Balance fault detection method, balance fault detection device and storage medium

Abstract

The invention provides a balance fault detection method, a balance fault detection device and a storage medium, which are applied to a washing machine, the washing machine comprises a plurality of balance elements and a plurality of sensors, each balance element corresponds to at least one sensor, and the fault detection method comprises the following steps: triggering an abnormal operation state of the washing machine; acquiring operation state data of the plurality of balance elements through the plurality of sensors; inputting the operation state data into a machine learning model to obtain the fault abnormity degree of each balance element; outputting a fault probability sequence according to the fault abnormity degree; and positioning a fault balance element through fault probability sorting. The balance fault detection method can quickly and accurately position the abnormity of the internal components without disassembling the machine.

Description

Technical Field

[0001] The present invention relates to the field of display technology, and in particular to a balance fault detection method, a balance fault detection device, and a storage medium. Background Art

[0002] In the field of washing machine technology, the dynamic balancing system is a core component that ensures a smooth spin cycle. Its performance stability is directly related to equipment efficiency and user experience. However, during long-term operation, the dynamic balancing system is prone to imbalance due to continuous vibration and impact or natural aging of components. This can lead to abnormal noise, unexpected interruptions in the spin cycle, and other malfunctions.

[0003] Traditionally, testing for these faults relies on engineers disassembling the washing machine's exterior and manually inspecting the physical condition and connections of each balancing system component. This approach is not only time-consuming, but the frequent disassembly and assembly can easily lead to increased component wear and loose connections, increasing the likelihood of secondary failures.

[0004] Therefore, how to develop a balanced fault detection method to quickly and accurately locate internal component anomalies without disassembling the machine becomes a difficult problem. Summary of the Invention

[0005] The embodiments of the present application provide a balance fault detection method, a balance fault detection device, and a storage medium, which can quickly and accurately locate internal component abnormalities without disassembling the device.

[0006] An embodiment of the present application provides a balancing fault detection method, which is applied to a washing machine. The washing machine includes multiple balancing elements and multiple sensors, each of the balancing elements corresponds to at least one of the sensors. The balancing fault detection method includes:

[0007] triggering an abnormal operating state of the washing machine;

[0008] Acquiring operating status data of a plurality of the balancing elements through a plurality of the sensors;

[0009] Inputting the operating status data into a machine learning model to obtain a fault abnormality degree of each balancing element;

[0010] Outputting a fault probability ranking according to the fault abnormality degree;

[0011] The fault balancing elements are located according to the fault probability ranking.

[0012] In some embodiments, triggering the abnormal operating state of the washing machine includes: placing an eccentric mass with adjustable mass in the washing machine, where the mass of the eccentric mass is determined based on the model or capacity of the washing machine; and starting the washing machine.

[0013] In some embodiments, the placement position of the eccentric block is adjusted according to a preset fault component, and the preset position of the eccentric block includes the front of the inner barrel, the rear of the inner barrel or the side of the inner barrel of the washing machine.

[0014] In some embodiments, after starting the washing machine, it also includes: if the washing machine stops running halfway, gradually reducing the mass of the eccentric block until the washing machine resumes operation, wherein the mass of the eccentric block is greater than a preset mass threshold; if the mass of the eccentric block drops to the preset mass threshold and the test program still cannot be completed, controlling the washing machine to reduce the speed or shorten the running time for repeated start-up; obtaining the operating status data of multiple balancing elements through multiple sensors includes: obtaining the operating status data of multiple balancing elements through the sensors at the moment of each start-up.

[0015] In some embodiments, outputting a fault probability ranking based on the fault abnormality includes: outputting a fault probability ranking based on the fault abnormality in combination with a time weight, a quality weight, a position weight and / or an additional weight; wherein the time weight includes the weight of the accumulated usage time of the balancing element, the quality weight includes the quality weight of the eccentric block, the position weight includes the position weight of the eccentric block, and the additional weight includes an additional weight when the fault abnormality exceeds a historical threshold.

[0016] In some embodiments, outputting a fault probability ranking according to the fault abnormality includes: modifying the fault probability ranking based on fault cases of the same type of washing machines in a historical database, and outputting the fault probability ranking.

[0017] In some embodiments, before inputting the operating status data into a machine learning model to obtain the fault abnormality degree of each balancing element, the method further includes: correcting the operating status data.

[0018] In some embodiments, after locating the faulty balancing element, the method further includes: after replacing or repairing the faulty balancing element, re-executing the balancing fault detection method to generate the fault probability ranking until the highest fault probability in the fault probability ranking is less than a preset safety threshold.

[0019] The present application also provides a balancing fault detection device for a washing machine. The washing machine includes a plurality of balancing elements and a plurality of sensors. Each balancing element corresponds to at least one sensor. The balancing fault detection device includes:

[0020] a trigger module, configured to trigger an abnormal operating state of the washing machine;

[0021] an acquisition module, configured to acquire operating status data of a plurality of the balancing elements through a plurality of the sensors;

[0022] a processing module, configured to input the operating status data into a machine learning model to obtain a fault abnormality degree of each of the balancing elements;

[0023] An analysis module, configured to output a fault probability ranking according to the fault abnormality degree;

[0024] A positioning module is used to locate the fault balancing element according to the fault probability ranking.

[0025] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which is loaded by a processor to execute the steps in the above-mentioned balancing fault detection method.

[0026] In the balancing fault detection method, balancing fault detection device, and storage medium provided in the embodiments of the present application, the execution process of the balancing fault detection method includes the following steps: first, triggering the washing machine to enter an abnormal operating state in a specific manner to simulate or capture potential fault scenarios; then, using multiple sensors to simultaneously collect operating status data of each balancing element in the abnormal operating state; then, inputting these collected operating status data into a pre-trained machine learning model, and calculating the fault abnormality of each balancing element through model analysis; further, ranking the balancing elements by fault probability based on the calculated fault abnormality to intuitively display the possibility of each element failing; finally, accurately locating the faulty balancing element based on the fault probability ranking result. By using the balancing fault detection method proposed in the embodiments of the present application, it is possible to quickly and accurately locate internal component abnormalities without disassembling the washing machine, thereby effectively improving fault troubleshooting efficiency, reducing maintenance costs, and enhancing user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] To more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present application. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0028] Figure 1 This is a first flow chart of the balancing fault detection method provided in an embodiment of the present application.

