Equipment fault high-precision detection system in strong noise environment

Abstract

The invention belongs to the technical field of equipment detection, and particularly relates to an equipment fault high-precision detection system in a strong noise environment, which comprises a multi-channel audio acquisition module, a self-adaptive noise suppression module, a fault feature extraction module, an intelligent fault diagnosis module and a remote monitoring alarm module, according to the method, various sound signals generated in the operation process of the equipment are captured, and real-time noise suppression, feature extraction and fault recognition and classification are performed, so that corresponding processing measures can be taken timely and reasonably, and equipment damage or long-time production interruption caused by further deterioration of faults is avoided; and the optimization decision output module analyzes the system repair optimization emergency degree, optimizes and repairs the system when the optimization repair emergency signal is generated, verifies the recognition performance of the system for the equipment fault in different noise scenes after the optimization repair of the system is completed, ensures that the system can accurately and quickly recognize the equipment fault, and improves the system reliability. And the operation safety of the equipment is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment detection, and in particular to a high-precision detection system for equipment failures in a strong noise environment. Background Art

[0002] In industrial production environments, mechanical equipment such as generators, compressors, and machine tools often need to operate in a high-noise environment. This noise comes from the equipment's own mechanical vibration, fluid flow, and the external environment. In a high-noise environment, the noise not only affects the normal operation of the equipment but also masks early signs of equipment failure, making it difficult to detect and address the failure in a timely manner.

[0003] Traditional equipment fault detection methods, such as vibration monitoring, perform well in low-noise environments but are susceptible to interference in high-noise environments. This affects the accuracy and timeliness of fault detection, and makes it difficult to reasonably determine the urgency of system repair and optimization and the need for strengthened equipment supervision, thus failing to significantly improve equipment operational safety and stability.

[0004] In view of the above technical defects, a solution is now proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a high-precision detection system for equipment failures in a strong noise environment, which solves the problem that the existing technology is easily interfered with in a strong noise environment, resulting in the accuracy and timeliness of equipment failure detection being affected, and it is difficult to reasonably judge the urgency of system repair and optimization and the necessity of strengthening equipment supervision, which is not conducive to ensuring the safe, stable and continuous operation of the equipment.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A high-precision detection system for equipment faults in a strong noise environment, including a multi-channel audio acquisition module, an adaptive noise suppression module, a fault feature extraction module, an intelligent fault diagnosis module, and a remote monitoring and alarm module;

[0008] The multi-channel audio acquisition module uses an array of multiple high-sensitivity, wide-bandwidth microphones placed at preset locations around the device. The microphones capture various sound signals generated by the device during operation and perform preliminary amplification and filtering through the signal conditioning circuit;

[0009] The adaptive noise suppression module uses an adaptive filtering algorithm based on deep learning to automatically adjust filtering parameters according to changes in the noise environment, performing real-time noise suppression on the collected sound signals to remove background noise. The fault feature extraction module uses audio signal processing technology to extract features from the noise-suppressed sound signals and sends the feature extraction results to the intelligent fault diagnosis module.

[0010] The intelligent fault diagnosis module uses a machine learning algorithm to learn and train historical equipment fault data, establish a mapping relationship between fault characteristics and fault types, and build a fault identification and classification model. The fault identification and classification model determines whether there is a fault in the equipment and its specific type based on the feature extraction results, and sends the fault diagnosis information to the remote monitoring and alarm module; if there is an equipment fault, the remote monitoring and alarm module will issue an early warning, and maintenance personnel will inspect and repair the equipment.

[0011] Furthermore, the remote monitoring alarm module is communicated with the optimization decision output module, which collects the date of the last optimization of the system and marks it as the adjacent table date, marks the interval between the current date and the adjacent table date as the adjacent interval time, compares the adjacent interval time with a preset adjacent interval time threshold, and generates an optimization repair emergency signal if the adjacent interval time exceeds the preset adjacent interval time threshold;

[0012] If the interval time does not exceed the preset interval time threshold, the monitoring period is L1, which is set by tracing back the current date and the end date. The optimization decision coefficient is obtained through comprehensive analysis of the urgency of optimization and repair. The optimization decision coefficient is numerically compared with the preset optimization decision coefficient threshold. If the optimization decision coefficient exceeds the preset optimization decision coefficient threshold, an optimization and repair emergency signal is generated and sent to the remote monitoring alarm module.

