Abnormal sound detection method, device, equipment and storage medium for equipment

Abstract

An embodiment of the present invention provides a method, apparatus, device, and storage medium for detecting abnormal sounds of a device. The device is provided with a sound sensor and a blower assembly. The method comprises: obtaining sound data collected by the sound sensor; determining whether the device has an abnormality based on the sound data; if it is determined that the device has an abnormality, blowing air to a preset area within the device through the blower assembly; after blowing air through the blower assembly, determining abnormal information about the device based on the sound data. The sound sensor monitors the operating sound of the device in real time to reduce the need for manual intervention, identifies device abnormalities in real time based on the sound data, and reduces missed detections and misjudgments. If an abnormality is detected, the blower assembly blows air to a preset area. If the abnormality within the device is caused by a movable foreign object, the airflow generated by the blower assembly guides the foreign object to move, thereby amplifying the sound data generated by the foreign object, further improving the accuracy of determining device abnormality information, and thus improving detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of device abnormality detection, and in particular to a method for detecting abnormal sound of a device, an abnormal sound detection device for a device, an electronic device, and a computer-readable storage medium. Background Art

[0002] With the increasing prevalence of household appliances, industrial equipment, and electronic products, the problem of noise caused by foreign objects such as insects is becoming increasingly prominent. These foreign objects can increase noise levels or cause psychological discomfort. Currently, the industry relies primarily on traditional methods such as manual inspection, visual testing, and vibration analysis, but these methods have significant limitations and cannot meet the needs of efficient, accurate, and real-time monitoring.

[0003] Currently, traditional methods mainly rely on manual inspection, but manual inspection is inefficient and difficult to detect foreign objects hidden in corners or gaps, resulting in low detection accuracy and poor efficiency, which cannot meet actual application needs. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide an abnormal sound detection method for a device, an abnormal sound detection apparatus for a device, an electronic device, and a computer-readable storage medium that overcome the above problems or at least partially solve the above problems.

[0005] To solve the above problem, a first aspect of an embodiment of the present invention provides a method for detecting abnormal sound of a device, wherein the device is provided with a sound sensor and a blower assembly, and the method comprises:

[0006] Acquiring sound data collected by the sound sensor;

[0007] determining whether the device has an abnormality according to the sound data;

[0008] If it is determined that the device is abnormal, blowing air to a preset area in the device through the blowing component;

[0009] After the air is blown by the air blowing assembly, abnormal information of the device is determined based on the sound data.

[0010] Optionally, the device is further provided with an expulsion component, and the method further comprises:

[0011] When the abnormal information of the device indicates that there is a living thing in the device, the foreign matter in the device is expelled by the expelling component.

[0012] Optionally, an air duct covering the preset area is provided in the device; the blowing assembly is used to blow air into the air duct, so that the airflow generated by the blowing assembly passes through the air duct to cover the preset area.

[0013] Optionally, a plurality of air blowing components are provided in the device; and the method further comprises:

[0014] determining an abnormal area of ​​the device where an abnormality exists based on the sound data;

[0015] determining a target air blast assembly close to the abnormal area;

[0016] The abnormal area is blown by the target blowing assembly so that the airflow generated by the target blowing assembly covers the abnormal area.

[0017] Optionally, determining abnormal information of the device according to the sound data includes:

[0018] Performing noise reduction processing on the sound data and amplifying and outputting the sound data through a microphone array deployed on the device;

[0019] Preprocessing the noise-reduced sound data, and extracting features from the preprocessed sound data to obtain sound features;

[0020] The sound features are input into a trained abnormal sound detection model, and abnormal information of the device is output.

[0021] Optionally, the performing noise reduction processing on the sound data includes:

[0022] Performing noise reduction processing on the sound data to reduce ambient noise when the device is running;

[0023] The sound data after noise reduction processing is input into a trained noise classification model, and the noise generated by the blowing component in the sound data is filtered to obtain filtered sound data.

[0024] Optionally, the abnormal sound detection model is trained in the following manner:

[0025] Acquire a sound data set collected by the sound sensor;

[0026] performing noise reduction processing on the sound data set;

[0027] Preprocessing the sound data set after noise reduction processing, and performing feature extraction on the preprocessed sound data set to obtain a sound feature set;

[0028] Inputting the sound feature set into a training model to train the training model to determine abnormal information of the device;

[0029] Obtaining the accuracy of the abnormal information of the device output by the training model;

[0030] The training model is adjusted according to the accuracy to obtain the abnormal sound detection model.

[0031] Optionally, the noise classification model is trained in the following manner:

[0032] Acquiring sound data samples; the sound data samples include normal sound samples and abnormal sound samples;

[0033] Preprocessing the normal sound samples and the abnormal sound samples, and performing feature extraction on the preprocessed normal sound samples and the abnormal sound samples to obtain a normal sound feature set and an abnormal sound feature set;

[0034] Inputting the normal sound feature set and the abnormal sound feature set into a classification model to train the classification model to filter the normal sound data from the abnormal sound data to obtain target sound data;

[0035] Obtaining the accuracy of the target sound data output by the classification model;

[0036] The classification model is adjusted according to the accuracy to obtain the noise classification model.