[0029] Figure 2 This is a schematic diagram of the first placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application.

[0030] Figure 3 This is a schematic diagram of the second placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application.

[0031] Figure 4 This is a schematic diagram of the third placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application.

[0032] Figure 5 It is a structural flow diagram of the balance fault detection device provided in an embodiment of the present application.

[0033] Figure 6 A second flow chart of the balancing fault detection method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0034] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0035] The present invention provides a method, device, and storage medium for detecting a balancing fault, which can quickly and accurately locate an internal component anomaly without disassembling the device. The following describes the method with reference to the accompanying drawings.

[0036] See also Figure 1 , Figure 1 This is a schematic diagram of the first flow chart of a balancing fault detection method provided in an embodiment of the present application. This embodiment of the present application provides a balancing fault detection method applicable to a washing machine. The washing machine includes multiple balancing elements and multiple sensors, with each balancing element corresponding to at least one sensor. This correspondence enables accurate monitoring of the operating status of each balancing element.

[0037] It's worth noting that in washing machines, balancing elements are key components used to maintain the balance of the drum during the spin cycle. For example, a common balancing ring typically contains salt water or a high-density liquid. When the drum spins at high speeds during the spin cycle, the liquid within the balancing ring is redistributed due to centrifugal force, compensating for the unbalanced torque caused by the uneven distribution of clothing within the drum, thereby ensuring smooth operation. There are also balancing weights, typically made of metal. Through appropriate installation and weight design, they help the washing machine achieve dynamic balance during rotation.

[0038] In the washing machine of the present application, the sensor is a device that can sense physical quantities related to the operating state of the balancing element and convert them into electrical signals or other processable signals.

[0039] The balancing fault detection method includes the following steps.

[0040] S100: triggering an abnormal operating state of the washing machine.

[0041] In practical applications, abnormal operating conditions can be triggered by simulating various abnormalities that may occur during the spin cycle. For example, an eccentric weight can be used to adjust the load, causing uneven distribution of clothing within the drum, thereby simulating balance issues caused by this. Alternatively, the speed of the washing machine can be adjusted beyond the normal operating range to observe the machine's operation at abnormal speeds. In this way, data on the operating status of the balancing element under abnormal conditions can be obtained, providing more comprehensive information for subsequent fault detection.

[0042] S200: Acquire operating status data of multiple balancing elements through multiple sensors.

[0043] Because each balancing element corresponds to at least one sensor, these sensors can collect real-time data on various physical quantities during the balancing element's operation. For example, an accelerometer can accurately measure the vibration acceleration of the balancing element during rotation; this data can reflect the balancing element's vibration characteristics. A displacement sensor can record changes in the balancing element's displacement to determine whether it is moving within its normal operating range. A pressure sensor can detect the pressure inside a balancing element (such as a gimbal). If the gimbal exhibits problems such as leakage, the pressure data will change. By combining multiple sensors, comprehensive and accurate data on the balancing element's operating status can be obtained.

[0044] S300: Input the operating status data into a machine learning model to obtain the fault abnormality degree of each balancing element.

[0045] The machine learning model used in this application is trained based on a large amount of data on the operating status of balancing elements in normal and faulty operating states of washing machines. During the training process, the model continuously learns the corresponding relationship between different operating status data and fault abnormality. When the actual collected operating status data is input into the trained machine learning model, the model can automatically analyze and calculate the fault abnormality of each balancing element based on the learned knowledge. For example, for a certain balancing element, if its corresponding sensor data is similar to the characteristics of a large amount of fault data stored in the model, then the fault abnormality of the balancing element will be high.

[0046] S400: Output the fault probability ranking according to the fault abnormality.

[0047] After determining the fault anomaly level for each balancing element, the system ranks them from highest to lowest, generating a failure probability ranking. This ranking method allows maintenance personnel to intuitively understand the probability of failure for each balancing element. For example, if the ranking results show that a balancing element has the highest fault anomaly level, then the probability of failure for that balancing element is the highest, and maintenance personnel can prioritize inspection and repair of that balancing element.

[0048] S500: Locate fault balancing components by sorting by fault probability.

[0049] Maintenance personnel can quickly locate the balancing components most likely to fail based on the failure probability ranking results. This method significantly improves troubleshooting efficiency compared to traditional disassembly inspection methods. For example, during maintenance, maintenance personnel can focus on testing balancing components with a high probability of failure, avoiding blind disassembly and reducing repair time and costs.

[0050] The balancing fault detection method provided in the embodiments of the present application can quickly and accurately locate internal component (balancing element) anomalies without disassembling the device. By analyzing the operating status data collected by sensors through a machine learning model, the failure risk of the balancing element can be assessed more scientifically and accurately.

[0051] In some embodiments, see Figures 2 to 4 , Figure 2 This is a schematic diagram of the first placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application. Figure 3 This is a schematic diagram of the second placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application. Figure 4 This is a schematic diagram of the third placement of the eccentric block during detection in the balance fault detection method provided in an embodiment of the present application.

[0052] In the step of triggering the abnormal operation state of the washing machine 10 , the balance fault detection method includes: placing an eccentric mass 11 with adjustable mass in the washing machine 10 , where the mass of the eccentric mass 11 is determined based on the model or capacity of the washing machine 10 ; and starting the washing machine 10 .

