein, The detecting whether the operating state of the washing machine is abnormal includes: Using the target detection model, determine whether the operating state of the washing machine is abnormal according to the current water level, the current rotation speed, the target audio information at the drum, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor.

10. The method according to claim 9, wherein, the target detection model is a model obtained by training an initial network model using a training sample set in a machine learning manner; the training sample set includes: a positive sample set, including a plurality of different positive samples, each positive sample including: a water level sample, a positive motor speed sample, and sample information of normal sound at the drum in the washing machine, sample information of normal vibration of the drum, sample information of normal sound of the motor, and sample information of normal vibration of the motor under the water level sample and the positive motor speed sample; each positive sample has its own first label for identifying that the output of the positive sample corresponding to the initial network model is normal; a negative sample set, including a plurality of negative samples, each negative sample including: the water level sample, a negative motor speed sample, and negative sample audio information at the drum in the washing machine, negative sample vibration information of the drum, negative sample audio information of the motor, and negative sample vibration information of the motor under the water level sample and the negative motor speed sample; wherein, at least one of the negative sample audio information at the drum in the washing machine, the negative sample vibration information of the drum, the negative sample audio information of the motor, and the negative sample vibration information of the motor is abnormal information; each negative sample has its own second label for identifying that the output of the negative sample corresponding to the initial network model is abnormal.

11. An abnormal detection device, wherein, comprising: an acquisition module, configured to acquire first target information of the washing machine, wherein the first target information includes: the current rotation speed and the current water level of the motor in the washing machine; a determination module, configured to determine second target information of the washing machine at the current water level and the current rotation speed; wherein the second target information at least includes the target audio information at the drum in the washing machine, the target vibration information of the drum, the target audio information of the motor, and the target vibration information of the motor; a detection module, configured to detect whether the operating state of the washing machine is abnormal according to the first target information and the second target information.

12. A washing machine, wherein, comprising: a processor and a memory for storing executable instructions that can run on the processor, wherein: when the processor is used to run the executable instructions, the executable instructions execute the method according to any one of claims 1 to 10 above.

13. A non-transitory computer-readable storage medium, wherein, the computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the method according to any one of claims 1 to 10 above is implemented.

Images