Wall breaking machine noise reduction method and wall breaking machine

Abstract

The invention discloses a wall breaking machine noise reduction method, and relates to the technical field of household appliances, and the method comprises the steps: collecting a noise signal generated in the working process of a wall breaking machine through a sound sensor, and preprocessing the noise signal to obtain standardized noise waveform data; converting the standardized noise waveform data from a time domain to a frequency domain by adopting a time-frequency analysis method, and extracting key frequency band characteristics reflecting noise characteristics; according to the key frequency band features, a machine learning algorithm is adopted to identify the noise type, and a noise feature template is matched to determine the noise type; calling a corresponding offset signal from a preset offset signal template library according to the noise category, and performing parameter optimization to generate an offset signal matched with the original noise; and playing the offset signal through a loudspeaker, superposing the offset signal with original noise to realize noise offset, and carrying out adaptive optimization on the offset signal to adapt to noise change.

Description

Technical Field

[0001] The present invention relates to the technical field of household appliances, and in particular to a noise reduction method for a wall-breaking machine and a wall-breaking machine. Background Art

[0002] During the use of the wall-breaking machine, the noise problem has always been an important factor affecting the user experience. Traditional noise reduction methods mainly rely on passive noise reduction technology, such as adding sound insulation materials, etc., but these methods have limited effects in practical applications and cannot effectively solve the noise problem of the wall-breaking machine under different working conditions. In recent years, active noise reduction technology has gradually attracted attention. It achieves noise reduction by generating a cancellation signal with a phase opposite to the noise, but the existing technology has technical bottlenecks in real-time dynamic adjustment of the cancellation signal.

[0003] When the wall breaker is working, its noise frequency and intensity will change with time, working status, materials and other factors. Simply using a fixed cancellation signal is difficult to effectively adapt to this dynamic change, resulting in poor noise cancellation effect. For example, the Chinese patent with publication number (CN118942437A) discloses an open active noise reduction wall breaker. Although it generates a noise reduction signal with the opposite phase to the noise through methods such as fast Fourier transform, in actual application, it is still impossible to dynamically adjust the cancellation signal in real time to adapt to the changing noise characteristics. Summary of the invention

[0004] In order to solve the above-mentioned technical problems, the present invention provides a noise reduction method for a wall breaking machine and a wall breaking machine.

[0005] The technical solution of the present invention is achieved in this way: A noise reduction method for a wall breaking machine comprises the following steps: S1. Collecting the noise signal generated during the working process of the wall breaking machine through a sound sensor, and preprocessing the noise signal to obtain standardized noise waveform data; S2. Converting the standardized noise waveform data from the time domain to the frequency domain using a time-frequency analysis method to extract key frequency band features reflecting noise characteristics; S3, using a machine learning algorithm to identify the noise type according to the key frequency band characteristics, and matching the noise feature template to determine the noise category; S4, retrieving a corresponding cancellation signal from a preset cancellation signal template library according to the noise category, and performing parameter optimization to generate a cancellation signal that matches the original noise; S5, playing the cancellation signal through a loudspeaker to superimpose it with the original noise to achieve noise cancellation, and adaptively optimizing the cancellation signal to adapt to noise changes; S6. Calculate the difference between the residual noise after cancellation and the original noise to obtain a noise suppression amount, and perform iterative optimization of the cancellation signal according to a comparison result between the noise suppression amount and the target threshold.

[0006] Furthermore, the process of extracting the key frequency band features reflecting the noise characteristics in step S2 is specifically as follows: Short-time Fourier transform is used to convert the standardized noise waveform data from the time domain to the frequency domain to obtain the energy distribution characteristics of the noise signal in different frequency bands; According to the energy distribution characteristics of the noise signal in different frequency bands, the key frequency band characteristics that can reflect the noise characteristics are extracted, including the energy mean and variance statistical characteristics of the low-frequency, medium-frequency and high-frequency bands; Build a noise recognition model and use the extracted key frequency band features as the input of the model.

[0007] Furthermore, the process of matching the noise feature template to determine the noise category in step S3 is specifically as follows: In the noise recognition process, if the key frequency band characteristics of the current noise have a high matching degree with the preset noise type feature template, the noise is judged to belong to the corresponding type and the category to which the noise belongs is obtained; Specifically, according to the material properties and working state of the wall-breaking machine, a corresponding typical frequency domain feature template is selected from a pre-established noise model library; The dynamic time warping algorithm is used to match and compare the frequency domain feature data of the noise with the selected typical frequency domain feature template, and the similarity between the two is calculated; if the similarity is greater than the preset threshold, the current noise is judged to belong to the noise category corresponding to the typical frequency domain feature template; if the similarity is less than the threshold, it continues to match with other typical frequency domain feature templates; The above process is repeated until a typical frequency domain feature template with the largest similarity and greater than the threshold is found to determine the category to which the noise belongs.