[0053] (1) An eccentric block 11 with adjustable mass is placed at a specific position inside the washing machine 10. The design and placement of the eccentric block 11 is to simulate the uneven distribution of clothes that may occur during the actual use of the washing machine 10, thereby triggering the abnormal operating state of the washing machine 10. The mass of the eccentric block 11 is not set arbitrarily, but is accurately determined based on the model or capacity of the washing machine 10. For different inner barrel volumes of the washing machine 10, the mass of the eccentric block 11 is applicable within a certain range. The greater the mass of the eccentric block 11, the more unstable the washing machine 10 is when running at high speed. When the limit that the balancing element can be adjusted is exceeded, the washing machine 10 will produce vibration, abnormal noise, and displacement. Therefore, in order to avoid risks, the washing machine 10 is generally required to stop actively when dehydrating at a certain eccentric mass.

[0054] Washing machines 10 of different models or capacities have different parameters such as internal structural design, drum size, and dehydration speed. For example, for a washing machine 10 with a larger capacity, its drum can accommodate more clothes and is subjected to greater centrifugal force during dehydration. Therefore, it is necessary to place an eccentric block 11 with a relatively large mass to effectively simulate a severe overload situation; while for a washing machine 10 with a smaller capacity, a similar effect can be achieved by placing an eccentric block 11 with a smaller mass. Through this method of determining the mass of the eccentric block 11 based on the model or capacity of the washing machine 10, it is possible to ensure that abnormal operating conditions can be accurately triggered on washing machines 10 of different models, providing a reliable data basis for subsequent fault detection. For example, for a washing machine 10 with a capacity of 8kg, the mass of the eccentric block 11 is 2kg; or, when the capacity of the washing machine 10 is 10kg, the mass of the eccentric block 11 is 3kg.

[0055] In actual operation, the eccentric mass 11 is usually made of a high-density, wear-resistant material to ensure that it will not be deformed or damaged during the high-speed rotation of the washing machine 10. At the same time, the design of the eccentric mass 11 also takes into account the adaptability to the internal structure of the washing machine 10, so that it can be conveniently placed in a designated position and easily removed when not needed.

[0056] In some cases, the eccentric block 11 can be an electromagnetically driven eccentric block, which realizes its movement and mass adjustment function through an electromagnetic drive system and supports remote control through an intelligent terminal. The intelligent terminal can be a smart phone, a tablet computer or other devices with network connection and data processing functions. Maintenance personnel can establish a communication connection with the washing machine 10 by installing a specific application on the intelligent terminal, so as to remotely control the mass, position and movement trajectory of the eccentric block 11. For example, the user can set the mass parameters of the eccentric block 11 in the application according to the model and capacity of the washing machine 10 and the fault scenario to be simulated, and the system will automatically adjust the mass of the eccentric block 11; at the same time, the user can also accurately control the position of the eccentric block 11 inside the washing machine 10 and place it in a position that can simulate the actual overload situation to the greatest extent.

[0057] The electromagnetically driven eccentric mass 11 has a variety of motion trajectory modes, including spiral descent, horizontal oscillation, or vertical impact. These different motion trajectory modes are intended to more comprehensively and realistically simulate various clothing distribution states that may occur during actual use of the washing machine 10.

[0058] The motion trajectory of the electromagnetically driven eccentric block 11 includes spiral descent, horizontal oscillation or vertical impact modes, which are used to simulate different clothing distribution states. When the eccentric block 11 moves in a spiral descent mode, it can simulate the situation where clothes gradually accumulate on one side in the drum of the washing machine 10. The horizontal oscillation mode is mainly used to simulate the uneven distribution of clothes in the horizontal direction in the drum. The vertical impact mode is used to simulate the situation where clothes suddenly generate vertical impact force on the drum due to certain reasons (such as clothes entanglement, loose parts, etc.) during the dehydration process. A variety of motion trajectory modes can more realistically and comprehensively simulate various clothing distribution states that may occur in actual use of the washing machine 10, so that the collected operating status data is closer to the actual situation, thereby improving the accuracy of fault detection.

[0059] After the eccentric mass 11 is placed, the washing machine 10 is started and enters the normal spin cycle. At this point, due to the presence of the eccentric mass 11, the mass distribution within the drum of the washing machine 10 is no longer uniform, generating an unbalanced torque. During the spin cycle, this unbalanced torque can cause abnormal vibrations and increased noise in the washing machine 10, causing it to enter an abnormal operating state.

[0060] After starting washing machine 10, its control system monitors various operating parameters in real time, such as vibration acceleration and speed fluctuations. Changes in these operating parameters serve as an important basis for subsequent sensor data collection. Triggering abnormal operating conditions in this way can more realistically simulate potential failure scenarios that washing machine 10 may encounter in actual use, making the collected operating status data more representative and valuable.

[0061] The placement position of the eccentric block 11 is adjusted according to the preset fault component. The preset position of the eccentric block 11 includes the front part of the inner drum of the washing machine 10 (such as Figure 2 ), the rear of the inner barrel (such as Figure 3 ) or the inner barrel side (such as Figure 4 ).

[0062] like Figure 2 When the eccentric block 11 is placed at the front of the inner tub, it can simulate the situation where clothes are piled up at the front of the inner tub, which will cause the front of the washing machine 10 to be subjected to a larger centrifugal force during dehydration, thereby generating a larger pressure and unbalanced torque on the front balancing element (such as the front balancing block, etc.).

[0063] For example, the eccentric mass 11 placed at the front of the inner drum can be used to detect faults in the shock-absorbing boom. If there is a fault in the shock-absorbing boom, such as aging of the internal spring, weakened elasticity, or loosening or wear of the connecting parts, it will not be able to withstand this additional force normally. Under the action of the unbalanced torque generated by the eccentric mass 11, the abnormal state of the shock-absorbing boom will be reflected in the operating performance of the washing machine 10. For example, the washing machine 10 may experience abnormal vibrations, increased noise, or even tilting of the inner drum. By collecting these operating status data through sensors and inputting them into the machine learning model for analysis, it is possible to accurately determine whether the shock-absorbing boom is faulty and the severity of the fault.