[0037] According to a second aspect of an embodiment of the present invention, there is provided an abnormal sound detection device for a device, wherein the device is provided with a sound sensor and a blower assembly, and the device comprises:

[0038] A sound acquisition module, used to acquire the sound data collected by the sound sensor;

[0039] an abnormality judgment module, configured to determine whether the device has an abnormality based on the sound data;

[0040] an equipment air blowing module, configured to blow air to a preset area within the equipment through the air blowing assembly if it is determined that the equipment is abnormal;

[0041] An information determination module is used to determine abnormal information of the device according to the sound data after the blowing component blows air.

[0042] Optionally, the device is further provided with an expulsion component, and the method further comprises:

[0043] The foreign body expulsion module is used to expel the foreign body in the device through the expulsion component when the abnormal information of the device indicates that there is a living thing in the device.

[0044] Optionally, an air duct covering the preset area is provided in the device; the blowing assembly is used to blow air into the air duct, so that the airflow generated by the blowing assembly passes through the air duct to cover the preset area.

[0045] Optionally, a plurality of air blowing components are provided in the device; and the method further comprises:

[0046] an abnormal area determination module, configured to determine an abnormal area where the device is abnormal based on the sound data;

[0047] an air blast component determination module, configured to determine a target air blast component close to the abnormal area;

[0048] The abnormal area blowing module is used to blow air to the abnormal area through the target blowing component, so that the airflow generated by the target blowing component covers the abnormal area.

[0049] Optionally, the information determination module includes:

[0050] A sound output submodule, configured to perform noise reduction processing on the sound data and amplify and output the sound data through a microphone array deployed on the device;

[0051] The feature extraction submodule is used to preprocess the sound data after noise reduction processing and perform feature extraction on the preprocessed sound data to obtain sound features;

[0052] The information output submodule is used to input the sound features into the trained abnormal sound detection model and output the abnormal information of the device.

[0053] Optionally, the sound output submodule includes:

[0054] a sound noise reduction unit, configured to perform noise reduction processing on the sound data to reduce ambient noise when the device is running;

[0055] The noise filtering unit is used to input the sound data after noise reduction processing into the trained noise classification model, filter the noise generated by the blowing component in the sound data, and obtain filtered sound data.

[0056] Optionally, the abnormal sound detection model is trained in the following manner:

[0057] A sound data set acquisition module, configured to acquire a sound data set collected by the sound sensor;

[0058] A sound data set denoising module, configured to perform denoising processing on the sound data set;

[0059] The sound data set preprocessing module is used to preprocess the sound data set after noise reduction processing and perform feature extraction on the preprocessed sound data set to obtain a sound feature set;

[0060] A model training module, configured to input the sound feature set into a training model to train the training model to determine abnormal information of the device;

[0061] An accuracy acquisition module, configured to acquire an accuracy rate of the abnormal information of the device output by the training model;

[0062] The first model adjustment module is used to adjust the training model according to the accuracy to obtain the abnormal sound detection model.

[0063] Optionally, the noise classification model is trained in the following manner:

[0064] A sound sample acquisition module is used to acquire sound data samples; the sound data samples include normal sound samples and abnormal sound samples;

[0065] A sound sample preprocessing module, configured to preprocess the normal sound sample and the abnormal sound sample, and perform feature extraction on the preprocessed normal sound sample and the abnormal sound sample to obtain a normal sound feature set and an abnormal sound feature set;

[0066] A filtering model training module, configured to input the normal sound feature set and the abnormal sound feature set into a classification model to train the classification model to filter the normal sound data from the abnormal sound data to obtain the target sound data;

[0067] An output accuracy acquisition module, used to obtain the accuracy of the target sound data output by the classification model;

[0068] The second model adjustment module is used to adjust the classification model according to the accuracy to obtain the noise classification model.

[0069] According to a third aspect of an embodiment of the present invention, an electronic device is provided, comprising: a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of a method for detecting abnormal sound of a device as described in any one of the above items are implemented.

[0070] According to a fourth aspect of an embodiment of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the abnormal sound detection method of a device as described in any one of the above are implemented.

[0071] The technical solutions provided by the embodiments of the present invention may have the following beneficial effects:

[0072] An embodiment of the present invention provides a method, apparatus, device, and storage medium for detecting abnormal sounds of a device. The device is provided with a sound sensor and a blower assembly. The method comprises: obtaining sound data collected by the sound sensor; determining whether the device has an abnormality based on the sound data; if it is determined that the device has an abnormality, blowing air to a preset area within the device through the blower assembly; after blowing air through the blower assembly, determining abnormal information about the device based on the sound data. The sound sensor monitors the operating sound of the device in real time to reduce the need for manual intervention, identifies device abnormalities in real time based on the sound data, and reduces missed detections and misjudgments. If an abnormality is detected, the blower assembly blows air to a preset area. If the abnormality within the device is caused by a movable foreign object, the airflow generated by the blower assembly guides the foreign object to move, thereby amplifying the sound data generated by the foreign object, further improving the accuracy of determining device abnormality information, and thus improving detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 This is a flowchart of a method for detecting abnormal sound of a device provided by an embodiment of the present invention;

[0074] Figure 2 This is a flowchart of the steps of an abnormal sound detection method of another device provided by an embodiment of the present invention;

[0075] Figure 3 This is a structural block diagram of an abnormal sound detection device of a device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0076] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0077] Currently, traditional methods mainly rely on manual inspection, but manual inspection is inefficient and difficult to detect foreign objects hidden in corners or gaps, resulting in low detection accuracy and poor efficiency, which cannot meet actual application needs.