[0008] Furthermore, the step S3 further includes: After determining the category to which the noise belongs, analyze the source and cause of the noise according to the category of the noise and the acoustic characteristic parameters of the noise of this category in the noise model library; Specifically, the association rules between the working state parameters of the wall-breaking machine and the characteristic vectors of the noise signal are mined through the data mining algorithm, and a mapping model between the two is established; According to the established mapping model, the changes in noise signals are monitored in real time to determine whether the working state of the wall-breaking machine is abnormal. If an abnormality is detected, an early warning prompt is triggered and the noise signal data under the abnormal state is recorded for subsequent fault diagnosis and model optimization.

[0009] Furthermore, the process of generating the cancellation signal matching the original noise in step S4 is specifically as follows: Adjust the parameters of the obtained cancellation signal template, match the frequency, amplitude, and phase parameters of the template with the actual characteristic parameters of the noise signal, and generate a cancellation signal that matches the original noise signal; The generated cancellation signal is superimposed on the original noise signal, and the strength of the original noise signal is weakened or eliminated through the waveform cancellation principle to obtain a superimposed and cancelled signal.

[0010] Furthermore, the process of adaptively optimizing the cancellation signal in step S5 is specifically as follows: After obtaining the signal after superposition and cancellation, the effect of the signal after superposition and cancellation is evaluated, the characteristic parameters of the residual noise are extracted, the noise reduction amount is calculated, and it is judged whether the noise cancellation effect reaches the preset threshold; if the noise cancellation effect does not reach the preset threshold, the parameters of the cancellation signal are further optimized until the noise cancellation effect reaches the preset threshold, and the final noise-reduced signal is output.

[0011] Furthermore, the specific process of iterative optimization of the cancellation signal in step S6 is as follows: Calculate the difference in energy between the residual noise signal and the original noise signal in each frequency band to obtain the noise suppression amount in each frequency band; The noise suppression amount in each frequency band is weighted averaged to obtain the total noise suppression amount, which is used as an indicator to measure the noise suppression effect; The total noise suppression amount is compared with a preset noise suppression target threshold value, and if the noise suppression amount is lower than the target threshold value, it is determined that the noise suppression effect is not ideal; If the noise suppression effect is not ideal, an adaptive filtering algorithm is used to optimize the cancellation signal, and the residual noise is minimized by adjusting the amplitude and phase of the cancellation signal; According to the optimized cancellation signal, the above process is repeated until the total noise suppression amount is higher than or equal to the target threshold. If the total noise suppression amount is higher than or equal to the target threshold, it is judged that the noise suppression effect reaches the expectation and the current cancellation signal is maintained.

[0012] Furthermore, the method further comprises: Real-time monitoring of the working status of the wall-breaking machine and changes in material properties triggers the noise feature extraction and analysis process, dynamically adjusts the noise recognition model and cancellation signal template, and realizes adaptive optimization of the cancellation signal.

[0013] A wall-breaking machine, comprising a main body, on which is mounted a sound sensor for real-time collection of noise data generated during the working process of the wall-breaking machine, and on which is also disposed a noise reduction device for reducing the noise of the wall-breaking machine, wherein the noise reduction device comprises a speaker, a signal processing module, a cancellation signal generation and adjustment module, and an adaptive filtering module; The signal processing module includes an analog-to-digital converter, a digital signal processor and a memory; The cancellation signal generation and adjustment module includes an electrically adjustable attenuator, a phase adjuster and a power synthesizer.