[0064] like Figure 3 When the eccentric mass 11 is placed at the rear of the inner tub, it can simulate a scenario where clothes accumulate at the rear. For example, during the spin cycle of washing machine 10, if clothes accumulate at the rear for some reason (e.g., limitations of the washing machine 10's internal structure, clothing material characteristics, etc.), the rear balancing elements (e.g., rear balancing rings, balancing weights, etc.) can be subjected to abnormal loads. This simulation method can be used to test the operation of the rear balancing elements of the inner tub under large unbalanced torques, allowing for the timely identification of potential faults.

[0065] For example, the eccentric weight 11 placed on the side of the inner barrel can be used to detect gimbal jams. If the gimbal is jammed, such as when the internal liquid flow is poor, the gimbal movement is blocked, or the connection between the gimbal and the inner barrel becomes stuck, it will not be able to respond promptly and effectively to the lateral unbalanced torque. Under the action of the eccentric weight 11, the gimbal jamming will cause the inner barrel to rotate unsteadily, resulting in periodic vibration and shaking. Sensors collect this vibration and shaking data in real time, and the machine learning model analyzes this data to accurately identify gimbal jams.

[0066] like Figure 4 The balancing elements on the sides of the inner drum play a key role in maintaining the lateral balance of the washing machine 10's rotating drum. Placing the eccentric weight 11 on the sides of the inner drum can simulate the situation where the clothes are offset laterally. In actual use, if the clothes are unevenly distributed in the washing machine 10's rotating drum, resulting in significant mass differences on the sides, the balancing elements on the sides will be subjected to additional forces and moments. By placing the eccentric weight 11 in this position, the performance of the balancing elements on the sides of the inner drum in dealing with lateral imbalance can be tested to determine whether they are functioning properly.

[0067] For example, the eccentric block 11 placed at the rear of the inner barrel can be used to detect the looseness of the motor bracket. If the motor bracket is loose, such as the fixing bolts are loose, there is a gap between the bracket and the outer shell of the washing machine 10, etc., the motor cannot be firmly fixed. Under the action of the unbalanced torque generated by the eccentric block 11, the loosening of the motor bracket will cause slight changes in the position of the motor, thereby affecting the rotation accuracy and stability of the motor. The washing machine 10 will experience unstable speed, increased vibration, abnormal noise, etc. After the sensor collects these operating data, the machine learning model can accurately analyze the loosening fault of the motor bracket.

[0068] The balancing fault detection method of this application offers the following advantages: It is highly targeted and comprehensive. Maintenance personnel can precisely adjust the position of the eccentric weight 11 based on the pre-determined fault component, allowing the simulated abnormal operating state to more directly affect the target balancing component, thereby improving the targeted and effective nature of fault detection. The diverse pre-determined positions can simulate uneven clothing distribution in different areas, comprehensively testing the performance and status of each balancing component within the washing machine 10 and identifying more potential fault hazards.

[0069] In some embodiments, after the step of starting the washing machine, the balancing fault detection method further includes the following steps.

[0070] If the washing machine stops running, the mass of the eccentric block is gradually reduced until the washing machine resumes running, wherein the mass of the eccentric block is greater than a preset mass threshold.

[0071] Washing machines often stop running mid-cycle because the unbalanced torque generated by the eccentric mass is too great to withstand, triggering the machine's internal balancing components. By gradually reducing the mass of the eccentric mass, the unbalanced torque can be effectively reduced, improving the internal stress conditions and potentially resuming operation.

[0072] Throughout the adjustment process, the eccentric mass remains above a preset mass threshold. This threshold was determined through extensive experimentation and data analysis, taking into account factors such as the washing machine model and specifications, the load capacity of the balancing components, and testing requirements. Maintaining the eccentric mass above this threshold ensures that sufficient unbalanced torque can be generated during the adjustment process to effectively trigger abnormal operating conditions in the washing machine, providing valuable data for subsequent fault detection.

[0073] The mass reduction step of the eccentric block can be 5% to 20% of the initial mass. This range can achieve a good balance between adjustment accuracy and detection efficiency, ensuring both adjustment accuracy and improved detection efficiency.

[0074] If the test procedure still fails even after the eccentric mass has dropped to the preset mass threshold, the washing machine is controlled to restart repeatedly at a reduced speed or shortened run time. Reducing the speed can reduce the unbalanced torque and centrifugal force generated during the washing machine's rotation, alleviating the burden on the balancing elements. For example, reducing the speed to 30%-50% of the rated speed, with a single run time of no more than 10 seconds. Shortening the run time can reduce the time the washing machine operates in an abnormal state, reducing the potential risk of damage to the washing machine's internal components. This control method can both reduce the difficulty of testing to a certain extent and continue to obtain operating status data for the balancing elements, facilitating fault detection. In this case, the step of obtaining operating status data for multiple balancing elements using multiple sensors includes obtaining operating status data for multiple balancing elements using sensors at the moment of each startup.

[0075] By acquiring data at each startup moment (the data at the startup moment includes at least one of the current transient waveform, vibration peak value or displacement mutation), a series of continuous and time-correlated data samples can be formed. These data samples not only contain the operating status information of the balancing element during the normal startup process, but also record the abnormal performance under the influence of the eccentric block. Through the comprehensive analysis of these data, the machine learning model can more accurately identify the fault characteristics of the balancing element and improve the accuracy and efficiency of fault detection. At the same time, this data collection method also provides comprehensive and detailed data support for subsequent fault diagnosis and maintenance, which helps maintenance personnel to locate faulty components more quickly and accurately and repair them.

