Method for predicting service life of tool bit of kitchen waste processor by using deep learning algorithm

Abstract

The invention discloses a method for predicting the service life of a tool bit of a kitchen waste processor by using a deep learning algorithm, and belongs to the technical field of data processing, and the method comprises the steps: obtaining working condition parameters of the kitchen waste processor, and generating a standardized data set containing the working condition parameters; vibration signal frequency spectrum features and motor current harmonic features are extracted from the standardized data set, and a multi-dimensional feature set is constructed; inputting the multi-dimensional feature set into a pre-trained long-short-term memory network model, and outputting an evaluation value of the wear degree of the tool bit; and vibration signal frequency spectrum features and motor current harmonic features corresponding to the evaluation value of the tool bit wear degree exceeding a preset threshold are extracted, historical records are matched according to the multi-dimensional feature set, and a sliding window weighting algorithm is adopted to calculate a residual life prediction value. The method for predicting the service life of the tool bit of the kitchen waste processor by using the deep learning algorithm solves the problem that the prediction accuracy of the service life of the tool bit is low due to the fact that the service life of the tool bit is difficult to predict through association between working condition parameters and residual service life at present.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm. Background Art

[0002] In the field of smart home appliances, kitchen garbage disposers are essential devices that enhance quality of life, and their lifespan directly impacts user experience and device safety. With the widespread adoption of intelligent technology, achieving predictive maintenance of these devices through data-driven approaches has become a critical issue in this field. The wear of the cutter head, one of the core components of a kitchen garbage disposer, directly determines its operating efficiency and service life. Therefore, accurately predicting the lifespan of the cutter head is crucial. However, current methods for predicting cutter head lifespan often rely on traditional periodic inspections or simple threshold judgments. These methods often fail to accurately capture the dynamic changes in cutter head wear and struggle to adapt to the complex operating conditions of various usage scenarios. This results in inaccurate predictions, potentially allowing the device to operate undetected with a defect, or even posing a safety hazard. Therefore, current methods for predicting the lifespan of kitchen garbage disposers struggle to predict the lifespan of the cutter head by correlating operating parameters with the remaining lifespan, resulting in low cutter head lifespan prediction accuracy. Summary of the Invention

[0003] In order to overcome the defects of the prior art, the present invention provides a method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm to solve the above problems.

[0004] The technical solution adopted by the present invention to solve the technical problem is: a method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm, comprising the following steps: S1: Acquire real-time collected operating parameters of a kitchen garbage disposer and generate a standardized data set containing the operating parameters, wherein the operating parameters include motor operating state data and cutter head vibration frequency data; S2: Performing a time-frequency joint analysis on the standardized data set to extract the vibration signal spectrum features and motor current harmonic features, and constructing a multidimensional feature set; S3: Inputting the multidimensional feature set including the vibration signal spectrum features and the motor current harmonic features into a pre-trained long short-term memory network model, and outputting an evaluation value of the degree of tool head wear; S4: Extract the vibration signal spectrum characteristics and motor current harmonic characteristics corresponding to the evaluation value of the tool head wear degree exceeding the preset threshold, match the historical records based on the multidimensional feature set, and use the sliding window weighted algorithm to calculate the remaining life prediction value.

[0005] Specifically, in step S1, after the operating parameters are acquired, noise elimination and outlier processing are performed on the operating parameters to generate a standardized data set including motor operating status data and tool head vibration frequency data with time stamp alignment.

[0006] Preferably, in step S2, short-time Fourier transform is used to perform time-frequency analysis on the tool head vibration frequency data in the standardized data set, and the window function is set to a Hanning window with a window length of a preset sampling point to obtain a time-frequency distribution matrix; The time-frequency distribution matrix is ​​divided into multiple frequency bands for energy integration, and the energy proportion of each frequency band is calculated as a preliminary spectrum feature; From the preliminary spectrum characteristics, a core vibration characteristic frequency band is extracted using wavelet decomposition technology, and the average value and variance of the core vibration characteristic frequency band are calculated to form the vibration signal spectrum characteristics.

[0007] Optionally, in step S2, a fast Fourier transform is performed on the motor current signal in the motor operating status data in the standardized data set, multiple harmonic components are analyzed, and the amplitude ratio of each harmonic is calculated to obtain the motor current harmonic characteristics.

[0008] It is worth noting that in step S2, based on the vibration signal spectrum characteristics and the motor current harmonic characteristics, the timestamps are aligned and a horizontal splicing method is used to generate a multidimensional feature set, which includes the vibration signal spectrum characteristics and the motor current harmonic characteristics.