[0014] Furthermore, the analog-to-digital converter is used to convert the analog noise signal collected by the sound sensor into a digital signal for subsequent digital signal processing; The digital signal processor is used to perform a time-frequency analysis method to convert the noise signal from the time domain to the frequency domain, and obtain the energy distribution characteristics of the noise in different frequency bands; it is also used to identify and classify the frequency domain characteristics of the noise according to a pre-established noise model library; and the matching technology is used to retrieve the cancellation signal template and adjust the amplitude and phase of the template; The electrically adjustable attenuator and phase adjuster are used to adjust the amplitude and phase of the cancellation signal template according to the actual frequency and intensity parameters of the noise, so as to generate a cancellation signal matching the original noise; The power synthesizer performs vector power synthesis on the adjusted cancellation signal and the reference signal to form a final cancellation signal; The adaptive filtering module is used to implement an improved variable step size minimum mean square adaptive filtering algorithm based on radial basis function neural network, optimize the cancellation signal in real time, and continuously correct the parameters of the cancellation signal to adapt to the dynamic changes of noise.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention collects the working noise of the wall-breaking machine in real time, performs time-frequency analysis and feature recognition on it, retrieves the matching template from the preset cancellation signal library, and uses the improved adaptive filtering algorithm. The present invention can dynamically optimize the cancellation signal to make it accurately match the original noise, and plays the optimized cancellation signal through the speaker to achieve active suppression of the wall-breaking machine noise; 2. The present invention also introduces a residual noise feedback mechanism, which continuously adjusts the offset strategy by comparing the noise suppression amount with the target threshold. At the same time, when the working state or material of the wall breaker changes, the present invention can automatically repeat the suppression process to ensure continuous and effective noise suppression under different working conditions. This intelligent and adaptive noise suppression method significantly improves the comfort of the wall breaker and the quality of the working environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1This is a flowchart of a noise reduction method for a wall breaking machine according to Example 1; Figure 2 is a schematic structural diagram of the noise reduction device in Example 2; Figure 3 is a framework diagram of the signal processing module in Example 2; Figure 4 This is a framework diagram of the cancellation signal generation and adjustment module in Example 2. DETAILED DESCRIPTION

[0017] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example 1 like Figure 1 As shown, this embodiment provides a noise reduction method for a wall breaking machine, comprising the following steps: S1. Collecting the noise signal generated during the working process of the wall breaking machine through a sound sensor, and preprocessing the noise signal to obtain standardized noise waveform data; S2. Converting the standardized noise waveform data from the time domain to the frequency domain using a time-frequency analysis method to extract key frequency band features reflecting noise characteristics; S3, using a machine learning algorithm to identify the noise type according to the key frequency band characteristics, and matching the noise feature template to determine the noise category; S4, retrieving a corresponding cancellation signal from a preset cancellation signal template library according to the noise category, and performing parameter optimization to generate a cancellation signal that matches the original noise; S5, playing the cancellation signal through a loudspeaker to superimpose it with the original noise to achieve noise cancellation, and adaptively optimizing the cancellation signal to adapt to noise changes; S6. Calculate the difference between the residual noise after cancellation and the original noise to obtain a noise suppression amount, and perform iterative optimization of the cancellation signal according to a comparison result between the noise suppression amount and the target threshold.

[0019] Furthermore, the process of extracting the key frequency band features reflecting the noise characteristics in step S2 is specifically as follows: Short-time Fourier transform is used to convert the standardized noise waveform data from the time domain to the frequency domain to obtain the energy distribution characteristics of the noise signal in different frequency bands; According to the energy distribution characteristics of the noise signal in different frequency bands, the key frequency band characteristics that can reflect the noise characteristics are extracted, including the energy mean and variance statistical characteristics of the low-frequency, medium-frequency and high-frequency bands; Build a noise recognition model and use the extracted key frequency band features as the input of the model; Specifically, as described below, 8 sound sensors are arranged around the wall breaker with a sampling frequency of 44.1kHz to collect the noise data generated during the operation of the wall breaker in real time and obtain the original waveform of the noise signal. The original noise waveform data is feature extracted using the wavelet analysis method, and 12 feature parameters such as the mean, variance, and energy of the wavelet coefficients are extracted to form the feature vector of the noise signal. The extracted feature vector is input into the noise classification model based on the convolutional neural network; For the preprocessing of the original noise data, the median filtering method can be used to remove outliers, and then the normalization method can be used to map the noise amplitude to the [0, 1] interval to obtain the standardized noise waveform data. Then, the standardized noise waveform is subjected to short-time Fourier transform using a Hamming window with a window length of 10ms and a frame shift of 5ms to obtain the energy distribution characteristics of the noise signal in the 0-8kHz frequency band. According to the energy distribution characteristics, the energy mean and variance of the three frequency bands of low frequency (0-2kHz), medium frequency (2-5kHz), and high frequency (5-8kHz) are extracted as key frequency band features; Next, a noise recognition model based on support vector machine is constructed. The key frequency band features are taken as input, and the discriminant function of the noise type is obtained through training. During the recognition process, if the Euclidean distance between the key frequency band features of the current noise and the preset five noise type feature templates is less than the threshold value of 0.1, the noise is judged to belong to the corresponding type.

