Online detection system and method for damage of filter cloth of filter
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-12
- Publication Date
- 2026-04-03
AI Technical Summary
In the existing technology, the detection of filter cloth damage in filter presses relies on manual periodic inspections, which cannot be monitored in real time and continuously. This results in filter cloth damage not being detected in time, affecting the filtration effect and increasing production costs.
The system uses industrial microphones to automatically collect sound signals of filter cloth damage. Through audio data feature extraction and AI model recognition, it achieves real-time online detection of filter cloth damage, including data acquisition, transmission, algorithm analysis, and display control modules.
It enables real-time detection of filter cloth damage, reduces manual inspection, improves production efficiency, accurately locates faults, reduces costs, and ensures the safety of inspection personnel.
Smart Images
Figure FT_1 
Figure FT_2 
Figure FT_3
Abstract
Description
Technical Field
[0001] This invention belongs to the field of filter machine testing technology, specifically an online detection system and method for filter cloth damage in filter machines. Background Technology
[0002] In the field of industrial filtration production, filter presses are key equipment for solid-liquid separation. Filter cloth, as the core filtration component of the filter press, directly affects filtration efficiency and product quality. During normal operation, the filter cloth, with its fine pore structure, effectively intercepts solid particles, achieving solid-liquid separation. However, during long-term operation, the filter cloth can be damaged due to various factors such as material wear and chemical corrosion.
[0003] Currently, the detection of filter cloth damage in filter presses mainly relies on regular manual inspections. This method is not only labor-intensive but also has significant limitations. Because manual inspections cannot monitor the filter cloth's condition in real time and continuously, it is difficult to detect damage in its early stages. Often, the damage is only discovered when the filter cloth is severely damaged, resulting in a significant decrease in filtration efficiency and material leakage. This not only leads to unstable filter product quality but also causes material waste and increased production costs.
[0004] With the development of industrial intelligence, industrial microphones can quickly and accurately collect sound signals. By extracting audio data features and building AI models to identify filter cloth damage, they can accurately judge the damage status of the filter cloth, locate the damaged filter disc, and issue timely damage warnings. Therefore, the use of advanced technology to achieve real-time online detection of filter cloth damage has become an urgent need. Summary of the Invention
[0005] To address the problems existing in the prior art, this invention provides an online detection system and method for filter cloth damage in filter presses. By automatically collecting, processing, and identifying sound signals of filter cloth damage, the system automatically locates and displays the position of the damaged filter disc, thus achieving real-time online detection of filter cloth damage.
[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: An online detection system for filter cloth damage in a filter press includes: The data acquisition module is used to acquire audio stream data of the multi-channel filter disc operation; The data transmission module is used to transmit the audio stream data acquired by the data acquisition module, extract multi-channel digital signals from multiple audio streams through analog-to-digital conversion, and then send them to the algorithm analysis module; The algorithm analysis module is used to receive multi-channel digital signals transmitted by the data transmission module and perform filter cloth damage analysis. The display control module is used to display charts and control the audible and visual alarms based on the filter cloth damage analysis results.
[0007] Furthermore, the data acquisition module includes multiple industrial microphones and a sound acquisition card, with the multiple industrial microphones connected to the sound acquisition card via BNC coaxial cables. The algorithm analysis module is deployed on an industrial control computer, and the industrial control computer and the sound acquisition card are connected by a Cat5e network cable or optical fiber.
[0008] Furthermore, the algorithm analysis module includes: a multi-channel audio data preprocessing submodule, an audio data feature extraction submodule, a filter cloth damage identification submodule, a damaged filter disc positioning submodule, and a filter cloth damage early warning submodule; The multi-channel audio data preprocessing submodule is used to receive multi-channel digital signals from the data transmission module, trim the real-time audio data, retain valid information, and perform pre-emphasis processing on the audio data; The audio data feature extraction submodule is used to apply Fourier transform to the preprocessed audio data to obtain the audio data spectrum, perform Mel filter bank filtering on the spectrum, and normalize the extracted features. The filter cloth damage identification submodule is used to input the real-time audio data after feature extraction and normalization into the AI model to output the filter cloth status signal at each filter plate position. The damaged filter disc positioning submodule is used to locate the specific filter disc where the filter cloth is damaged based on the filter cloth status signal output by the AI model. The filter cloth damage early warning submodule calculates the number of filter cloth breaks in each cycle based on the filter cloth status signal, the position of the broken filter disc, and the rotation cycle calculated according to the frequency of the filter machine spindle. If the number of breaks exceeds the user-set threshold, an alarm message is issued.