[0078] One of the core concepts of the embodiments of the present invention is to monitor the operating sound of the equipment in real time through sound sensors, reduce the need for manual intervention, identify equipment abnormalities in real time based on sound data, and reduce missed detections and misjudgments. If an abnormality is detected, the blowing component actively blows air to a preset area, and the equipment abnormality information is further determined based on the sound data after the blowing component blows, thereby improving detection accuracy.

[0079] Reference Figure 1, which shows a flowchart of a method for detecting abnormal sound of a device provided by an embodiment of the present invention. The device is provided with a sound sensor and a blast assembly. The method may specifically include the following steps:

[0080] Step 101, obtaining sound data collected by the sound sensor;

[0081] Currently, the detection of foreign objects and abnormal noises inside equipment primarily relies on traditional methods such as manual inspection, visual inspection, or vibration detection. However, these methods are inefficient, highly variable, and difficult to detect foreign objects hidden in corners or crevices. The operation is complex, with numerous interfering and influencing factors, and real-time monitoring is difficult. An embodiment of the present invention provides a method for detecting abnormal sounds in equipment that can sensitively and in real time detect the presence of foreign objects, abnormal load jitter, and abnormal noise inside the equipment, thereby improving the accuracy and reliability of detection.

[0082] The device of the embodiment of the present invention can be a fan product such as a cooling fan, a circulating fan, a tower fan, or various electrical appliances. In this embodiment, a cooling fan equipped with a sound sensor and a blower assembly is used as an example to illustrate the principles of the present invention. However, this description is merely exemplary and the scope of the present invention is not limited thereto. The principles of the present invention can also be applied to household appliances equipped with sound sensors and blower assemblies.

[0083] An acoustic sensor is an electronic device that can detect, receive, and convert sound waves into electrical signals. It is widely used in environmental monitoring, smart homes, industrial control, and other fields. It converts sound wave vibrations into electrical signals through a microphone.

[0084] Sound sensors are placed near special locations requiring testing, sensitive areas, and specific locations within the device that require testing, such as motors, bearings, and air ducts. This is because most users prefer not to have foreign objects in these locations. The presence of foreign objects can cause discomfort and even contamination, impacting the user experience. Furthermore, foreign objects in live controllers can easily cause short circuits and damage the controllers.

[0085] A blower assembly is a device that mechanically generates or enhances air flow and is widely used in ventilation, cooling, combustion support, material conveying, and other fields. Its core components include a motor that provides power; an impeller / fan that converts mechanical energy into airflow; and a housing / duct that directs airflow and reduces turbulence and noise.

[0086] In the embodiment of the present invention, multi-channel synchronous acquisition is performed by setting sound sensors at special positions or near sensitive parts that need to be detected, and the internal sound field information of the device is captured in real time to ensure that the sound signal can be effectively obtained.

[0087] Step 102, determining whether the device has an abnormality based on the sound data;

[0088] Determining whether there is an abnormality in the device indicates whether the sound data generated when the device is running is the same as the sound data generated when the device is operating normally. If they are different, it means that a component in the device is damaged or there is a foreign object in the device.

[0089] In an embodiment of the present invention, real-time analysis is performed based on the acoustic data collected by the sound sensor. By comparing it with the sound database of the normal operation of the equipment, the unique signals of foreign objects and abnormal mechanical vibration characteristics can be accurately identified. When abnormal characteristics of the sound characteristics exceeding the threshold are detected, it is determined that there is an abnormality in the equipment.

[0090] Step 103: If it is determined that the device is abnormal, the blower assembly blows air to a preset area within the device;

[0091] Setting a predefined area in a device or system typically involves defining a specific range through physical boundaries, electronic fences, or software parameters. This is used to control the device's operating range, safety zone, or function trigger conditions. Predefined areas cover specific locations or sensitive areas within the device that require detection.

[0092] In this embodiment of the present invention, when an equipment anomaly is detected based on acoustic data collected by a sound sensor, the intelligent air blowing component is activated to generate directional airflow in a pre-set area, leveraging the airflow effect to stimulate the movement of potential foreign objects. The sound sensor also monitors in real time, capturing the sound data of foreign objects colliding or moving under the airflow disturbance.

[0093] Step 104 : After the air blowing component blows air, determine abnormal information of the device according to the sound data.

[0094] Device abnormality information refers to abnormal status indications caused by malfunctions, environmental disturbances, or improper operation during device operation. This information is crucial for fault diagnosis, preventive maintenance, and safety control. Device abnormality information indicates damage to a component or the presence of a moving foreign object within the device.

[0095] In this embodiment of the present invention, after the blower assembly is activated, an enhanced sound data detection mode is activated. Sound sensors collect real-time internal device sound data for in-depth analysis. Based on the sound data collected after the blower assembly is activated, abnormal patterns such as foreign object collisions (such as the scraping sound of cockroach legs against metal surfaces), structural looseness (such as resonance caused by a loose screw), or mechanical jamming are identified. The system dynamically compares the acoustic signatures before and after the blower assembly is activated, ultimately outputting device anomaly information, including the type of anomaly.

[0096] Reference Figure 2, shows a flowchart of the steps of an abnormal sound detection method of another device provided by an embodiment of the present invention. The device is also provided with an expulsion component. The method may specifically include the following steps:

[0097] Step 201, obtaining sound data collected by the sound sensor;

[0098] In the embodiment of the present invention, multi-channel synchronous acquisition is performed by setting sound sensors at special positions or near sensitive parts that need to be detected, and the internal sound field information of the device is captured in real time to ensure that the sound signal can be effectively obtained.