[0020] According to the identified noise category, a convolutional neural network algorithm is used to perform template matching in a pre-established library of 100 cancellation signal templates to select the cancellation signal template that best matches the current noise category; For the selected cancellation signal template, the frequency, amplitude, phase and other parameters of the noise signal are fitted by the least square method to generate a cancellation signal that matches the original noise signal; Finally, the generated cancellation signal is superimposed on the original noise signal to weaken the noise intensity through the waveform cancellation principle. The effect of the superimposed cancellation signal is evaluated, the frequency domain energy characteristics of the residual noise are extracted, and the noise reduction amount is calculated. If the noise reduction amount is less than 20dB, the cancellation signal generation step is returned to further optimize the cancellation signal parameters through the gradient descent method until the noise reduction amount reaches 20dB, and the final noise reduction signal is output.

[0021] Furthermore, the process of matching the noise feature template to determine the noise category in step S3 is specifically as follows: In the noise recognition process, if the key frequency band characteristics of the current noise have a high matching degree with the preset noise type feature template, the noise is judged to belong to the corresponding type and the category to which the noise belongs is obtained; Specifically, according to the material properties and working state of the wall-breaking machine, a corresponding typical frequency domain feature template is selected from a pre-established noise model library; The dynamic time warping algorithm is used to match and compare the frequency domain feature data of the noise with the selected typical frequency domain feature template, and the similarity between the two is calculated; if the similarity is greater than the preset threshold, the current noise is judged to belong to the noise category corresponding to the typical frequency domain feature template; if the similarity is less than the threshold, it continues to match with other typical frequency domain feature templates; The above process is repeated until a typical frequency domain feature template with the largest similarity and greater than the threshold is found to determine the category to which the noise belongs.

[0022] Furthermore, the step S3 further includes: After determining the category to which the noise belongs, analyze the source and cause of the noise according to the category of the noise and the acoustic characteristic parameters of the noise of this category in the noise model library; Specifically, the association rules between the working state parameters of the wall-breaking machine and the characteristic vectors of the noise signal are mined through the data mining algorithm, and a mapping model between the two is established; According to the established mapping model, by real-time monitoring of the changes in noise signals, it is determined whether the working state of the wall-breaking machine is abnormal. If an abnormality is detected, an early warning prompt is triggered, and the noise signal data under the abnormal state is recorded for subsequent fault diagnosis and model optimization; Specifically, as described below, according to the material properties (stainless steel, glass, etc.) and working state (no load, load, etc.) of the wall breaking machine, the corresponding typical frequency domain feature template is selected from the pre-established noise model library containing 500 samples, and the dynamic time warping algorithm is used. With 10ms as the time window, the Euclidean distance between the noise frequency domain feature data and the typical frequency domain feature template is calculated to obtain the similarity between the two. If the similarity is greater than 0.85, it is determined that the current noise belongs to the noise category corresponding to the typical frequency domain feature template; if the similarity is less than 0.85, it is continued to match with other typical frequency domain feature templates until a template with the largest similarity and greater than 0.85 is found to determine the category to which the noise belongs; According to the noise category, combined with the acoustic characteristic parameters of the noise in the noise model library, such as the sound pressure level and frequency range, a logistic regression algorithm is used to analyze the source (mechanical failure, environmental noise, etc.) and cause (bearing wear, rotor imbalance, etc.) of the noise; Through the Apriori association rule mining algorithm, with a minimum support of 0.05 and a minimum confidence of 0.8, the association rules between the working state parameters of the wall-breaking machine, such as speed and vibration, and the frequency domain feature vector of the noise signal are mined, and a nonlinear mapping model between the two is established; Based on this model, the changes in the frequency domain characteristics of the noise signal are monitored in real time. When the deviation between the characteristic vector and the model prediction value exceeds 10%, the working state of the wall-breaking machine is judged to be abnormal, an early warning prompt is triggered, and the noise signal data under the abnormal state is recorded for subsequent fault diagnosis and model parameter optimization.

[0023] Furthermore, the process of generating the cancellation signal matching the original noise in step S4 is specifically as follows: Adjust the parameters of the obtained cancellation signal template, match the frequency, amplitude, and phase parameters of the template with the actual characteristic parameters of the noise signal, and generate a cancellation signal that matches the original noise signal; The generated cancellation signal is superimposed on the original noise signal, and the strength of the original noise signal is weakened or eliminated through the waveform cancellation principle to obtain a superimposed and cancelled signal; Specifically, as described below, according to the identified noise category, a convolutional neural network is used to perform template matching in the cancellation signal template library to find the cancellation signal template that is most similar to the current noise category, and then the frequency, amplitude, phase and other parameters of the matched cancellation signal template are optimized to match the characteristic parameters of the original noise signal, thereby generating a targeted cancellation signal; The generated cancellation signal is superimposed on the original noise signal. The waveforms of the two signals are canceled out by the principle of acoustic interference, thereby weakening the noise intensity. The spectrum of the canceled signal is analyzed to calculate the noise reduction amount. If the reduction amount does not reach the preset -20dB threshold, the cancellation signal optimization step is returned to further adjust the cancellation signal parameters through the gradient descent algorithm until the noise reduction amount meets the requirements. Finally, the optimized cancellation signal is played in real time through the speaker on the wall breaker and superimposed with the original noise to achieve active noise reduction. At the same time, an adaptive filtering algorithm based on radial basis function neural network is adopted to dynamically adjust the amplitude and phase parameters of the cancellation signal according to the correlation between the residual noise and the cancellation signal to adapt to the real-time changes of noise and maintain a stable noise reduction effect.