[0009] Furthermore, the display control module includes: an audio display submodule, a database submodule, a data processing and display submodule, a log submodule, and a control submodule; The audio display submodule is used to receive the algorithm analysis results from the algorithm analysis module, draw a visualization image, and show the user the real-time status of all filter discs and filter cloths on the filter machine. The database submodule is used to store the data sent by the algorithm analysis module; The data processing and display submodule is used to draw charts from the data stored in the database module. The log submodule is used to allow users to query historical damage data and algorithm output information; The control submodule controls the audible and visual alarm to emit a warning sound based on the alarm information output by the filter cloth damage early warning submodule.
[0010] A method for online detection of filter cloth damage in a filter press includes: The audio stream data of the multi-channel filter disc operation is acquired through the data acquisition module; The data transmission module performs analog-to-digital conversion on the audio stream data acquired by the data acquisition module, extracts multi-channel digital signals from multiple audio streams, and then sends them to the algorithm analysis module. The algorithm analysis module receives audio data and performs filter cloth damage analysis. The display control module displays the analysis results from the algorithm analysis module and controls the audible and visual alarms.
[0011] Furthermore, the algorithm analysis module receives and analyzes audio data, including: The multi-channel audio data preprocessing submodule receives multi-channel digital signals from the data transmission module, trims the real-time audio data to retain valid information, and pre-emphasizes the audio data. The audio data spectrogram is obtained through the audio data feature extraction submodule, the spectrogram is filtered by Mel filter bank, and the extracted features are normalized. The filter cloth damage identification submodule inputs the real-time audio data after feature extraction and normalization into the AI model to output the filter cloth status signal at each filter plate position. The damaged filter disc location submodule locates the specific location of the damaged filter disc based on the filter cloth status signal output by the AI model. The filter cloth damage early warning submodule uses the filter cloth status signal, the position of the damaged filter disc, and the rotation cycle calculated based on the frequency of the filter machine spindle to obtain the number of filter cloth breaks in each cycle. If the number of breaks exceeds the user-set threshold, an alarm message is issued.
[0012] Furthermore, the display control module utilizes the filter cloth damage analysis results for result display and controls the audible and visual alarms, including: The audio display submodule receives the algorithm analysis results from the algorithm analysis module, draws a visual image, and displays the real-time status of all filter discs and filter cloths on the filter machine to the user. The database submodule stores the data sent by the algorithm analysis module. The data processing and display submodule draws the data stored in the database module into charts; The log submodule allows users to query historical damage data and algorithm output information; The alarm information output by the filter cloth damage early warning submodule is controlled by the control submodule to make the audible and visual alarm sound.
[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1) Real-time audio monitoring can detect filter cloth damage in advance, avoiding production accidents and quality losses; 2) Automated monitoring reduces manual inspections, lowers labor costs, and improves production efficiency; 3) The microphone placement design facilitates quick identification of damaged filter discs, accurate fault location, and shortens repair time; 4) Intuitive data display, providing an interface for operating status and damage conditions, facilitating equipment management and maintenance decisions; 5) Using automated inspection instead of manual inspection under harsh working conditions effectively protects the health and safety of inspection personnel. Attached Figure Description
[0014] Figure 1 This is an architecture diagram of the filter cloth damage online detection system provided in Embodiment 1 of this application; Figure 2 This is a schematic diagram illustrating the application scenario of the online detection system for filter cloth damage in a filter press provided in Embodiment 1 of this application; Figure 3 This is a data processing flowchart of the filter cloth damage online detection system provided in Embodiment 1 of this application; Figure 4 This is a flowchart illustrating the online detection system for filter cloth damage in a filter press provided in Embodiment 1 of this application. Figure 5 This is a flowchart of the online detection method for filter cloth damage in a filter press provided in Embodiment 2 of this application. Detailed Implementation
[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.
[0016] Example 1 See Figures 1-4 The filter cloth damage online detection system for filter presses includes: S1. Data acquisition module, which acquires audio stream data of the multi-channel filter disc operation.