[0099] Step 202, determining whether the device has an abnormality based on the sound data;

[0100] In an embodiment of the present invention, real-time analysis is performed based on the acoustic data collected by the sound sensor. By comparing it with the sound database of the normal operation of the equipment, the unique signals of foreign objects and abnormal mechanical vibration characteristics can be accurately identified. When abnormal characteristics of the sound characteristics exceeding the threshold are detected, it is determined that there is an abnormality in the equipment.

[0101] Step 203: If it is determined that the device is abnormal, the blower assembly blows air to a preset area within the device;

[0102] In this embodiment of the present invention, when an equipment anomaly is detected based on acoustic data collected by a sound sensor, the intelligent air blowing component is activated to generate directional airflow in a pre-set area, leveraging the airflow effect to stimulate the movement of potential foreign objects. The sound sensor also monitors in real time, capturing the sound data of foreign objects colliding or moving under the airflow disturbance.

[0103] Step 204 : After the air blowing component blows air, determine abnormal information of the device according to the sound data.

[0104] In this embodiment of the present invention, after the blower assembly is activated, an enhanced sound data detection mode is activated. Sound sensors collect real-time internal device sound data for in-depth analysis. Based on the sound data collected after the blower assembly is activated, abnormal patterns such as foreign object collisions (such as the scraping sound of cockroach legs against metal surfaces), structural looseness (such as resonance caused by a loose screw), or mechanical jamming are identified. The system dynamically compares the acoustic signatures before and after the blower assembly is activated, ultimately outputting device anomaly information, including the type of anomaly.

[0105] In some embodiments, step 204 may include the following sub-steps:

[0106] Sub-step S11, performing noise reduction processing on the sound data, and amplifying and outputting the sound data through a microphone array deployed on the device;

[0107] In equipment monitoring and fault diagnosis, noise reduction is used to filter out interference from the original signal and extract effective abnormal information to improve detection accuracy and reliability.

[0108] A microphone array is a system composed of multiple microphones arranged in a specific geometric structure. It uses beamforming and sound source localization technology to achieve functions such as noise suppression, speech enhancement, and sound source tracking.

[0109] In this embodiment of the present invention, multi-stage noise reduction and intelligent enhancement technology are employed to optimize the processing of raw sound data. An improved spectral subtraction method and adaptive filtering (LMS algorithm) are used to eliminate steady-state noise generated by the device's operation (such as motor hum and airflow noise), preserving key audio frequency bands and effectively improving the clarity and intelligibility of the target signal. The enhanced audio signal is then output through a microphone array deployed at key locations on the device, ensuring effective output of abnormal sounds and providing high-fidelity audio input for subsequent anomaly detection algorithms.

[0110] Sub-step S12, pre-processing the sound data after the noise reduction processing, and extracting features from the pre-processed sound data to obtain sound features;

[0111] Preprocessing noise-reduced sound data is a key step in tasks such as speech recognition and acoustic analysis. Its purpose is to further eliminate residual noise, enhance effective signals, and extract features. Preprocessing involves standardizing the sound data to ensure comparability between samples, segmenting the sound data into small time windows (e.g., 1-second intervals), and labeling each window for abnormal sounds.

[0112] Feature extraction from preprocessed sound data is a core step in tasks such as speech recognition and acoustic event detection. This process involves calculating statistical features of the sound signal, such as the mean, variance, kurtosis, and peak value, using time-domain features; extracting spectral features of the sound signal, such as the power spectrum and frequency center, using Fourier transforms or wavelet transforms; and extracting time-frequency features using short-time Fourier transforms (STFTs) or Mel-spectral coefficients (MFCCs).

[0113] Sound features are quantitative indicators that describe the characteristics of audio signals and are widely used in speech recognition, music analysis, environmental sound detection and other fields. They include time domain features and frequency domain features.

[0114] In this embodiment of the present invention, a standardized preprocessing process is performed on the noise-reduced sound data. First, the sound data is standardized, then segmented and labeled. The sound data is divided into small time windows (e.g., 1-second windows), and each window is labeled to determine whether it contains abnormal sound. During the feature extraction phase, three key features are extracted: time domain features, frequency domain features, and time-frequency domain features. These features are then normalized to produce sound features, which provide input data for the subsequent determination of device abnormality information.

[0115] Sub-step S13: input the sound features into a trained abnormal sound detection model and output abnormal information of the device.

[0116] Abnormal sound detection is a technology that identifies abnormal events (such as mechanical failures and hazardous environmental sounds) by analyzing sound signals. A trained abnormal sound detection model is obtained by collecting a large amount of sound data to train a baseline model.

[0117] In this embodiment of the present invention, the extracted fused sound feature vector is fed into a trained anomaly detection model for real-time inference. This model strengthens the feature weights of key frequency bands and employs a multi-task learning framework to output device anomaly information. The results are then transmitted to the device's main control system, triggering an audible and visual alarm, and uploaded to a cloud-based operations and maintenance platform.

[0118] Through noise reduction processing of sound data, abnormal sound features are made clearer, reducing misjudgments. Multi-microphone beamforming technology is used to focus on the sound in the target area and suppress irrelevant noise. The pre-processed and feature-extracted sound data is input into a trained abnormal sound detection model, which outputs device abnormality information. Automated processing meets the needs of real-time device monitoring.