[0024] Furthermore, the process of adaptively optimizing the cancellation signal in step S5 is specifically as follows: After obtaining the signal after superposition and cancellation, the effect of the signal after superposition and cancellation is evaluated, the characteristic parameters of the residual noise are extracted, the noise reduction amount is calculated, and it is judged whether the noise cancellation effect reaches the preset threshold; if the noise cancellation effect does not reach the preset threshold, the parameters of the cancellation signal are further optimized until the noise cancellation effect reaches the preset threshold, and the final noise-reduced signal is output.

[0025] Furthermore, the specific process of iterative optimization of the cancellation signal in step S6 is as follows: Calculate the difference in energy between the residual noise signal and the original noise signal in each frequency band to obtain the noise suppression amount in each frequency band; The noise suppression amount in each frequency band is weighted averaged to obtain the total noise suppression amount, which is used as an indicator to measure the noise suppression effect; The total noise suppression amount is compared with a preset noise suppression target threshold value, and if the noise suppression amount is lower than the target threshold value, it is determined that the noise suppression effect is not ideal; If the noise suppression effect is not ideal, an adaptive filtering algorithm is used to optimize the cancellation signal, and the residual noise is minimized by adjusting the amplitude and phase of the cancellation signal; According to the optimized cancellation signal, the above process is repeated until the total noise suppression amount is higher than or equal to the target threshold. If the total noise suppression amount is higher than or equal to the target threshold, it is judged that the noise suppression effect reaches the expected value, and the current cancellation signal is maintained; Specifically, as described below, by performing fast Fourier transform (FFT) analysis on the noise signal collected during the operation of the wall-breaking machine, characteristic parameters such as the main noise frequency of 1.5kHz, amplitude of -20dB, and phase of 45° are extracted, and the extracted characteristic parameters are input into a pre-trained support vector machine (SVM) model for noise category recognition. The recognition result shows that the noise belongs to the motor noise category with a confidence level of 95%; According to the identified motor noise category, the convolutional neural network (CNN) algorithm is used to search for the most matching cancellation signal template in the cancellation signal template library, with a matching degree of 98%. The parameters of the matched cancellation signal template are fine-tuned to adjust the template frequency to 1.5kHz, the amplitude to 20dB, and the phase to 225° to generate a cancellation signal that matches the original noise signal. The improved variable step-size least mean square (LMS) adaptive filtering algorithm based on radial basis function neural network (RBFNN) was used to optimize the cancellation signal in real time with an initial step size of 0.01. After 100 iterations, the mean square error converged to less than 0.001, realizing the dynamic correction of the cancellation signal parameters. The optimized cancellation signal is played through the speaker on the wall-breaking machine and superimposed with the original noise signal. Through the principle of waveform cancellation, the intensity of the noise signal is weakened by 20dB. The effect of the signal after superposition and cancellation was evaluated. The frequency of the residual noise was 1.8kHz, the amplitude was -40dB, and the phase was 60°. The noise reduction was calculated to be 20dB, reaching the preset 18dB noise cancellation threshold. The final noise-reduced signal was output, completing the active cancellation process of the wall-breaking machine noise. Furthermore, the method further comprises: Real-time monitoring of the working status of the wall-breaking machine and changes in material properties, triggering the noise feature extraction and analysis process, dynamically adjusting the noise recognition model and the cancellation signal template, and achieving adaptive optimization of the cancellation signal; Specifically, as described below, the original noise waveform data generated by the wall-breaking machine during operation is obtained through a sensor, and the noise signal is converted into a digital signal for processing; According to the pre-established noise signal characteristic model of the wall-breaking machine, feature extraction is performed on the acquired original noise waveform data to obtain the feature vector of the noise signal; The extracted noise signal feature vector is input into the adaptive filter, and the adaptive filter dynamically adjusts the weight coefficient of the filter according to the characteristics of the noise signal to generate a cancellation signal with high correlation with the noise signal; Perform correlation analysis on the generated cancellation signal and the original noise signal, and calculate the correlation coefficient between the two. If the correlation is higher than a preset threshold, the cancellation signal is output to the speaker of the wall-breaking machine for playback; The sensor obtains the residual noise after the speaker plays the cancellation signal, and feeds the residual noise signal back to the adaptive filter. The adaptive filter optimizes the weight coefficient of the filter in real time according to the residual noise signal, and continuously corrects the parameters of the cancellation signal to adapt to the dynamic changes of the noise. Repeat the above process until the residual noise is less than the preset threshold or reaches the set number of iterations, completing the active cancellation process of the wall-breaking machine noise; perform correlation analysis on the working state parameters of the wall-breaking machine and the characteristic vector of the noise signal, and use the data mining algorithm to mine the correlation rules between the two to determine the relationship between the noise and the working state of the wall-breaking machine; According to the results of the correlation analysis, a mapping model between the working state of the wall breaker and the noise signal is established. The working state of the wall breaker is monitored in real time through the changes in the noise signal. When the working state of the wall breaker is detected to be abnormal, an early warning prompt is triggered in time, and the noise signal data under the abnormal state is recorded for subsequent fault diagnosis and model optimization. When the working state of the wall breaker changes or the material used changes, the above step 6 is repeated to achieve dynamic adjustment of the offset signal to ensure that the noise can be effectively suppressed under different working conditions.