[0017] A high-sensitivity industrial microphone is installed diagonally above the midpoint between every two adjacent filter discs in the filter press to acquire high-quality audio stream data. Each industrial microphone is equipped with a protective fixture. The industrial microphone model is MNP21. The microphone is wrapped in foam sponge and then inserted into the cylinder housing the industrial microphone fixture bracket. The foam sponge must be fully filled, completely covering the inner wall of the cylinder. A sponge cover is also added to the outside of the cylinder to prevent corrosion of internal equipment and dust from entering the fixture and clogging the industrial microphone. The connection between the fixture and the flexible conduit used to protect the cable is wrapped with tape. This protective fixture is made of corrosion-resistant material, protecting the microphone from corrosion and effectively preventing corrosion from surrounding dust, moisture, and humid gases, thus extending the microphone's lifespan.
[0018] To achieve remote data transmission, a single audio acquisition card manages and uses all installed industrial microphones. Coaxial cables provide a stable power supply to the audio acquisition card and microphones, and the card works in conjunction with an industrial control computer within the electrical control box. The industrial microphones and the audio acquisition card are connected via BNC coaxial cables to transmit audio stream data.
[0019] Choose between using network cable or fiber optic cable to transmit audio stream data based on the working distance. Use shielded Cat5e network cable when the industrial control computer is close to the sound acquisition card, and use fiber optic cable when the industrial control computer is far from the sound acquisition card. Install audible and visual alarms near each filter cloth inspection device to alert personnel to handle damaged filter cloths.
[0020] S2. Data transmission module: Deployed on the industrial control computer, it transmits audio stream data from multiple industrial microphones, extracts multi-channel digital signals from multiple audio streams through analog-to-digital conversion, and then sends them to the algorithm analysis module.
[0021] In industrial environments, multiple industrial microphones continuously capture analog sound signals generated by equipment operation and environmental noise, forming multiple independent audio streams. The data transmission module on the industrial control computer incorporates a high-precision analog-to-digital converter (ADC), model DSA-9234. It first precisely samples each audio stream at a pre-set fixed sampling frequency, converting the continuous analog sound signal into a discrete sequence of voltage samples at different time points. Then, it meticulously divides the voltage range into numerous quantization intervals, each corresponding to a specific digital code, thereby converting continuous voltage values into discrete digital quantities, ultimately forming a multi-channel digital signal reflecting sound information from different locations.
[0022] S3. Algorithm Analysis Module: Deployed on an industrial control computer, it receives audio data and performs analysis.
[0023] This module includes: a multi-channel audio data preprocessing submodule, an audio data feature extraction submodule, a filter cloth damage identification submodule, a damaged filter disc location submodule, and a filter cloth damage early warning submodule.
[0024] The multi-channel audio data preprocessing submodule receives multi-channel digital signals from the data transmission module, trims the real-time audio data to retain only valid information, and finally performs pre-emphasis processing on the audio data to enhance high-frequency signals and reduce the loss of high-frequency information.
[0025] In multi-channel audio data, effective information refers to the sound segments and related features that are truly valuable for subsequent tasks such as algorithm analysis, fault diagnosis, and sound feature extraction. Raw audio data in industrial environments often contains a large amount of redundant and interfering content, such as brief periods of silence before and after equipment startup, and constant low-frequency noise that persists in the environmental background noise and is not helpful to the analysis target. Effective information, on the other hand, consists of characteristic sounds emitted by specific components of equipment during normal operation, such as the specific frequency sound waves generated by the rotation of a motor; and abnormal sounds generated when equipment malfunctions, such as the friction sound of bearing wear or the airflow sound of a leaking pipe. These sounds contain key information about the equipment status, and by trimming and removing irrelevant parts, data processing efficiency and analysis accuracy can be improved.
[0026] Pre-emphasis processing aims to enhance the performance of the high-frequency components of the audio signal. Specifically, it references the values of adjacent sampling points and combines the current sampling point's value with the previous value. After processing, the intensity of the high-frequency signal is targeted to be enhanced, while the intensity of the low-frequency signal is relatively weakened. This effectively compensates for any attenuation that may occur in the high-frequency signal during the initial acquisition and transmission processes, resulting in a more balanced distribution of the audio signal across various frequency bands. This lays a solid foundation for subsequent accurate analysis of audio characteristics and accurate assessment of device status.
[0027] The audio data feature extraction submodule is used to apply Fourier transform to each audio data segment to obtain the spectrum of that segment.