[0119] In some embodiments, step S11 may include the following sub-steps:

[0120] Sub-step S111, performing noise reduction processing on the sound data to reduce the ambient noise when the device is running;

[0121] Use noise reduction algorithms (such as spectral subtraction and adaptive filters) to reduce background noise during fan operation and improve sound signal clarity. First, apply spectral subtraction to reduce the fan's continuous background noise and minimize interference. Then, adjust the filter parameters in real time to further optimize the noise reduction effect.

[0122] In an embodiment of the present invention, a multi-stage adaptive noise reduction technology is used to reduce the noise of the original sound signal: spectral subtraction is used to eliminate the steady-state noise generated by the operation of the equipment (such as motor hum, wind noise, etc.). At the same time, the adaptive filter is combined to specifically retain the key abnormal characteristic frequency bands, while suppressing low-frequency mechanical noise, providing high-fidelity audio input for subsequent abnormality detection.

[0123] Sub-step S112: inputting the noise-reduced sound data into a trained noise classification model, filtering the noise generated by the air blowing component in the sound data, and obtaining filtered sound data.

[0124] The noise classification model is designed to automatically identify and classify different types of noise (such as environmental noise, mechanical failure sounds, speech interference, etc.), and is widely used in industrial detection, speech enhancement, environmental monitoring and other fields.

[0125] In this embodiment of the present invention, the noise-reduced sound data is input into a trained noise classification model for real-time inference. The model identifies the characteristics of the wind noise, recognizes the broadband continuous characteristics of the wind noise, and outputs the frequency band of the filtered wind noise in real time.

[0126] By combining noise reduction preprocessing and noise classification model filtering to filter out wind noise, the interference of environmental noise on anomaly detection is reduced, the detection rate of weak anomalies is improved, and the accuracy and anti-interference ability of abnormal sound detection are significantly improved.

[0127] Step 205 : When the abnormal information of the device indicates that there is a living thing in the device, the foreign matter in the device is expelled by the expelling component.

[0128] The expulsion component, deployed within the device, automatically executes expulsion operations upon detecting a living foreign object. This component includes an odor expulsion module and a sound expulsion module. The odor expulsion module releases odors to expel living foreign objects within the device, while the sound expulsion module emits different sound frequencies for different moving foreign objects, guiding them to move and expel them.

[0129] In this embodiment of the present invention, when a device's abnormal information indicates the presence of a living foreign object, an alarm mechanism is triggered, notifying the user or control system of the possible presence of a foreign object. The system then immediately activates an expelling component, releasing an odor or sound to expel the foreign object. This effectively addresses the issue of unusual noises, unsanitary hazards, or equipment damage caused by the intrusion of living creatures (such as insects and small rodents) within the device.

[0130] In some embodiments, an air duct covering the preset area is provided in the device; the air blowing component is used to blow air into the air duct so that the airflow generated by the air blowing component passes through the air duct to cover the preset area.

[0131] Air duct designs are tailored to the product's appearance, ensuring even airflow across all key areas. Current devices have varying appearances, such as traditional floor fans, tower fans, bladeless fans, and multi-function fans, as well as varying internal controller placement. Therefore, the internal air duct design for each product must be tailored to its structure and appearance.

[0132] The device's internal air duct design ensures efficient airflow coverage of all critical areas. The duct consists of a main duct and multiple branch ducts. A blower assembly blows air into the duct, ensuring that the airflow flows through the duct to cover the designated area. The duct's diversion structure avoids blind spots caused by direct airflow, achieving full coverage. The blower airflow also effectively clears obstructions around the sensor, improving the quality of sound data and significantly enhancing the accuracy and efficiency of internal anomaly detection and live animal removal.

[0133] In some embodiments, the device is provided with a plurality of air blowing assemblies; the method further comprises:

[0134] determining an abnormal area of ​​the device where an abnormality exists based on the sound data;

[0135] determining a target air blast assembly close to the abnormal area;

[0136] The abnormal area is blown by the target blowing assembly so that the airflow generated by the target blowing assembly covers the abnormal area.

[0137] The device is provided with preset areas covering multiple key locations, and the abnormal area indicates an area where an abnormality exists at a key location within the device.

[0138] An air blowing assembly is provided near each key position, and the target air blowing assembly near the abnormal area is determined by determining the abnormal area.

[0139] In an embodiment of the present invention, sound data collected by sound sensors installed near various key locations are used to identify key locations within the device where anomalies exist, and the target blower assembly closest to the abnormal area is intelligently matched. After the target blower assembly is determined, the target blower assembly is used to blow air, and the airflow generated by the target blower assembly flows through the air duct, thereby covering the abnormal area. The airflow generated by the blower assembly is used to guide the movement of foreign objects in the abnormal area and amplify their sound signals. Through the precise control process of "abnormal area positioning, target blower assembly selection, and directional airflow processing," efficient diagnosis and treatment of equipment anomalies are achieved, significantly improving the efficiency of abnormality diagnosis.

[0140] In some embodiments, the abnormal sound detection model is trained in the following manner:

[0141] Acquire a sound data set collected by the sound sensor;

[0142] performing noise reduction processing on the sound data set;

[0143] Preprocessing the sound data set after noise reduction processing, and performing feature extraction on the preprocessed sound data set to obtain a sound feature set;

[0144] Inputting the sound feature set into a training model to train the training model to determine abnormal information of the device;

[0145] Obtaining the accuracy of the abnormal information of the device output by the training model;

[0146] The training model is adjusted according to the accuracy to obtain the abnormal sound detection model.