[0076] The sensor in this application can also obtain the instantaneous current of multiple balancing elements. Taking the driving components in the balancing ring (if the balancing ring has a driving function) and the motor connected to the motor bracket as an example, at the moment of startup, due to the unbalanced torque generated by the eccentric block, these components need to overcome additional loads to start and run. At this time, the current required by the balancing element will increase instantaneously, and the magnitude and change of the current can directly reflect the load status and operating performance of the component. If there is a fault in the balancing element, such as internal short circuit, poor contact, mechanical jamming, etc., it will cause the instantaneous current to be abnormal, such as excessive current, too small current, severe fluctuations, etc. By obtaining instantaneous current data and comparing and analyzing it with the current data under normal conditions, the health status of the balancing element can be evaluated more comprehensively and accurately, providing important supplementary information for fault detection.

[0077] In some embodiments, in the step of outputting a fault probability ranking based on the fault abnormality, the balanced fault detection method includes: outputting a fault probability ranking based on the fault abnormality in combination with a time weight, a quality weight, a location weight, and / or an additional weight. Specifically, in this step, the fault abnormality, time weight, quality weight, location weight, and / or additional weight are input into a machine learning model to generate the fault probability ranking. The machine learning model in the embodiments of the present application is a deep neural network analysis model deployed in the cloud.

[0078] The time weighting factor includes the cumulative usage time of the balancing components. Over the long-term use of a washing machine, each balancing component will gradually wear out and age, and the probability of failure will increase accordingly. Therefore, incorporating cumulative usage time into the weighting factor can more accurately reflect the potential failure risk of balancing components due to their long-term use.

[0079] The mass weight includes the mass of the eccentric mass, which directly affects the balance state and stress conditions during operation. Eccentric masses of varying masses exert varying degrees of force on the balancing element, affecting its workload and risk of failure.

[0080] Position weights include the position weight of the eccentric mass. Balancing elements at different locations within the washing machine may experience different forces and operating environments during operation. For example, balancing elements near the washing machine's rotational center and those further away may experience different forces and vibration characteristics under the same unbalanced load, resulting in different failure probabilities.

[0081] Additional weighting includes an additional weighting when the fault anomaly exceeds a historical threshold. This additional weighting is primarily used to address situations where the fault anomaly exceeds a historical threshold. When the fault anomaly of a balancing component exceeds a pre-set threshold, it indicates a potentially serious fault and warrants increased attention. In this case, adding an additional weight to the component will give it a higher priority in the fault probability ranking, allowing maintenance personnel to prioritize inspection and repairs.

[0082] In some embodiments, in the step of outputting the fault probability ranking according to the fault abnormality, the balanced fault detection method includes: correcting the fault probability ranking based on fault cases of the same type of washing machines in a historical database, and outputting the fault probability ranking.

[0083] The historical database stores a large number of failure cases that occurred during the past operation of similar washing machines. These cases cover various fault types, occurrence conditions, and corresponding repair plans. By comparing and analyzing the abnormality of the fault detected in the current washing machine with the cases in the historical database, we can identify correlations and similarities between the two and then refine the initial fault probability ranking. This correction mechanism fully utilizes the value of historical data, making fault diagnosis results more accurate and effective, and improving the scientific nature and effectiveness of repair decisions.

[0084] The correction method can be a probability weight adjustment based on the number of similar failure cases.

[0085] For example, check whether the current washing machine model has more than 50 similar failure cases in the historical database. If so, this indicates that this type of failure is common and regular for this model. In this case, the probability weight of this type of failure is increased by 20%-50%. The specific increase can be set based on actual conditions. For example, the increase can be larger when there are more similar failure cases or when the impact of this type of failure on the washing machine's performance is more severe.

[0086] By increasing the probability weight, this type of fault will be significantly ranked higher in the fault probability ranking. This means that maintenance personnel will pay more attention to this type of fault during troubleshooting, giving it priority for inspection and repair. This approach allows for quicker identification of common faults and improves maintenance efficiency.

[0087] The correction method can also be to quote the maintenance plan based on the similarity of the fault abnormality.

[0088] By comparing the feature vector of the current fault abnormality with the feature vector of the historical case, the similarity value between the two is calculated. If the similarity between the current fault abnormality and a historical case exceeds 90%, the two are considered to have a high degree of similarity.

[0089] When a historical case with a similarity of more than 90% is found, the system will directly reference the maintenance plan corresponding to that historical case. This is because in highly similar situations, the current fault is likely to be of the same type as the fault in the historical case, and the same maintenance plan can effectively solve the problem.

[0090] In some embodiments, before the step of inputting the operating status data into the machine learning model to obtain the fault abnormality degree of each balancing element, the balancing fault detection method further includes correcting the operating status data.

[0091] Ways to correct the operating status data include eliminating noise interference or performing compensation processing.

[0092] During washing machine operation, the collected operating status data is often contaminated by a significant amount of noise due to various factors, including mechanical vibration, electromagnetic interference, and ambient noise. This noise can seriously interfere with the authenticity and accuracy of the data, and thus affect fault detection results. To effectively eliminate this noise interference, this application can employ wavelet transforms or Kalman filtering algorithms.

[0093] Wavelet transform is a time-frequency analysis method that can perform threshold processing on the wavelet coefficients of the high-frequency part by setting an appropriate threshold, that is, suppressing or removing the wavelet coefficients corresponding to the noise, while retaining or appropriately enhancing the effective signal coefficients of the low-frequency part. Finally, an inverse wavelet transform is performed to obtain the denoised operating status data.

[0094] The Kalman filter is a recursive filtering algorithm based on a state-space model. It recursively calculates the optimal state estimate for the current moment based on the state estimate at the previous moment and the observed value at the current moment. During this process, the algorithm automatically suppresses noise in the observed values, resulting in a more accurate estimate of the operating state data.

[0095] The compensation process may include at least one of ground tilt compensation, voltage fluctuation compensation, and ambient temperature compensation.