[0009] Specifically, in step S3, the multidimensional feature data set is processed using a standardization method to obtain a standardized feature data set; The standardized feature data set is input into a pre-trained long short-term memory network model, and a time series analysis is performed on the standardized feature data set to obtain an evaluation value of the degree of tool head wear.

[0010] Preferably, in step S4, the step of matching historical maintenance records according to the multidimensional feature set includes: extracting vibration signal spectrum features and motor current harmonic features corresponding to the evaluation value of the tool head wear degree exceeding a preset threshold from the multidimensional feature set to generate an abnormal feature data set; Historical records are retrieved from a pre-established database, where the historical records include historical data. A similarity matching method is used to determine historical data that matches the abnormal feature data set as a historical data subset, where the historical data subset includes historical vibration signal spectrum features and historical motor current harmonic features.

[0011] It is worth noting that, in step S4, the step of calculating the remaining life prediction interval using the sliding window weighted algorithm includes: for the matched historical data subset, segmenting the features in the historical data using a sliding window method to obtain a segmented feature sequence; According to the segmented feature sequence, a weighted feature value corresponding to each segmented feature sequence is calculated using a weighted method according to the weights of the historical vibration signal spectrum characteristics and the historical motor current harmonic characteristics in the historical data subset; Using a time series analysis method to process the timestamp data corresponding to the weighted feature value and the corresponding segmented feature sequence to obtain a feature change trend parameter; The historical records also include historical maintenance records, and the weighted characteristic values ​​when the cutter head is replaced in the historical maintenance records are obtained. The time required to change from the current weighted characteristic value to the weighted characteristic value when the cutter head is replaced in the historical maintenance records is calculated according to the characteristic change trend parameters as the remaining life prediction value.

[0012] The beneficial effects of the present invention are as follows: in the method for predicting the life of the kitchen garbage disposer blade using a deep learning algorithm, the operating parameters of the processor, including the motor operating status and blade vibration frequency data, are collected in real time, the vibration signal spectrum characteristics and motor current harmonic characteristics are extracted, and a multidimensional feature set is constructed. The multidimensional feature set is input into a pre-trained long short-term memory network model, and an evaluation value of the degree of blade wear is output. For wear evaluation values ​​that exceed the threshold, a sliding window weighted algorithm is used to predict the remaining life prediction value in combination with historical maintenance records. The present invention predicts the blade life through the correlation between operating parameters and remaining life through multidimensional data analysis and deep learning methods, thereby achieving accurate evaluation of the wear status of the kitchen garbage disposer blade and reliable prediction of the remaining life, which helps in timely maintenance and replacement, and improves the operating efficiency and safety of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a flow chart of a method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm in one embodiment of the present invention; Figure 2 A step-by-step flow chart of step S2 in one embodiment of the present invention; Figure 3 FIG. 4 is a step-by-step flow chart of step S4 in one embodiment of the present invention. DETAILED DESCRIPTION

[0014] The following is a further description of specific embodiments of the present invention in conjunction with the accompanying drawings. It should be noted that the description of these embodiments is intended to facilitate understanding of the present invention and does not constitute a limitation of the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0015] like Figure 1-3 As shown, a method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm includes the following steps: S1: Acquire real-time collected operating parameters of a kitchen garbage disposer and generate a standardized data set containing the operating parameters, wherein the operating parameters include motor operating state data and cutter head vibration frequency data; S2: Performing a time-frequency joint analysis on the standardized data set to extract the vibration signal spectrum features and motor current harmonic features, and constructing a multidimensional feature set; S3: Inputting the multidimensional feature set including the vibration signal spectrum features and the motor current harmonic features into a pre-trained long short-term memory network model, and outputting an evaluation value of the degree of tool head wear; S4: Extract the vibration signal spectrum characteristics and motor current harmonic characteristics corresponding to the evaluation value of the tool head wear degree exceeding the preset threshold, match the historical records based on the multidimensional feature set, and use the sliding window weighted algorithm to calculate the remaining life prediction value.