[0028] The process involves performing analog-to-digital conversion and preprocessing on the analog audio signals collected by industrial microphones, then converting the time-domain data into a frequency-domain signal using a fast Fourier transform. After calculating the power spectral density, a two-dimensional heat map is plotted along the time-frequency-energy intensity dimension to generate an audio spectrum diagram for filter cloth damage detection.
[0029] The spectrogram is filtered using a Mel filter bank to obtain Mel spectral features. Finally, the extracted features are normalized to eliminate dimensional differences between different features.
[0030] The filter cloth damage recognition submodule is used to input the real-time audio data after feature extraction into the trained AI model. The AI model processes the multi-channel audio data synchronously in real time and outputs the status (normal and damaged) of the filter cloth at each filter position, and outputs the filter cloth status signal at each filter position.
[0031] For AI models, to maximize their generalization ability, it is necessary to repeatedly clean, filter, and segment the audio data collected from various environmental conditions showing the filter machines operating normally, constructing a filter cloth damage detection dataset containing sufficient feature information. The audio data of the filter machines operating normally, as described here, refers to the audio processed by the multi-channel audio data preprocessing submodule and the audio data feature extraction submodule.
[0032] The cleaning process involves first standardizing the format, ensuring all audio is encoded in a high-fidelity format and that the sampling rate and number of channels are consistent; then, noise reduction is performed, using algorithms such as spectral subtraction, Wiener filters, or wavelet transform to reduce background noise based on different noise sources.
[0033] The screening criteria must ensure the validity of the audio, retaining only segments from when the filter is operating normally and excluding invalid data from non-steady-state phases (abnormal data collected by industrial microphones); ensure completeness, so that the audio segments contain the complete operating cycle; take into account diversity, covering audio data under different environmental conditions; and set a minimum signal-to-noise ratio threshold to ensure that the audio quality meets the analysis requirements.
[0034] Segment labeling can be done manually by listening and labeling, with professionals labeling the equipment status, damage type, and timestamp one by one according to the audio content.
[0035] Sufficient feature information means that the training set should contain at least 1,000 independent samples to allow the model to learn the diverse features of filter cloth damage; in terms of class balance, the proportion of samples of each type of damage should be balanced to avoid the model being biased towards the majority class.
[0036] The dataset was then divided into training, validation, and test sets in a 6:2:2 ratio. The training set was then fed into a convolutional neural network for training for 100 epochs, and the highest-scoring AI model was evaluated using the validation set.
[0037] For each piece of audio data in the training, validation, and test sets, a unified preprocessing operation is required. This involves repeated cleaning, filtering, and segment labeling after processing by the multi-channel audio data preprocessing and audio data feature extraction submodules in the S3 algorithm analysis module. In daily work, the data input to the trained AI model only undergoes processing by the multi-channel audio data preprocessing and audio data feature extraction submodules in the S3 algorithm analysis module, without repeated cleaning, filtering, and segment labeling. During neural network training, each complete traversal of all data in the training set is called an epoch. Here, 100 epochs are set, meaning the network will learn from the training data for 100 rounds. After each training round, the validation set is used to evaluate the currently trained model. The validation set does not participate in model parameter updates; it is only used to check the model's performance at the current stage, such as accuracy and recall. By comparing the model's evaluation scores on the validation set at different epochs, the model with the highest score is selected as the final trained AI model. The model with the highest score is considered to have the best generalization ability under the current training strategy and can better handle unseen data. Finally, the selected best model will be used for final performance evaluation using a test set. The test set is also not used in model training; its purpose is to simulate real-world data and comprehensively and objectively evaluate the model's performance in practical applications, ensuring that the model can stably and accurately complete the filter cloth damage detection task in real-world use.
[0038] The damaged filter disc location submodule is used to locate the specific filter disc where the filter cloth has been damaged, based on the filter cloth status signal output by the AI model. Due to the continuous installation location of the industrial microphones, this allows for real-time monitoring of filter cloth damage characteristics, providing precise location information for timely repair or replacement.
[0039] The filter cloth damage early warning submodule monitors all filter cloths on the filter press in real time based on filter cloth status signals and the location of damaged filter discs. It calculates the rotation cycle based on the filter press spindle frequency and obtains the number of damaged filter cloths within each cycle. If the number of damaged filter cloths on a filter disc exceeds a user-defined threshold, an alarm is issued.