[0147] In this embodiment of the present invention, a sound sensor collects a dataset of raw sound during device operation, obtaining audio signals from the device in different states. This raw sound data is then subjected to noise reduction processing to remove irrelevant signals such as ambient noise, thereby improving data quality. The noise-reduced data is then preprocessed, including operations such as framing, windowing, and normalization. Time-domain, frequency-domain, or time-frequency-domain features are then extracted to form a sound feature set.

[0148] The feature set is input into a training model (such as a support vector machine (SVM), random forest, or deep learning model) and trained through supervised or self-supervised learning to identify normal and abnormal sound patterns. After training, a test set is used to evaluate the accuracy of the model's output of abnormal information, using metrics such as precision, recall, and F1 score. If the model performance is insufficient, optimization is performed by adjusting the model structure, parameters (such as the learning rate), or data augmentation, ultimately resulting in a highly accurate abnormal sound detection model. This automated analysis, from raw audio to intelligent diagnosis, can be widely applied in scenarios such as industrial equipment fault warning and home appliance anomaly detection.

[0149] In order to enable those skilled in the art to better understand the embodiments of the present invention, the embodiments of the present invention are described below using an example:

[0150] Identify key locations within the fan that require monitoring, such as the motor, bearings, and air ducts. Identify the types of anomalies that require detection, such as small animals or insects, and collect relevant sound samples within the device. Collect normal operating sound: When the fan is operating normally, collect internal sound data, including sounds at different speeds and loads. For abnormal sounds, simulate the situation where small animals or insects enter the fan and collect abnormal sound samples. This can be accomplished by introducing small animals or simulating their sounds. Collect environmental noise during device operation to train the model's anti-interference capabilities.

[0151] Perform noise reduction on the collected sound; use a noise reduction algorithm to reduce background noise when the fan is running and improve the clarity of the sound signal. Standardize the sound data to ensure comparability between different samples. Divide the sound data into small time windows and mark whether each window contains abnormal sound. Perform feature extraction on the noise-reduced sound; calculate statistical features such as the mean, variance, kurtosis, and peak of the sound signal. Use Fourier transform or wavelet transform to extract the spectral features of the sound signal, such as the power spectrum and frequency center. Use short-time Fourier transform or Mel spectrum coefficients to extract the time-frequency domain features of the sound signal. Time-frequency domain features: Use short-time Fourier transform or Mel spectrum coefficients to extract the time-frequency domain features of the sound signal.

[0152] Train the model based on the processed sound data; select an appropriate machine learning or deep learning algorithm based on the task requirements. Use the preprocessed dataset to train the selected model. Ensure that the training and test sets are properly divided to avoid overfitting. Optimize model parameters through methods such as cross-validation and grid search to improve model accuracy and generalization. Test the trained model to further optimize it; test its performance on the validation set, evaluating metrics such as accuracy, recall, and F1 score. Based on the test results, adjust model parameters or improve data preprocessing methods to further optimize model performance.

[0153] Integrate the trained model into the fan's AI voice module to ensure it can quickly detect abnormal sounds during real-time operation. Continuously monitor internal device sound data and use the trained model for real-time anomaly detection. Regularly collect new sound data for model updates and optimization, ensuring the model adapts to different environments and usage scenarios.

[0154] In some embodiments, the noise classification model is trained in the following manner:

[0155] Acquiring sound data samples; the sound data samples include normal sound samples and abnormal sound samples;

[0156] Preprocessing the normal sound samples and the abnormal sound samples, and performing feature extraction on the preprocessed normal sound samples and the abnormal sound samples to obtain a normal sound feature set and an abnormal sound feature set;

[0157] Inputting the normal sound feature set and the abnormal sound feature set into a classification model to train the classification model to filter the normal sound data from the abnormal sound data to obtain target sound data;

[0158] Obtaining the accuracy of the target sound data output by the classification model;

[0159] The classification model is adjusted according to the accuracy to obtain the noise classification model.

[0160] Normal sound samples include the sound data generated during normal equipment operation and the noise generated by the blower. Abnormal sound samples include the sound data generated during equipment operation, the noise generated by the blower, and the additional sound data generated by foreign objects within the equipment. The additional sound data generated by foreign objects within the equipment is obtained by filtering the normal sound data from the abnormal sound data through the trained classification model.

[0161] In this embodiment of the present invention, the system is primarily used to filter normal sound components from abnormal sound data to improve the accuracy of anomaly detection. A dataset containing normal sound samples (e.g., sounds generated during normal equipment operation) and abnormal sound samples (e.g., sounds generated during malfunctions or noise interference) is collected. These two types of samples are preprocessed, including noise reduction, framing, and normalization. Effective acoustic features are then extracted to form a normal sound feature set and an abnormal sound feature set.

[0162] These two feature sets are fed into a classification model for training, enabling it to identify and eliminate normal sound components from abnormal data, thereby generating purer target sound data (i.e., truly abnormal sounds). After training, the accuracy of the model's target data output (such as precision and recall) is evaluated. If performance is insufficient, model parameters are adjusted or feature extraction methods are optimized, ultimately resulting in a high-performance noise classification model. This method can be used in scenarios such as industrial equipment fault diagnosis and smart home anomaly detection, improving the reliability of the detection system by accurately separating normal and abnormal sounds.