[0016] In the method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm, the operating parameters of the processor, including the motor operating status and blade vibration frequency data, are collected in real time, the vibration signal spectrum characteristics and motor current harmonic characteristics are extracted, and a multidimensional feature set is constructed. The multidimensional feature set is input into a pre-trained long short-term memory network model, and an evaluation value of the degree of blade wear is output. For wear evaluation values ​​that exceed the threshold, a sliding window weighted algorithm is used to predict the remaining life prediction value in combination with historical maintenance records. The present invention predicts the blade life through the correlation between operating parameters and remaining life through multidimensional data analysis and deep learning methods, thereby achieving accurate evaluation of the wear status of the kitchen garbage disposer blade and reliable prediction of the remaining life, which facilitates timely maintenance and replacement, and improves equipment operation efficiency and safety.

[0017] It is worth noting that in step S1, after the operating parameters are obtained, noise elimination and outlier processing are performed on the operating parameters to generate a standardized data set including motor operating status data and tool head vibration frequency data with timestamp alignment.

[0018] In this embodiment, the original data set of the real-time operating parameters of the kitchen garbage disposer is obtained through a sensor network. The original data set contains multiple time series sampling points, thereby obtaining the motor operating status data and the blade vibration frequency data aligned with the timestamp. Through the timestamp alignment, the correspondence between the two types of data at the same time point can be ensured, which provides the premise for the subsequent feature fusion. The sliding window method is used to smooth the original data set to generate a first intermediate data set after denoising. For the first intermediate data set, the dynamic threshold range of each sampling point is calculated, and the dynamic threshold range determines the upper threshold and the lower threshold based on the interquartile range. If the sampling point in the first intermediate data set exceeds the dynamic threshold range, the sampling point is eliminated to generate a second intermediate data set with blanks. For the blanks in the second intermediate data set, the linear interpolation method is used to fill them to obtain a complete standardized data set.

[0019] For example, the raw data set obtained from a kitchen garbage disposer through a sensor network often contains multiple time series sampling points, which reflect the operating status of the kitchen garbage disposer in different time periods. The raw data set may be interfered by noise, such as sudden vibrations during equipment operation or abnormal values ​​collected by sensors, so it needs to be smoothed. The sliding window method is a commonly used denoising method. Assuming that the window size is 5 sampling points, the data fluctuations at each point are smoothed by calculating the average value of the data in the window. For example, the blade vibration frequency data collected in a certain time period are 10Hz, 12Hz, 50Hz, 11Hz, and 9Hz. Obviously, 50Hz is an abnormal value. After smoothing with a sliding window, the value of this point may be adjusted to be close to the average value of the surrounding data, such as 10.4Hz, thereby generating the first intermediate data set.

[0020] Specifically, for the first intermediate data set, calculating the dynamic threshold range is a key step. The dynamic threshold range determines the upper and lower thresholds based on the interquartile range, and can adapt to changes in data distribution. For example, assuming that the interquartile range calculation result of a certain segment of vibration frequency data is 8Hz for the first quartile and 12Hz for the third quartile, then the upper threshold may be set to 15Hz and the lower threshold to 5Hz. If the value of a certain sampling point is 16Hz, which exceeds the upper threshold, it is regarded as an outlier and removed, generating a second intermediate data set with gaps. This method can dynamically adapt to changes in the operating status of the equipment and avoid misjudgments caused by fixed thresholds.

[0021] Linear interpolation is a simple and effective way to fill in gaps in the second intermediate dataset. For example, suppose that in a time series, the vibration frequency data for a certain point in time is missing. The data for the previous point in time is 10Hz, and the data for the next point in time is 12Hz. Using linear interpolation, we can estimate the missing data point to be 11Hz. This method maintains data continuity and generates a complete, standardized dataset.

[0022] Specifically, the processing methods for each of the above steps are closely centered around the specific characteristics of the kitchen garbage disposer's operating parameters. Sliding window smoothing effectively reduces noise interference with subsequent analysis, dynamic threshold range settings allow for flexible response to varying device operating conditions, and linear interpolation ensures data integrity. These processing methods work together to improve data reliability during the conversion from raw data to a standardized dataset, providing a solid foundation for subsequent device status analysis.

[0023] Preferably, in step S2, short-time Fourier transform is used to perform time-frequency analysis on the tool head vibration frequency data in the standardized data set, and the window function is set to a Hanning window with a window length of a preset sampling point to obtain a time-frequency distribution matrix; The time-frequency distribution matrix is ​​divided into multiple frequency bands for energy integration, and the energy proportion of each frequency band is calculated as a preliminary spectrum feature; From the preliminary spectrum characteristics, a core vibration characteristic frequency band is extracted using wavelet decomposition technology, and the average value and variance of the core vibration characteristic frequency band are calculated to form the vibration signal spectrum characteristics.