[0096] Floor tilt compensation uses a built-in gyroscope to detect the washing machine's tilt angle and correct the vibration data amplitude. Specifically, by establishing a mathematical model between the tilt angle and the vibration data amplitude, using the tilt angle as an input parameter, the deviation in the vibration data amplitude caused by floor tilt is calculated. This deviation is then subtracted from the collected vibration data amplitude to obtain the corrected vibration data amplitude. After floor tilt compensation, the vibration data can more accurately reflect the actual vibration conditions inside the washing machine caused by factors such as unbalanced loads, thereby improving data reliability.

[0097] Voltage fluctuation compensation adjusts the current data's baseline value by monitoring the grid voltage in real time. Specifically, during normal operation of the washing machine, current data from the motor's stable operation at different voltages is collected in advance. Through data analysis, a voltage-current baseline curve or model is established. During actual operation, the grid voltage is monitored in real time, and the corresponding current baseline value for the current voltage is determined based on this curve or model. The actual current data collected is then compared with this baseline value to calculate the current data deviation. Finally, the collected current data is adjusted using this baseline value as a reference to obtain compensated current data. After voltage fluctuation compensation, the current data more accurately reflects the operating status of the washing machine motor under actual load, improving data accuracy.

[0098] In order to eliminate the influence of ambient temperature on operating status data, the present application monitors the ambient temperature in real time by setting a temperature sensor inside the washing machine. Based on the correction relationship between the ambient temperature and the operating status data obtained in advance through experiments or theoretical analysis, the collected operating status data is compensated. For example, for vibration data, the relationship between the output signal of the vibration sensor and the actual vibration at different temperatures is determined through experiments, and a temperature-vibration correction coefficient table is established. During actual operation, according to the ambient temperature monitored in real time, the corresponding correction coefficient is found in the correction coefficient table, and the collected vibration data is multiplied by the correction coefficient to obtain the compensated vibration data. Through ambient temperature compensation, the interference of ambient temperature changes on the operating status data can be reduced, and the stability and reliability of the data can be improved.

[0099] The embodiment of the present application further proposes a specific solution of weighting according to the degree of fault abnormality. Therefore, the step of outputting the fault probability ranking according to the degree of fault abnormality further includes the following steps.

[0100] Compare the fault abnormality degree h of each balancing element with the normal threshold x of the corresponding position in the cloud database.

[0101] After obtaining the fault anomaly degree h for each balancing element, it is compared with the normal threshold value x at the corresponding location in the cloud database. The cloud database stores a large amount of normal operating data and related parameters for the same model of washing machine. By analyzing and processing this data, the normal threshold value x for balancing elements at different locations can be determined.

[0102] If h>x, it is determined that the balancing element is abnormal and an additional weight is added to the element.

[0103] At this point, in order to highlight the fault risk of the component, an additional weight is added to it. The size of the additional weight can be dynamically adjusted according to the degree to which the fault abnormality exceeds the normal threshold. For example, the greater the abnormality exceeds the normal threshold, the greater the additional weight.

[0104] If h≤x, it is determined that there is no abnormality and weights are evenly assigned to all relevant balancing elements.

[0105] The specific rules are as follows: If h is ≤ x for all relevant balancing elements, the total weight is divided equally among the number of elements. For example, if there are three relevant balancing elements with a total weight of 3, each element is assigned a weight of 1. If h is greater than x for some elements, only those elements with h ≤ x are equally weighted. This ensures that the weight distribution of normal elements is not affected by abnormal elements, while also ensuring the rationality of the fault probability ranking.

[0106] The normal threshold x is dynamically generated using the following methods: One method involves extracting historical sensor data from a cloud database for the same model of washing machine under normal conditions. This data, which includes various parameters of the washing machine under different operating conditions, such as vibration, current, and temperature, serves as an important basis for analyzing the normal state of the balancing element. Another method involves calculating the range of characteristic values ​​for each balancing element within a preset confidence interval based on statistical distribution, with the upper limit taken as x. For example, for the vibration amplitude characteristic value of a balancing element under normal conditions, a reasonable confidence interval is determined by calculating its statistical distribution, and the upper limit of this interval is used as the normal threshold for that element. This means that when the fault abnormality of a balancing element exceeds this upper limit, it can be considered that its operating state has deviated from the normal range and a fault may exist. This method of dynamically generating the normal threshold fully accounts for individual differences between washing machines and changes in the operating environment, making the normal threshold more accurate.

[0107] In some embodiments, after the step of locating the faulty balancing element, the balancing fault detection method further includes: after replacing or repairing the faulty balancing element, re-executing the balancing fault detection method to generate a fault probability ranking until the highest fault probability in the fault probability ranking is less than a preset safety threshold.

[0108] If the highest failure probability is less than the preset safety threshold, the repair is deemed successful and the washing machine can resume normal operation; if the highest failure probability is still higher than the preset safety threshold, it means that the repair effect has not met expectations and there may be other potential problems that need further investigation.

[0109] If the highest failure probability after retesting remains above the preset safety threshold, the system will upload all relevant data detected so far, including the operating status of each balancing component, the degree of fault anomaly, and the ranking of fault probabilities, to a cloud database. This cloud database, with its powerful storage and analysis capabilities, enables long-term storage and in-depth analysis of this data, providing a reference for subsequent fault diagnosis and repair. At the same time, the system will trigger a manual intervention command, which notifies maintenance personnel or relevant technicians to conduct further inspection and repair of the washing machine.

[0110] The machine learning model is a simple neural network classification model deployed in the cloud, and its input data includes at least one of time domain waveform, frequency domain spectrum and time-frequency joint features.

[0111] The time-domain waveform directly shows how the sensor's collected signal changes over time. By analyzing the time-domain waveform, we can observe characteristics such as the signal's amplitude, period, and trend. These characteristics can reflect the dynamic behavior of the washing machine's balancing element during operation.

[0112] The frequency domain spectrum is a spectrum diagram obtained by converting the time domain signal to the frequency domain through methods such as Fourier transform. The frequency domain spectrum can clearly show the distribution of different frequency components in the signal.