[0163] An embodiment of the present invention provides a method for detecting abnormal sounds of a device, wherein the device is provided with a sound sensor and a blower assembly. The method comprises: obtaining sound data collected by the sound sensor; determining whether the device has an abnormality based on the sound data; if it is determined that the device has an abnormality, blowing air to a preset area within the device through the blower assembly; after blowing air through the blower assembly, determining abnormal information of the device based on the sound data. The device operating sound is monitored in real time by the sound sensor, reducing the need for manual intervention, and device abnormalities are identified in real time based on the sound data, reducing missed detections and misjudgments. If an abnormality is detected, the preset area is blown through the blower assembly. If the abnormality within the device is caused by a movable foreign object, the airflow generated by the blower assembly guides the foreign object to move, thereby amplifying the sound data generated by the foreign object, further improving the accuracy of determining device abnormality information, and thus improving detection accuracy.

[0164] It should be noted that for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0165] Reference Figure 3 , shows a structural block diagram of an abnormal sound detection device of a device provided by an embodiment of the present invention. The device is provided with a sound sensor and a blast assembly, and specifically may include the following modules:

[0166] A sound acquisition module 301 is used to acquire sound data collected by the sound sensor;

[0167] An abnormality determination module 302 is used to determine whether the device has an abnormality based on the sound data;

[0168] The device air blowing module 303 is configured to blow air to a preset area within the device through the air blowing component if it is determined that the device is abnormal;

[0169] The information determination module 304 is configured to determine abnormal information of the device according to the sound data after the blowing component blows air.

[0170] In some embodiments, the device is further provided with an expulsion component, and the method further comprises:

[0171] The foreign body expulsion module is used to expel the foreign body in the device through the expulsion component when the abnormal information of the device indicates that there is a living thing in the device.

[0172] In some embodiments, an air duct covering the preset area is provided in the device; the air blowing component is used to blow air into the air duct so that the airflow generated by the air blowing component passes through the air duct to cover the preset area.

[0173] In some embodiments, the device is provided with a plurality of air blowing assemblies; the method further comprises:

[0174] an abnormal area determination module, configured to determine an abnormal area where the device is abnormal based on the sound data;

[0175] an air blast component determination module, configured to determine a target air blast component close to the abnormal area;

[0176] The abnormal area blowing module is used to blow air to the abnormal area through the target blowing component, so that the airflow generated by the target blowing component covers the abnormal area.

[0177] In some embodiments, the information determination module 304 includes:

[0178] A sound output submodule, configured to perform noise reduction processing on the sound data and amplify and output the sound data through a microphone array deployed on the device;

[0179] The feature extraction submodule is used to preprocess the sound data after noise reduction processing and perform feature extraction on the preprocessed sound data to obtain sound features;

[0180] The information output submodule is used to input the sound features into the trained abnormal sound detection model and output the abnormal information of the device.

[0181] In some embodiments, the sound output submodule includes:

[0182] a sound noise reduction unit, configured to perform noise reduction processing on the sound data to reduce ambient noise when the device is running;

[0183] The noise filtering unit is used to input the sound data after noise reduction processing into the trained noise classification model, filter the noise generated by the blowing component in the sound data, and obtain filtered sound data.

[0184] In some embodiments, the abnormal sound detection model is trained in the following manner:

[0185] A sound data set acquisition module, configured to acquire a sound data set collected by the sound sensor;

[0186] A sound data set denoising module, configured to perform denoising processing on the sound data set;

[0187] The sound data set preprocessing module is used to preprocess the sound data set after noise reduction processing and perform feature extraction on the preprocessed sound data set to obtain a sound feature set;

[0188] A model training module, configured to input the sound feature set into a training model to train the training model to determine abnormal information of the device;

[0189] An accuracy acquisition module, configured to acquire an accuracy rate of the abnormal information of the device output by the training model;

[0190] The first model adjustment module is used to adjust the training model according to the accuracy to obtain the abnormal sound detection model.

[0191] In some embodiments, the noise classification model is trained in the following manner:

[0192] A sound sample acquisition module is used to acquire sound data samples; the sound data samples include normal sound samples and abnormal sound samples;

[0193] A sound sample preprocessing module, configured to preprocess the normal sound sample and the abnormal sound sample, and perform feature extraction on the preprocessed normal sound sample and the abnormal sound sample to obtain a normal sound feature set and an abnormal sound feature set;

[0194] A filtering model training module, configured to input the normal sound feature set and the abnormal sound feature set into a classification model to train the classification model to filter the normal sound data from the abnormal sound data to obtain the target sound data;

[0195] An output accuracy acquisition module, used to obtain the accuracy of the target sound data output by the classification model;

[0196] The second model adjustment module is used to adjust the classification model according to the accuracy to obtain the noise classification model.

[0197] An embodiment of the present invention provides an abnormal sound detection device for a device, wherein the device is provided with a sound sensor and a blower assembly, and the device comprises: a sound acquisition module for acquiring sound data collected by the sound sensor; an abnormality judgment module for determining whether the device has an abnormality based on the sound data; a device blower module for blowing air to a preset area within the device through the blower assembly if it is determined that the device has an abnormality; and an information determination module for determining abnormal information of the device based on the sound data after blowing air through the blower assembly. The device operating sound is monitored in real time by a sound sensor to reduce the need for manual intervention, and device abnormalities are identified in real time based on the sound data to reduce missed detections and misjudgments. If an abnormality is detected, the preset area is blown through the blower assembly. If the abnormality within the device is caused by a movable foreign object, the airflow generated by the blower assembly guides the foreign object to move, thereby amplifying the sound data generated by the foreign object, further improving the accuracy of determining device abnormality information, and thus improving detection accuracy.