[0024] For time-frequency analysis of cutter head vibration frequency data, the short-time Fourier transform (SFT) is a tool that decomposes time-domain signals into two dimensions: time and frequency. Setting the Hanning window function with a window length of 256 sampling points achieves a balance between time and frequency resolution. The time-frequency distribution matrix can intuitively reflect the frequency variations of the vibration signal over different time periods. For example, when a kitchen garbage disposer is processing hard waste, the frequency may be concentrated in a higher frequency band during a certain period, while it may be concentrated in a lower frequency band when processing soft waste. This analysis method helps capture the dynamic changes in the operating status of the kitchen garbage disposer.

[0025] Based on the time-frequency distribution matrix, preliminary spectral features can be further extracted by dividing the frequency range into multiple bands and calculating the energy proportion of each band. Assuming the frequency range is divided into multiple bands such as 0 to 50 Hz and 50 to 100 Hz, the ratio of the sum of the energy of each band to the total energy is calculated. If the proportion of a certain frequency band is significantly higher than that of other frequency bands, it may reflect the vibration characteristics of the kitchen garbage disposer under specific operating conditions. This method can provide targeted data support for subsequent feature extraction.

[0026] Based on the preliminary spectral characteristics, wavelet decomposition technology is used to extract the core vibration characteristic frequency bands, allowing for a more detailed analysis of the signal's local characteristics. Assume that the wavelet decomposition breaks the signal into multiple layers, extracting the core vibration characteristic frequency bands associated with the key vibration modes of the kitchen garbage disposer. The mean and variance of these frequency bands are then calculated as statistical indicators. These indicators reflect the stability and changing trends of the vibration signal, providing important evidence for determining the device's status.

[0027] Optionally, in step S2, a fast Fourier transform is performed on the motor current signal in the motor operating status data in the standardized data set, multiple harmonic components are analyzed, and the amplitude ratio of each harmonic is calculated to obtain the motor current harmonic characteristics.

[0028] Fast Fourier transform (FFT) can decompose the current signal from motor operating status data into multiple harmonic components. Assuming the analysis results show a fundamental amplitude of 10 and a second harmonic amplitude of 2, the ratio of the second harmonic to the fundamental is 0.2. If this ratio increases significantly during certain time periods, it may indicate a change in motor load or a potential operating anomaly. Extracting this harmonic signature provides insight into the operating status of the kitchen garbage disposer blade from a current perspective.

[0029] Specifically, in step S2, based on the vibration signal spectrum characteristics and the motor current harmonic characteristics, a multidimensional feature set is generated by aligning the timestamps and adopting a horizontal splicing method. The multidimensional feature set includes the vibration signal spectrum characteristics and the motor current harmonic characteristics.

[0030] Based on the spectral characteristics of the vibration signal and the harmonic characteristics of the motor current, a multidimensional feature set is generated by horizontally splicing them together through timestamp alignment. This method integrates features from different sources into a unified feature matrix. Assuming that at a certain point in time, the statistical indicators of the vibration signal are one set of values, and the current harmonic ratios are another set of values, these are spliced ​​together to form a multidimensional vector. This feature set construction method can comprehensively characterize the operating status of the equipment, laying the foundation for subsequent analysis.

[0031] It is worth noting that, in step S3, the multidimensional feature data set is processed using a standardization method to obtain a standardized feature data set; The standardized feature data set is input into a pre-trained long short-term memory network model, and a time series analysis is performed on the standardized feature data set to obtain an evaluation value of the degree of tool head wear.

[0032] The spectral characteristics of the vibration signal can reflect the dynamic response of the cutter head during operation, while the motor current harmonic characteristics can reveal changes in the motor load. The purpose of normalizing multidimensional feature datasets is to eliminate the influence of different feature dimensions and ensure that the data can be compared on the same scale. In one possible implementation, a mean-standard deviation normalization method can be used to adjust the spectral characteristic values ​​of the vibration signal and the current harmonic characteristic values ​​to a distribution range with a mean of 0 and a standard deviation of 1. Assuming that the average value of each core vibration characteristic frequency band of the cutter head vibration frequency data ranges from 10 to 50, and the amplitude ratio of each harmonic of the motor current signal ranges from 0.1 to 0.5, after normalization, both are mapped to similar numerical ranges. This method can effectively prevent a feature with a large numerical range from having an excessive impact on subsequent models.