[0113] Joint time-frequency features combine information from both the time and frequency domains to more comprehensively describe signal characteristics. Time-frequency analysis methods (such as short-time Fourier transform and wavelet transform) can decompose signals into two dimensions: time and frequency, to determine the energy distribution of the signal at different times and frequencies.

[0114] The training data of the simple neural network classification model includes multi-dimensional sensor data under normal conditions, artificially simulated fault scenario data, and actual fault case data in historical maintenance records.

[0115] Multi-dimensional sensor data under normal conditions forms one of the foundational data sets for model training. This data consists of measurements of various physical quantities, such as vibration acceleration, displacement, and rotational speed, collected by multiple sensors during the normal operation of the washing machine. By analyzing and learning from this large amount of data under normal conditions, a simple neural network classification model can establish an accurate understanding of the normal operation of the washing machine, serving as a reference for determining fault conditions.

[0116] To enable the model to accurately identify various fault types, this application specifically constructs artificially simulated fault scenario data. This data is obtained by simulating various possible fault conditions of washing machines in an experimental environment. Specific scenarios include but are not limited to the following.

[0117] Damping rod fracture: The damping rod is a critical component in a washing machine. When it breaks, the machine's damping effect is significantly reduced, leading to increased vibration. By simulating a damping rod fracture scenario and collecting corresponding sensor data, the model can learn the characteristic behaviors of this fault condition.

[0118] Gimbal stuck: The gimbal is responsible for maintaining the balance of the washing machine during operation. If the gimbal becomes stuck, the machine's balance will be disrupted, causing abnormal vibration and noise. Simulating a gimbal stuck scenario and collecting data can help the model identify this type of fault.

[0119] Counterweight loss: Counterweights are used to adjust the center of gravity of a washing machine to ensure stable operation. A falling counterweight can cause the washing machine's center of gravity to shift, leading to unbalanced vibration. By simulating a falling counterweight, we provide the model with learning samples for this fault type.

[0120] Loose motor brackets: Loose motor brackets can affect the stability of the motor's installation, leading to abnormal vibration and noise during operation. Simulating loose motor brackets and collecting relevant data can enrich the model's fault signature library.

[0121] Other scenarios: In addition to the common scenarios mentioned above, you can also simulate other possible fault scenarios based on actual needs, such as bearing wear and loose belts, to further improve the model's fault identification capabilities.

[0122] Historical maintenance records contain a large amount of data on actual washing machine failures. This data, verified through actual repairs, is highly authentic and reliable. By analyzing and organizing this data, extracting relevant sensor data and fault characteristics and incorporating them into the training dataset, the model can be better adapted to real-world application scenarios, improving the accuracy and practicality of fault diagnosis.

[0123] See also Figure 5 , Figure 5 It is a structural flow diagram of the balance fault detection device provided in an embodiment of the present application.

[0124] In some scenarios, the balancing fault detection method includes the following steps.

[0125] Start the washing machine and place the eccentric mass according to the preset rules to trigger possible abnormal operating conditions.

[0126] During the operation of the washing machine, multiple sensors are used to obtain the operating status data of multiple balancing elements. These data will serve as the basis for subsequent input into the machine learning model.

[0127] If the washing machine stops running, the mass of the eccentric block is gradually reduced until the washing machine resumes running, wherein the mass of the eccentric block is greater than a preset mass threshold.

[0128] The acquired operating status data is corrected, including noise elimination and compensation, to obtain accurate operating status data. The corrected data can more realistically reflect the operating status of the balancing element, improving the accuracy of subsequent fault diagnosis.

[0129] The corrected operating status data is fed into a simple neural network classification model deployed in the cloud. Based on at least one of the input time-domain waveform, frequency-domain spectrum, and time-frequency joint features, combined with the normal state characteristics and various fault characteristics learned during training, the model automatically calculates the fault anomaly degree h for each balancing element.

[0130] Output the fault probability ranking according to the fault abnormality: compare the fault abnormality h of each balancing element with the normal threshold x of the corresponding position in the cloud database.

[0131] If h>x, the balancing element is determined to be abnormal and an additional weight is added to the element.

[0132] If h≤x, it is determined that there is no abnormality and the weights are evenly distributed to all relevant balancing elements.

[0133] The deep neural network analysis model deployed in the cloud comprehensively considers the fault abnormality, time weight, quality weight, location weight and / or additional weight, and outputs the fault probability ranking.

[0134] The faulty balancing component is located by ranking by probability of failure. Based on the ranking results, maintenance personnel can dismantle and inspect the washing machine in a targeted manner to accurately identify the faulty balancing component and perform appropriate repairs or replacements.

[0135] Re-execute the balancing fault detection method (if necessary). After completing the replacement or repair of the faulty balancing element, re-acquire the operating status data and perform correction processing.

[0136] The corrected data is input into a simple neural network classification model on the cloud to generate a new fault probability ranking.

[0137] Verification and feedback (if necessary): Compare the highest failure probability in the new failure probability ranking with the preset safety threshold.

[0138] If the highest failure probability is less than the preset safety threshold, the repair is determined to be successful and the washing machine resumes normal operation.

[0139] If the highest failure probability is still higher than the preset safety threshold, the current data will be uploaded to the cloud database, and a manual intervention instruction will be triggered to notify maintenance personnel to conduct further inspection and repairs.

[0140] See also Figure 6 , Figure 6This is a second flow diagram of the balancing fault detection method provided in an embodiment of the present application. This embodiment of the present application also provides a balancing fault detection device 20 for use in a washing machine. The washing machine includes multiple balancing elements and multiple sensors, each balancing element corresponding to at least one sensor. The balancing fault detection device 20 includes a trigger module 201, an acquisition module 202, a processing module 203, an analysis module 204, and a positioning module 205.

[0141] The trigger module 201 is used to trigger an abnormal operation state of the washing machine.

[0142] The acquisition module 202 is configured to acquire operating status data of a plurality of balancing elements through a plurality of sensors.

[0143] The processing module 203 is used to input the operating status data into the machine learning model to obtain the fault abnormality degree of each balancing element.