[0198] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0199] An embodiment of the present invention further provides an electronic device comprising: a sound sensor, a blower assembly, a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When executed by the processor, the computer program implements the various processes of the abnormal sound detection method embodiment of the aforementioned device and can achieve the same technical effect. To avoid repetition, the description is not repeated here. The electronic device can be a fan product such as a cooling fan, a circulating fan, a tower fan, or various household electrical appliances. The embodiment of the present invention does not impose any restrictions on the specific type of electronic device.

[0200] An embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the abnormal sound detection method embodiment of the above-mentioned device are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0201] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0202] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, embodiments of the present invention may take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0203] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems) and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0204] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0205] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0206] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0207] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0208] The above is a detailed introduction to the abnormal sound detection method of a device and the abnormal sound detection device of a device provided by the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the ideas of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for detecting abnormal sound of a device, characterized in that: The device is provided with a sound sensor, a blast component, and an expelling component, and the method includes: Acquiring sound data currently collected by the sound sensor and reference sound data; the reference sound data is sound data generated when the device is operating normally; determining whether the device has an abnormality according to the currently collected sound data and the reference sound data; If it is determined that the device is abnormal, blowing air to a preset area in the device through the blowing component; After the air is blown by the air blowing assembly, abnormal information of the device is determined based on the sound data; the abnormal information indicates that a component inside the device is damaged or there is a movable foreign object in the device; When the abnormal information of the device indicates that there is a living thing in the device, the foreign matter in the device is expelled by the expelling component; the expelling component is used to release an odor or make a sound.

2. The abnormal sound detection method of a device according to claim 1, characterized in that: An air duct covering the preset area is provided in the device; the air blowing component is used to blow air into the air duct, so that the airflow generated by the air blowing component passes through the air duct to cover the preset area.

3. The abnormal sound detection method of a device according to claim 1, characterized in that: The device is provided with a plurality of air blowing components; the method further comprises: determining an abnormal area of ​​the device where an abnormality exists based on the sound data; determining a target air blast assembly close to the abnormal area; The abnormal area is blown by the target blowing assembly so that the airflow generated by the target blowing assembly covers the abnormal area.

4. The abnormal sound detection method of a device according to claim 1, characterized in that: Determining abnormal information of the device according to the sound data includes: Performing noise reduction processing on the sound data and amplifying and outputting the sound data through a microphone array deployed on the device; Preprocessing the noise-reduced sound data, and extracting features from the preprocessed sound data to obtain sound features; The sound features are input into a trained abnormal sound detection model, and abnormal information of the device is output.

5. The abnormal sound detection method of a device according to claim 4, characterized in that: The performing noise reduction processing on the sound data includes: Performing noise reduction processing on the sound data to reduce ambient noise when the device is running; The sound data after noise reduction processing is input into a trained noise classification model, and the noise generated by the blowing component in the sound data is filtered to obtain filtered sound data.

6. The abnormal sound detection method of a device according to claim 4, characterized in that: The abnormal sound detection model is trained in the following way: Acquire a sound data set collected by the sound sensor; performing noise reduction processing on the sound data set; Preprocessing the sound data set after noise reduction processing, and performing feature extraction on the preprocessed sound data set to obtain a sound feature set; Inputting the sound feature set into a training model to train the training model to determine abnormal information of the device; Obtaining the accuracy of the abnormal information of the device output by the training model; The training model is adjusted according to the accuracy to obtain the abnormal sound detection model.

7. The abnormal sound detection method of a device according to claim 5, characterized in that: The noise classification model is trained in the following way: Acquiring sound data samples; the sound data samples include normal sound samples and abnormal sound samples; Preprocessing the normal sound samples and the abnormal sound samples, and performing feature extraction on the preprocessed normal sound samples and the abnormal sound samples to obtain a normal sound feature set and an abnormal sound feature set; Inputting the normal sound feature set and the abnormal sound feature set into a classification model to train the classification model to filter the normal sound data from the abnormal sound data to obtain target sound data; Obtaining the accuracy of the target sound data output by the classification model; The classification model is adjusted according to the accuracy to obtain the noise classification model.

8. An abnormal sound detection device for equipment, characterized in that: The device is provided with a sound sensor, a blast component, and an expulsion component, and the device includes: A sound acquisition module, configured to acquire the sound data currently collected by the sound sensor and reference sound data; the reference sound data is the sound data generated during normal operation of the device; an abnormality judgment module, configured to determine whether the device has an abnormality based on the currently collected sound data and the reference sound data; an equipment air blowing module, configured to blow air to a preset area within the equipment through the air blowing assembly if it is determined that the equipment is abnormal; an information determination module, configured to determine abnormal information of the device based on the sound data after the blowing assembly blows air; the abnormal information indicating damage to a component within the device or the presence of a movable foreign object within the device; The foreign body expulsion module is used to expel the foreign body in the device through the expulsion component when the abnormal information of the device indicates that there is a living thing in the device; the expulsion component is used to release an odor or make a sound.

9. An electronic device, characterized in that: include: A sound sensor, a blower assembly, a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the abnormal sound detection method of an apparatus as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the abnormal sound detection method of a device according to any one of claims 1 to 7 are implemented.

Images