[0026] Exemplarily, the original noise waveform data of the wall-breaking machine during operation is obtained by a sensor at a sampling rate of 44.1kHz, and 12 MFCC feature vectors are extracted by short-time Fourier transform of the noise signal. The feature vectors are input into an adaptive LMS filter, and the 32 weight coefficients of the filter are dynamically adjusted by the minimum mean square error criterion to generate a cancellation signal; Calculate the Pearson correlation coefficient between the cancellation signal and the original noise. If it is higher than 0.8, output the cancellation signal to the speaker for playback. Obtain the residual noise and feed it back to the filter. Optimize the weights in real time with a learning rate of 0.01, and iterate 50 times until the residual noise is less than 30dB. The Apriori algorithm is used to mine the association rules between noise characteristics and working parameters such as wall-breaking machine speed and material viscosity, with a support degree of 0.5 and a confidence level of 0.8; A BP neural network mapping model is established based on association rules to monitor the working condition of the wall-breaking machine in real time according to the changes in noise characteristics. If the noise energy exceeds the normal value by 20%, an early warning is triggered and the abnormal data is recorded.

[0027] In this embodiment, multiple sound sensors are arranged around the wall breaking machine to collect noise signals in real time, and pre-process to obtain standardized noise waveform data. Then, short-time Fourier transform and other time-frequency analysis methods are used to convert the noise signals from the time domain to the frequency domain, extract key frequency band features, and construct a noise recognition model. Use machine learning algorithms to identify noise types and match noise feature templates to determine noise categories. Based on the identification results, retrieve the corresponding cancellation signal from the preset cancellation signal template library, and perform parameter optimization to generate a cancellation signal that matches the original noise. Noise reduction is achieved by superimposing the cancellation signal played by the speaker with the original noise, and an adaptive filtering algorithm is used to optimize the cancellation signal in real time to adapt to the dynamic changes of the noise. At the same time, the difference between the residual noise after cancellation and the original noise is calculated to obtain the noise suppression amount, and it is iteratively optimized based on the comparison result with the target threshold; In addition, the present invention also monitors the working status and material property changes of the wall breaking machine in real time, dynamically adjusts the noise recognition model and the offset signal template, and ensures that noise can be effectively suppressed under different working conditions, which significantly improves the noise reduction effect of the wall breaking machine and the user experience.

[0028] Example 2 like Figure 2-4As shown, this embodiment provides a wall breaking machine for implementing the above-mentioned wall breaking machine noise reduction method, comprising a main body, on which is mounted a sound sensor for real-time collection of noise data generated during the working process of the wall breaking machine, and on which is also disposed a noise reduction device for reducing the noise of the wall breaking machine, wherein the noise reduction device comprises a speaker, a signal processing module, a cancellation signal generation and adjustment module, and an adaptive filtering module; The signal processing module includes an analog-to-digital converter, a digital signal processor and a memory; The cancellation signal generation and adjustment module includes an electrically adjustable attenuator, a phase adjuster and a power synthesizer.