[0013] Furthermore, the specific analysis process of optimizing the comprehensive analysis of repair urgency is as follows:

[0014] The moment when the multi-channel audio acquisition module collects sound signals is obtained and marked as the first moment, and the moment when the intelligent fault diagnosis module outputs fault diagnosis information is marked as the second moment, and the interval between the first moment and the second moment is marked as the analysis duration; all analysis durations within the monitoring period are obtained and the average is calculated to obtain the analysis timeliness value, and the proportion of analysis durations that exceed the preset analysis duration threshold is marked as the analysis inefficiency value;

[0015] When the intelligent fault diagnosis module determines that the equipment has a fault, the maintenance personnel obtain the inspection and maintenance information after completing the inspection and maintenance. If the equipment does not have the corresponding fault, the diagnostic abnormality symbol KP-1 is assigned; the number of times the diagnostic abnormality symbol KP-1 is assigned during the monitoring period is obtained and the ratio of the number of times the diagnostic abnormality symbol KP-1 is assigned to the adjacent interval time is marked as the diagnostic abnormality detection value;

[0016] The number of times that maintenance personnel find faults in the equipment during routine inspection and maintenance of the equipment during the monitoring period but the intelligent fault diagnosis module fails to detect the corresponding fault is marked as a fault identification value, and the fault identification value is calculated by ratioing the fault identification value with the inspection and maintenance frequency during the monitoring period to obtain the identification detection value; the optimized decision coefficient is obtained by numerically calculating the analysis timeliness value, the analysis inefficiency value, the diagnosis detection value and the identification detection value.

[0017] Furthermore, when the optimization decision output module generates an optimization repair emergency signal, the management personnel optimize and repair the system performance as needed. After completing the system optimization and repair, the system's recognition accuracy, missed alarm rate, false alarm rate and analysis efficiency for equipment failures under different noise scenarios are verified. If the system's recognition accuracy, missed alarm rate, false alarm rate and analysis efficiency for equipment failures under different noise scenarios meet the corresponding requirements, the optimization and repair are judged to be qualified; otherwise, the optimization and repair are judged to be unqualified and the remote monitoring alarm module issues an early warning.

[0018] Furthermore, the optimization decision output module communicates with the verification environment construction module, and the optimization decision output module sends the optimization repair emergency signal to the verification environment construction module. Before verifying the system performance, the verification environment construction module builds the required verification environment, and performs a construction accuracy analysis on the constructed verification environment to obtain a construction abnormality coefficient, and compares the construction abnormality coefficient with the corresponding preset construction abnormality coefficient threshold. If the construction abnormality coefficient exceeds the preset construction abnormality coefficient threshold, a construction alarm signal is generated, and the construction alarm signal is sent to the remote monitoring alarm module.

[0019] Furthermore, the specific analysis process of building accuracy analysis is as follows:

[0020] The noise decibel value, ambient temperature, ambient humidity, ambient air pressure and ambient cleanliness of the constructed environment are collected. The noise decibel value is calculated by subtracting the median value of the corresponding preset noise decibel value range and taking the absolute value to obtain the noise abnormality value. Similarly, the temperature abnormality value, humidity abnormality value, air pressure abnormality value and cleanliness abnormality value are obtained.

[0021] Assign corresponding preset weight coefficients to the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value and cleanliness abnormal value, multiply the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value and cleanliness abnormal value with the corresponding preset weight coefficients respectively, and sum the five groups of product results to obtain the abnormality coefficient.

[0022] Furthermore, the optimization decision output module is communicatively connected to the strong management necessity decision module. If the optimization repair emergency signal is not generated, the strong management necessity decision module will analyze the operating performance of the equipment during the monitoring period, and evaluate the degree of necessity of strengthening equipment management based on this, and determine whether to generate a strong management high necessity signal, and send the strong management high necessity signal to the remote monitoring alarm module. When the remote monitoring alarm module receives the strong management high necessity signal, it reminds the management personnel to strengthen the subsequent supervision of the equipment.

[0023] Furthermore, the specific analysis process of the mandatory management decision module is as follows:

[0024] All faults that occurred during the monitoring period are obtained and classified. The number of occurrences of the corresponding type of fault during the monitoring period is collected and marked as a frequency coefficient. The frequency coefficient is numerically compared with the preset frequency coefficient threshold for the corresponding type of fault. If a fault type with a frequency coefficient exceeding the corresponding preset frequency coefficient threshold occurs during the monitoring period, a strong control high necessary signal is generated.