[0033] When inputting a standardized feature dataset into a pre-trained long-short-term memory (LSTM) network model for time series analysis, the focus is on capturing the data's dependencies along the time dimension. LSTM models excel at processing time series data, remembering long-term dependencies and ignoring irrelevant noise. Assuming that within a certain time period, the vibration signal spectrum characteristics of the tool head vibration frequency data exhibit periodic high-frequency fluctuations, while the current harmonic characteristics show a gradually increasing trend, the LSTM model can analyze these time series changes and infer that the tool head may be showing signs of wear due to long-term high-load operation. The model outputs an evaluation value between 0 and 1, such as 0.7, indicating a high degree of tool head wear.

[0034] In this embodiment, historical operating data and corresponding tool head wear evaluation values ​​are obtained. The historical operating data includes historical vibration signal spectral characteristics and historical motor current harmonic characteristics. A timestamp alignment method is used to align the evaluation values ​​and the historical operating data in the time dimension to obtain a time series basic data set. The time series basic data set is input into a long short-term memory network model for training, resulting in a pre-trained long short-term memory network model. The model uses the historical vibration signal spectral characteristics and historical motor current harmonic characteristics as inputs and the historical tool head wear evaluation values ​​as outputs.

[0035] Preferably, in step S4, the step of matching historical maintenance records according to the multidimensional feature set includes: extracting vibration signal spectrum features and motor current harmonic features corresponding to the evaluation value of the tool head wear degree exceeding a preset threshold from the multidimensional feature set to generate an abnormal feature data set; Historical records are retrieved from a pre-established database, where the historical records include historical data. A similarity matching method is used to determine historical data that matches the abnormal feature data set as a historical data subset, where the historical data subset includes historical vibration signal spectrum features and historical motor current harmonic features.

[0036] When retrieving historical feature vectors from the database and performing similarity matching, the Euclidean distance method can be used to calculate the proximity between the current anomaly feature dataset and the historical data. The closest historical data is then selected as the historical data subset. If a historical data subset exists in the historical data whose feature value deviates from the current anomaly feature by less than 5%, it is selected as the matching result. This matching method can quickly screen relevant historical data and provide a reference for subsequent analysis.

[0037] Optionally, in step S4, the step of calculating the remaining life prediction interval using a sliding window weighted algorithm includes: for the matched historical data subset, segmenting the features in the historical data using a sliding window method to obtain a segmented feature sequence; According to the segmented feature sequence, a weighted feature value corresponding to each segmented feature sequence is calculated using a weighted method according to the weights of the historical vibration signal spectrum characteristics and the historical motor current harmonic characteristics in the historical data subset; Using a time series analysis method to process the timestamp data corresponding to the weighted feature value and the corresponding segmented feature sequence to obtain a feature change trend parameter; The historical records also include historical maintenance records, and the weighted characteristic values ​​when the cutter head is replaced in the historical maintenance records are obtained. The time required to change from the current weighted characteristic value to the weighted characteristic value when the cutter head is replaced in the historical maintenance records is calculated according to the characteristic change trend parameters as the remaining life prediction value.

[0038] Sliding window segmentation of historical data subsets can be used to segment the data into 10-second windows, generating multiple segmented feature sequences. This method captures local variations in the data over time, laying the foundation for subsequent weight calculations. For example, if the vibration characteristics fluctuate significantly within a window, this may indicate that the tool head was subjected to high loads during that time period.

[0039] When calculating the weights of each feature and generating weighted eigenvalues, weights can be assigned based on the feature's contribution to wear. For example, if the weight of the historical vibration signal spectrum feature is 0.6 and the weight of the historical current harmonic feature is 0.4, the final weighted eigenvalue is 0.52. This weighting method can highlight the impact of key features and improve the targeted nature of the analysis.

[0040] When building a trend prediction model, since the segmented feature sequences corresponding to weighted eigenvalues ​​have unique timestamp data, we can combine these weighted eigenvalues ​​and timestamp data to analyze how features change over time. For example, suppose the weighted eigenvalue gradually increased from 0.2 to 0.4 over the past week. Time series analysis yields a trend parameter for the feature change, which is an average daily growth rate of 0.03. This trend parameter provides important support for subsequent forecasts.

[0041] Obtain historical maintenance records that match historical data corresponding to weighted eigenvalues ​​in the historical records, extract weighted eigenvalues ​​when the cutter head is replaced in the historical maintenance records, calculate the difference between the current weighted eigenvalue and the weighted eigenvalue when the cutter head is replaced, and obtain the time for the current weighted eigenvalue to change to the weighted eigenvalue when the cutter head is replaced based on the difference and the current feature change trend parameter based on the current feature change trend parameter, and use this time as the remaining life prediction value.