[0029] Furthermore, the analog-to-digital converter is used to convert the analog noise signal collected by the sound sensor into a digital signal for subsequent digital signal processing; The digital signal processor is used to perform a time-frequency analysis method to convert the noise signal from the time domain to the frequency domain, and obtain the energy distribution characteristics of the noise in different frequency bands; it is also used to identify and classify the frequency domain characteristics of the noise according to a pre-established noise model library; and the matching technology is used to retrieve the cancellation signal template and adjust the amplitude and phase of the template; The electrically adjustable attenuator and phase adjuster are used to adjust the amplitude and phase of the cancellation signal template according to the actual frequency and intensity parameters of the noise, so as to generate a cancellation signal matching the original noise; The power synthesizer performs vector power synthesis on the adjusted cancellation signal and the reference signal to form a final cancellation signal; The adaptive filtering module is used to implement an improved variable step size minimum mean square adaptive filtering algorithm based on radial basis function neural network, optimize the cancellation signal in real time, and continuously correct the parameters of the cancellation signal to adapt to the dynamic changes of noise.

[0030] This embodiment provides a wall-breaking machine with a noise reduction function. A sound sensor is installed on the main body to collect noise data in real time during the working process. The noise reduction device includes a speaker, a signal processing module, a cancellation signal generation and adjustment module, and an adaptive filtering module; The analog-to-digital converter in the signal processing module converts the analog noise signal into a digital signal. The digital signal processor performs time-frequency analysis, identifies the noise characteristics and retrieves the cancellation signal template. The electrically adjustable attenuator and phase adjuster adjust the cancellation signal according to the noise parameters. The power synthesizer synthesizes the final cancellation signal. The adaptive filtering module uses an improved algorithm based on the radial basis function neural network to optimize the cancellation signal in real time and adapt to noise changes, thereby effectively reducing the noise of the wall breaker during operation and improving the user experience.

[0031] The specific embodiments of the invention are described in detail above, but they are only examples. The present invention is not limited to the specific embodiments described above. Those skilled in the art should understand that the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A noise reduction method for a wall breaking machine, characterized in that: The following steps are involved: S1. Collecting the noise signal generated during the working process of the wall breaking machine through a sound sensor, and preprocessing the noise signal to obtain standardized noise waveform data; S2. Converting the standardized noise waveform data from the time domain to the frequency domain using a time-frequency analysis method to extract key frequency band features reflecting noise characteristics; S3, using a machine learning algorithm to identify the noise type according to the key frequency band characteristics, and matching the noise feature template to determine the noise category; S4, retrieving a corresponding cancellation signal from a preset cancellation signal template library according to the noise category, and performing parameter optimization to generate a cancellation signal that matches the original noise; S5, playing the cancellation signal through a loudspeaker to superimpose it with the original noise to achieve noise cancellation, and adaptively optimizing the cancellation signal to adapt to noise changes; S6. Calculate the difference between the residual noise after cancellation and the original noise to obtain a noise suppression amount, and perform iterative optimization of the cancellation signal according to a comparison result between the noise suppression amount and the target threshold.

2. A wall breaking machine noise reduction method according to claim 1, characterized in that: The process of extracting the key frequency band features reflecting the noise characteristics in step S2 is specifically as follows: Short-time Fourier transform is used to convert the standardized noise waveform data from the time domain to the frequency domain to obtain the energy distribution characteristics of the noise signal in different frequency bands; According to the energy distribution characteristics of the noise signal in different frequency bands, the key frequency band characteristics that can reflect the noise characteristics are extracted, including the energy mean and variance statistical characteristics of the low-frequency, medium-frequency and high-frequency bands; Build a noise recognition model and use the extracted key frequency band features as the input of the model.

3. A wall breaking machine noise reduction method according to claim 1, characterized in that: The process of matching the noise feature template to determine the noise category in step S3 is specifically as follows: In the noise recognition process, if the key frequency band characteristics of the current noise have a high matching degree with the preset noise type feature template, the noise is judged to belong to the corresponding type and the category to which the noise belongs is obtained; Specifically, according to the material properties and working state of the wall-breaking machine, a corresponding typical frequency domain feature template is selected from a pre-established noise model library; The dynamic time warping algorithm is used to match and compare the frequency domain feature data of the noise with the selected typical frequency domain feature template, and the similarity between the two is calculated; if the similarity is greater than the preset threshold, the current noise is judged to belong to the noise category corresponding to the typical frequency domain feature template; if the similarity is less than the threshold, it continues to match with other typical frequency domain feature templates; The above process is repeated until a typical frequency domain feature template with the largest similarity and greater than the threshold is found to determine the category to which the noise belongs.