[0144] The analysis module 204 is configured to output a fault probability ranking according to the fault abnormality degree.

[0145] The positioning module 205 is used to locate the faulty balancing element by sorting the fault probabilities.

[0146] An embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. The computer program is loaded by a processor to execute the steps in the above-mentioned balancing fault detection method.

[0147] It should be noted that, for the balancing fault detection method of the embodiment of the present application, a person skilled in the art will understand that all or part of the process of the balancing fault detection method of the embodiment of the present application can be achieved by controlling the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, such as a memory, and executed by at least one processor. During the execution process, it may include the process of the embodiment of the balancing fault detection method. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), etc. The above describes in detail the balancing fault detection method, device, storage medium, and washing machine provided in the embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and its core concept of the present application. At the same time, for those skilled in the art, based on the concept of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present application.

[0148] An embodiment of the present application provides a washing machine, which includes multiple balancing elements, multiple sensors and a processor. Each of the balancing elements corresponds to at least one of the sensors. The processor is configured to execute the above-mentioned balancing fault detection method by calling a computer program stored in a memory.

[0149] The processor may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor may be implemented in at least one of the following hardware forms: a DSP (Digital Signal Processing), an FPGA (Field-Programmable Gate Array), or a PLA (Programmable Logic Array).

[0150] The processor may also include a main processor and a coprocessor. The main processor is a processor for processing data in the awake state, also known as a CPU (Central Processing Unit); the coprocessor is a low-power processor for processing data in the standby state.

[0151] In some embodiments, the processor may be integrated with a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content required to be displayed on the display screen. The processor may also include an AI (Artificial Intelligence) processor, which is used to handle the control method operations related to the balance fault detection system, allowing the model in the balance fault detection system to autonomously train and learn, improving efficiency and accuracy.

[0152] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0153] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Therefore, features specified as "first" or "second" may explicitly or implicitly include one or more features.

[0154] The above describes in detail the balancing fault detection method, balancing fault detection device, and storage medium provided in the embodiments of the present application. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is intended only to facilitate understanding of the present application. Furthermore, those skilled in the art will appreciate that variations in the specific implementation methods and scope of application may occur based on the concepts of the present application. In summary, the contents of this specification should not be construed as limiting the present application.

Claims

1. A method for detecting a balancing fault, characterized in that: Applied to a washing machine, the washing machine includes a plurality of balancing elements and a plurality of sensors, each balancing element corresponds to at least one of the sensors, and the balancing fault detection method includes: triggering an abnormal operating state of the washing machine; Acquiring operating status data of a plurality of the balancing elements through a plurality of the sensors; Inputting the operating status data into a machine learning model to obtain a fault abnormality degree of each balancing element; Outputting a fault probability ranking according to the fault abnormality degree; The fault balancing elements are located according to the fault probability ranking.

2. The balance fault detection method according to claim 1, characterized in that: Triggering the abnormal operating state of the washing machine includes: placing an eccentric mass with adjustable mass in the washing machine, where the mass of the eccentric mass is determined based on the model or capacity of the washing machine; and starting the washing machine.

3. The balance fault detection method according to claim 2, characterized in that: The placement position of the eccentric block is adjusted according to a preset fault component, and the preset position of the eccentric block includes the front of the inner barrel, the rear of the inner barrel or the side of the inner barrel of the washing machine.

4. The method for detecting a balancing fault according to claim 2, wherein: After starting the washing machine, it also includes: if the washing machine stops running halfway, gradually reducing the mass of the eccentric block until the washing machine resumes operation, wherein the mass of the eccentric block is greater than a preset mass threshold; if the mass of the eccentric block drops to the preset mass threshold and the test program still cannot be completed, controlling the washing machine to reduce the speed or shorten the running time for repeated start-up; obtaining the operating status data of multiple balancing elements through multiple sensors includes: obtaining the operating status data of multiple balancing elements through the sensors at the moment of each start-up.

5. The method for detecting a balancing fault according to claim 2, wherein: Outputting a fault probability ranking according to the fault abnormality includes: outputting a fault probability ranking according to the fault abnormality in combination with a time weight, a quality weight, a position weight and / or an additional weight; wherein the time weight includes the weight of the accumulated usage time of the balancing element, the quality weight includes the quality weight of the eccentric block, the position weight includes the position weight of the eccentric block, and the additional weight includes an additional weight when the fault abnormality exceeds a historical threshold.

6. The method for detecting a balance failure according to any one of claims 1 to 4, characterized in that: Outputting a fault probability ranking according to the fault abnormality includes: correcting the fault probability ranking based on fault cases of the same type of washing machines in a historical database, and outputting the fault probability ranking.

7. The balance fault detection method according to any one of claims 1 to 4, characterized in that: Before inputting the operating status data into a machine learning model to obtain the fault abnormality degree of each balancing element, the method further includes: correcting the operating status data.

8. The balance fault detection method according to any one of claims 1 to 4, characterized in that: After locating the faulty balancing element, the method further includes: after replacing or repairing the faulty balancing element, re-executing the balancing fault detection method to generate the fault probability ranking until the highest fault probability in the fault probability ranking is less than a preset safety threshold.

9. A balance fault detection device, characterized in that: Applied to a washing machine, the washing machine includes a plurality of balancing elements and a plurality of sensors, each of the balancing elements corresponds to at least one of the sensors, and the balancing fault detection device includes: a trigger module, configured to trigger an abnormal operating state of the washing machine; an acquisition module, configured to acquire operating status data of a plurality of the balancing elements through a plurality of the sensors; a processing module, configured to input the operating status data into a machine learning model to obtain a fault abnormality degree of each of the balancing elements; An analysis module, configured to output a fault probability ranking according to the fault abnormality degree; A positioning module is used to locate the fault balancing element according to the fault probability ranking.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps in the balance fault detection method according to any one of claims 1 to 8.

Images