[0042] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. It is apparent to those skilled in the art that various changes, modifications, substitutions, and variations to these embodiments may be made without departing from the principles and spirit of the present invention, and these changes and modifications still fall within the scope of protection of the present invention.

Claims

1. A method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm, characterized in that: The following steps are involved: S1: Acquire real-time collected operating parameters of a kitchen garbage disposer and generate a standardized data set containing the operating parameters, wherein the operating parameters include motor operating state data and cutter head vibration frequency data; S2: Performing a time-frequency joint analysis on the standardized data set to extract the vibration signal spectrum features and motor current harmonic features, and constructing a multidimensional feature set; S3: Inputting the multidimensional feature set including the vibration signal spectrum features and the motor current harmonic features into a pre-trained long short-term memory network model, and outputting an evaluation value of the degree of tool head wear; S4: Extract the vibration signal spectrum characteristics and motor current harmonic characteristics corresponding to the evaluation value of the tool head wear degree exceeding the preset threshold, match the historical records based on the multidimensional feature set, and use the sliding window weighted algorithm to calculate the remaining life prediction value.

2. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 1, characterized in that: In step S1, after the operating parameters are acquired, noise elimination and outlier processing are performed on the operating parameters to generate a standardized data set including motor operating status data and tool head vibration frequency data with time stamp alignment.

3. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 1, characterized in that: In step S2, short-time Fourier transform is used to perform time-frequency analysis on the tool head vibration frequency data in the standardized data set, and the window function is set to Hanning window, and the window length is the preset sampling point to obtain a time-frequency distribution matrix; The time-frequency distribution matrix is ​​divided into multiple frequency bands for energy integration, and the energy proportion of each frequency band is calculated as a preliminary spectrum feature; From the preliminary spectrum characteristics, a core vibration characteristic frequency band is extracted using wavelet decomposition technology, and the average value and variance of the core vibration characteristic frequency band are calculated to form the vibration signal spectrum characteristics.

4. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 3, characterized in that: In step S2, a fast Fourier transform is performed on the motor current signal in the motor operating state data in the standardized data set, multiple harmonic components are analyzed, and the amplitude ratio of each harmonic is calculated to obtain the motor current harmonic characteristics.

5. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 4, characterized in that: In step S2, based on the vibration signal spectrum characteristics and the motor current harmonic characteristics, a multidimensional feature set is generated by aligning the timestamps and adopting a horizontal splicing method. The multidimensional feature set includes the vibration signal spectrum characteristics and the motor current harmonic characteristics.

6. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 1, characterized in that: In the step S3, the multidimensional feature data set is processed using a standardization method to obtain a standardized feature data set; The standardized feature data set is input into a pre-trained long short-term memory network model, and a time series analysis is performed on the standardized feature data set to obtain an evaluation value of the degree of tool head wear.

7. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 1, characterized in that: In step S4, the step of matching historical maintenance records according to the multidimensional feature set includes: extracting vibration signal spectrum features and motor current harmonic features corresponding to the evaluation value of the tool head wear degree exceeding a preset threshold from the multidimensional feature set to generate an abnormal feature data set; Historical records are retrieved from a pre-established database, where the historical records include historical data. A similarity matching method is used to determine historical data that matches the abnormal feature data set as a historical data subset, where the historical data subset includes historical vibration signal spectrum features and historical motor current harmonic features.

8. The method for predicting the life of a kitchen garbage disposer blade using a deep learning algorithm according to claim 7, characterized in that: In step S4, the step of calculating the remaining life prediction interval using a sliding window weighted algorithm includes: for a matched historical data subset, segmenting the features in the historical data using a sliding window method to obtain a segmented feature sequence; According to the segmented feature sequence, a weighted feature value corresponding to each segmented feature sequence is calculated using a weighted method according to the weights of the historical vibration signal spectrum characteristics and the historical motor current harmonic characteristics in the historical data subset; Using a time series analysis method to process the timestamp data corresponding to the weighted feature value and the corresponding segmented feature sequence to obtain a feature change trend parameter; The historical records also include historical maintenance records, and the weighted characteristic values ​​when the cutter head is replaced in the historical maintenance records are obtained. The time required to change from the current weighted characteristic value to the weighted characteristic value when the cutter head is replaced in the historical maintenance records is calculated according to the characteristic change trend parameters as the remaining life prediction value.

Images