[0040] S4. Display and Control Module: Deployed on an industrial computer, it uses the analysis results for display and control.
[0041] This module includes: an audio display submodule, a database submodule, a data processing and display submodule, a log submodule, and a control submodule.
[0042] The audio display submodule is used to receive the algorithm analysis results from the algorithm analysis module, draw a visualization image, and show the user the real-time status of all filter discs and filter cloths on the filter machine.
[0043] The database submodule is used to store the data sent by the algorithm analysis module.
[0044] The data processing and display submodule is used to further process the data stored in the database module and draw it into a chart (curve of the number of broken filter cloths for each filter disc), which is updated once after each filter disc rotates once.
[0045] The log submodule is used by users to query historical damage data and algorithm output information.
[0046] The control submodule is used to install audible and visual alarms near each filter cloth inspection device. Based on the alarm information output by the filter cloth damage early warning submodule, the audible and visual alarms are controlled to prompt personnel to handle the damaged filter cloth.
[0047] Example 2 See Figure 5 Based on the above technical solutions, this application also provides an online detection method for filter cloth damage in a filter press. This method is applicable to online detection of filter cloth damage in a filter press. This method can be implemented by the online detection system for filter cloth damage in a filter press as described in the embodiments of this invention. This system can be implemented by software and / or hardware and is specifically configured in an electronic device.
[0048] The method includes: The audio stream data of the multi-channel filter disc operation is acquired through the data acquisition module; The data transmission module performs analog-to-digital conversion on the audio stream data acquired by the data acquisition module, extracts multi-channel digital signals from multiple audio streams, and then sends them to the algorithm analysis module. The algorithm analysis module receives and analyzes audio data. The display control module uses the analysis results from the algorithm analysis module to display the results and control the audible and visual alarms.
[0049] The algorithm analysis module receives and analyzes audio data, including: The multi-channel audio data preprocessing submodule receives multi-channel digital signals from the data transmission module, trims the real-time audio data to retain valid information, and performs pre-emphasis processing on the audio data. The audio data feature extraction submodule obtains the audio data spectrogram, performs Mel filter bank filtering on the spectrogram, and normalizes the extracted features. The filter cloth damage identification submodule inputs the real-time audio data after feature extraction and normalization into the AI model to output the filter cloth status signal at each filter plate position. The damaged filter disc positioning submodule locates the specific location of the damaged filter disc based on the filter cloth status signal output by the AI model. The filter cloth damage early warning submodule uses the filter cloth status signal, the position of the damaged filter disc, and the rotation cycle calculated based on the frequency of the filter machine spindle to obtain the number of filter cloth breaks in each cycle. If the number of breaks exceeds the user-set threshold, an alarm message is issued.
[0050] The display control module utilizes the filter cloth damage analysis results for result display and controls the audible and visual alarms, including: The audio display submodule receives the algorithm analysis results from the algorithm analysis module, draws a visualization image, and displays the real-time status of all filter discs and filter cloths on the filter machine to the user. The database submodule stores the data sent by the algorithm analysis module; The data processing and display submodule uses the data processing and display submodule to draw charts from the data stored in the database module. The log submodule allows users to query historical damage data and algorithm output information. The control submodule controls the audible and visual alarm to emit a warning sound based on the alarm information output by the filter cloth damage early warning submodule.
[0051] Finally, it should be noted that the above embodiments are merely preferred embodiments of the present invention and are only used to illustrate the technical solutions of the present invention, not to limit them. Any modifications, improvements, or other modifications or equivalent substitutions made by those skilled in the art to the technical solutions of the present invention, as long as they do not depart from the spirit and scope of the technical solutions of the present invention, should be included within the protection scope of the present invention.
Claims
1. An online detection system for filter cloth damage in a filter press, characterized in that, include: The data acquisition module is used to acquire audio stream data of the multi-channel filter disc operation; The data transmission module is used to transmit the audio stream data acquired by the data acquisition module, extract multi-channel digital signals from multiple audio streams through analog-to-digital conversion, and then send them to the algorithm analysis module; The algorithm analysis module is used to receive multi-channel digital signals transmitted by the data transmission module and perform filter cloth damage analysis. The display control module is used to display charts and control the audible and visual alarms based on the filter cloth damage analysis results.