4. A wall breaking machine noise reduction method according to claim 3, characterized in that: The step S3 further comprises: After determining the category to which the noise belongs, analyze the source and cause of the noise according to the category of the noise and the acoustic characteristic parameters of the noise of this category in the noise model library; Specifically, the association rules between the working state parameters of the wall-breaking machine and the characteristic vectors of the noise signal are mined through the data mining algorithm, and a mapping model between the two is established; According to the established mapping model, the changes in noise signals are monitored in real time to determine whether the working state of the wall-breaking machine is abnormal. If an abnormality is detected, an early warning prompt is triggered and the noise signal data under the abnormal state is recorded for subsequent fault diagnosis and model optimization.

5. A wall breaking machine noise reduction method according to claim 1, characterized in that: The process of generating the cancellation signal matching the original noise in step S4 is specifically as follows: Adjust the parameters of the obtained cancellation signal template, match the frequency, amplitude, and phase parameters of the template with the actual characteristic parameters of the noise signal, and generate a cancellation signal that matches the original noise signal; The generated cancellation signal is superimposed on the original noise signal, and the strength of the original noise signal is weakened or eliminated through the waveform cancellation principle to obtain a superimposed and cancelled signal.

6. A wall breaking machine noise reduction method according to claim 1, characterized in that: The process of adaptively optimizing the cancellation signal in step S5 is specifically as follows: After obtaining the signal after superposition and cancellation, the effect of the signal after superposition and cancellation is evaluated, the characteristic parameters of the residual noise are extracted, the noise reduction amount is calculated, and it is determined whether the noise cancellation effect reaches a preset threshold; If the noise cancellation effect does not reach the preset threshold, the parameters of the cancellation signal are further optimized until the noise cancellation effect reaches the preset threshold, and the final noise-reduced signal is output.

7. A wall breaking machine noise reduction method according to claim 1, characterized in that: The specific process of iterative optimization of the cancellation signal in step S6 is: Calculate the difference in energy between the residual noise signal and the original noise signal in each frequency band to obtain the noise suppression amount in each frequency band; The noise suppression amount in each frequency band is weighted averaged to obtain the total noise suppression amount, which is used as an indicator to measure the noise suppression effect; The total noise suppression amount is compared with a preset noise suppression target threshold value, and if the noise suppression amount is lower than the target threshold value, it is determined that the noise suppression effect is not ideal; If the noise suppression effect is not ideal, an adaptive filtering algorithm is used to optimize the cancellation signal, and the residual noise is minimized by adjusting the amplitude and phase of the cancellation signal; According to the optimized cancellation signal, the above process is repeated until the total noise suppression amount is higher than or equal to the target threshold. If the total noise suppression amount is higher than or equal to the target threshold, it is judged that the noise suppression effect reaches the expectation and the current cancellation signal is maintained.

8. A wall breaking machine noise reduction method according to claim 1, characterized in that: Also includes: Real-time monitoring of the working status of the wall-breaking machine and changes in material properties triggers the noise feature extraction and analysis process, dynamically adjusts the noise recognition model and cancellation signal template, and realizes adaptive optimization of the cancellation signal.

9. A wall-breaking machine, used to implement the wall-breaking machine noise reduction method according to any one of claims 1 to 8, characterized in that: The invention comprises a main body, on which a sound sensor for real-time acquisition of noise data generated during the working process of the wall-breaking machine is installed, and a noise reduction device for reducing the noise of the wall-breaking machine is also arranged on the main body, and the noise reduction device comprises a speaker, a signal processing module, a cancellation signal generation and adjustment module and an adaptive filtering module; The signal processing module includes an analog-to-digital converter, a digital signal processor and a memory; The cancellation signal generation and adjustment module includes an electrically adjustable attenuator, a phase adjuster and a power synthesizer.

10. A wall breaking machine according to claim 9, characterized in that: The analog-to-digital converter is used to convert the analog noise signal collected by the sound sensor into a digital signal for subsequent digital signal processing; The digital signal processor is used to perform a time-frequency analysis method to convert the noise signal from the time domain to the frequency domain, and obtain the energy distribution characteristics of the noise in different frequency bands; it is also used to identify and classify the frequency domain characteristics of the noise according to a pre-established noise model library; and the matching technology is used to retrieve the cancellation signal template and adjust the amplitude and phase of the template; The electrically adjustable attenuator and phase adjuster are used to adjust the amplitude and phase of the cancellation signal template according to the actual frequency and intensity parameters of the noise, so as to generate a cancellation signal matching the original noise; The power synthesizer performs vector power synthesis on the adjusted cancellation signal and the reference signal to form a final cancellation signal; The adaptive filtering module is used to implement an improved variable step size minimum mean square adaptive filtering algorithm based on radial basis function neural network, optimize the cancellation signal in real time, and continuously correct the parameters of the cancellation signal to adapt to the dynamic changes of noise.

Images