2. The filter cloth damage online detection system for filter presses according to claim 1, characterized in that: The data acquisition module includes multiple industrial microphones and a sound acquisition card, with the multiple industrial microphones connected to the sound acquisition card via BNC coaxial cables. The algorithm analysis module is deployed on an industrial control computer, and the industrial control computer and the sound acquisition card are connected by a Cat5e network cable or optical fiber.
3. The online detection system for filter cloth damage of a filter press according to claim 2, characterized in that, The algorithm analysis module includes: a multi-channel audio data preprocessing submodule, an audio data feature extraction submodule, a filter cloth damage identification submodule, a damaged filter disc positioning submodule, and a filter cloth damage early warning submodule; The multi-channel audio data preprocessing submodule is used to receive multi-channel digital signals from the data transmission module, trim the real-time audio data, retain valid information, and perform pre-emphasis processing on the audio data; The audio data feature extraction submodule is used to apply Fourier transform to the preprocessed audio data to obtain the audio data spectrum, perform Mel filter bank filtering on the spectrum, and normalize the extracted features. The filter cloth damage identification submodule is used to input the real-time audio data after feature extraction and normalization into the AI model to output the filter cloth status signal at each filter plate position. The damaged filter disc positioning submodule is used to locate the specific filter disc where the filter cloth is damaged based on the filter cloth status signal output by the AI model. The filter cloth damage early warning submodule calculates the number of filter cloth breaks in each cycle based on the filter cloth status signal, the position of the broken filter disc, and the rotation cycle calculated according to the frequency of the filter machine spindle. If the number of breaks exceeds the user-set threshold, an alarm message is issued.
4. The filter cloth damage online detection system for filter presses according to claim 3, characterized in that: The display control module includes: an audio display submodule, a database submodule, a data processing and display submodule, a log submodule, and a control submodule; The audio display submodule is used to receive the algorithm analysis results from the algorithm analysis module, draw a visualization image, and show the user the real-time status of all filter discs and filter cloths on the filter machine. The database submodule is used to store the data sent by the algorithm analysis module; The data processing and display submodule is used to draw charts from the data stored in the database module. The log submodule is used to allow users to query historical damage data and algorithm output information; The control submodule controls the audible and visual alarm to emit a warning sound based on the alarm information output by the filter cloth damage early warning submodule.
5. A method for online detection of filter cloth damage in a filter press, characterized in that, include: The audio stream data of the multi-channel filter disc operation is acquired through the data acquisition module; The data transmission module performs analog-to-digital conversion on the audio stream data acquired by the data acquisition module, extracts multi-channel digital signals from multiple audio streams, and then sends them to the algorithm analysis module. The algorithm analysis module receives audio data and performs filter cloth damage analysis. The display control module displays the results of the filter cloth damage analysis and controls the audible and visual alarms.
6. The online detection method for filter cloth damage in a filter press according to claim 5, characterized in that, The algorithm analysis module receives and analyzes audio data, including: The multi-channel audio data preprocessing submodule receives multi-channel digital signals from the data transmission module, trims the real-time audio data to retain valid information, and pre-emphasizes the audio data. The audio data spectrogram is obtained through the audio data feature extraction submodule, the spectrogram is filtered by Mel filter bank, and the extracted features are normalized. The filter cloth damage identification submodule inputs the real-time audio data after feature extraction and normalization into the AI model to output the filter cloth status signal at each filter plate position. The damaged filter disc location submodule locates the specific location of the damaged filter disc based on the filter cloth status signal output by the AI model. The filter cloth damage early warning submodule uses the filter cloth status signal, the position of the damaged filter disc, and the rotation cycle calculated based on the frequency of the filter machine spindle to obtain the number of filter cloth breaks in each cycle. If the number of breaks exceeds the user-set threshold, an alarm message is issued.
7. The online detection method for filter cloth damage in a filter press according to claim 5, characterized in that, The display control module utilizes the filter cloth damage analysis results for result display and controls the audible and visual alarms, including: The audio display submodule receives the algorithm analysis results from the algorithm analysis module, draws a visual image, and displays the real-time status of all filter discs and filter cloths on the filter machine to the user. The database submodule stores the data sent by the algorithm analysis module. The data processing and display submodule draws the data stored in the database module into charts; The log submodule allows users to query historical damage data and algorithm output information; The alarm information output by the filter cloth damage early warning submodule is controlled by the control submodule to make the audible and visual alarm sound.