[0025] If there is no fault type whose frequency coefficient exceeds the corresponding preset frequency coefficient threshold during the monitoring period, then the corresponding preset impact coefficient is assigned to it based on the degree of harm of the corresponding type of fault to the equipment, and the frequency coefficient of the corresponding type of fault is calculated by ratio with the corresponding preset frequency coefficient threshold, and the product of the ratio result and the corresponding preset impact coefficient is marked as the fault harm value;

[0026] The fault hazard value of all types of faults occurring during the monitoring period is summed up to obtain the fault hazard value, which is then compared with the preset fault hazard threshold. If the fault hazard value exceeds the preset fault hazard threshold, a necessary signal for strong control is generated.

[0027] Furthermore, if the fault danger meter value does not exceed the preset fault danger meter threshold, the total duration of equipment suspension due to the fault during the monitoring period is collected and marked as the operation impact value, the total amount of loss caused by the equipment due to the fault during the monitoring period is collected and marked as the fault loss value, and the ratio of the fault danger meter value to the corresponding preset fault danger meter threshold is marked as the fault detection value;

[0028] The necessary coefficient for strong pipe operation is obtained by numerically calculating the fault detection value, the operation impact value, and the fault loss value. The necessary coefficient for strong pipe operation is numerically compared with the preset necessary coefficient threshold for strong pipe operation. If the necessary coefficient for strong pipe operation exceeds the preset necessary coefficient threshold for strong pipe operation, a high necessary signal for strong pipe operation is generated.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. In the present invention, by capturing various sound signals generated by the equipment during operation and performing real-time noise suppression, feature extraction, and fault identification and classification, high-precision detection of equipment faults in strong noise environments and timely warnings are achieved, thereby preventing further deterioration of the faults from causing equipment damage or long-term production interruptions, reducing the risk of equipment operation, and analyzing the urgency of system repair optimization through the optimization decision output module. When the optimization and repair emergency signal is generated, the system is optimized and repaired. After the system optimization and repair is completed, the system's recognition performance for equipment faults in different noise scenarios is verified, ensuring that the system can accurately and quickly identify equipment faults, further improving the safety of equipment operation;

[0031] 2. In the present invention, the required verification environment is built before the system performance verification through the verification environment building module, and the built verification environment is analyzed for construction accuracy to determine whether the simulated environmental conditions meet the verification requirements under the corresponding scenario, thereby avoiding adverse effects on the system performance verification under the corresponding scenario and ensuring the accuracy of the verification results. In addition, the operation performance of the equipment during the monitoring period is analyzed by the strong control necessity decision module before the optimization repair emergency signal is generated to determine whether it is necessary to strengthen the subsequent supervision of the equipment, which is conducive to the reasonable adjustment of the subsequent supervision planning plan to ensure the subsequent safe and stable operation of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0033] Figure 1 This is a system block diagram of Embodiment 1 of the present invention;

[0034] Figure 2 This is a system block diagram of Embodiment 2 and Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0035] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments 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 making creative efforts are within the scope of protection of the present invention.

[0036] Example 1: Figure 1 As shown, the present invention proposes a high-precision detection system for equipment faults in a strong noise environment, which includes a multi-channel audio acquisition module, an adaptive noise suppression module, a fault feature extraction module, an intelligent fault diagnosis module, an optimized decision output module and a remote monitoring alarm module;

[0037] The multi-channel audio acquisition module uses an array of multiple high-sensitivity, wide-bandwidth microphones arranged at preset positions around the device. The microphones capture various sound signals generated by the device during operation (the microphones have a wide frequency response range and low distortion characteristics, and can capture weak signals generated by equipment failures) and perform preliminary amplification and filtering through signal conditioning circuits. Multi-channel audio acquisition helps to obtain the spatial distribution information of sound signals, providing data support for subsequent noise suppression and fault feature extraction.

[0038] The adaptive noise suppression module uses an adaptive filtering algorithm based on deep learning to perform real-time noise suppression on the collected sound signals. The algorithm can automatically adjust the filtering parameters according to changes in the noise environment, effectively remove background noise, and improve the signal-to-noise ratio of the sound signal. Through adaptive noise suppression, the clarity of the fault characteristic signal can be significantly improved, providing a reliable data basis for subsequent fault detection.

[0039] The fault feature extraction module uses audio signal processing technologies such as short-time Fourier transform (STFT) and Mel-frequency cepstral coefficients (MFCC) to extract features from the noise-suppressed sound signal and sends the feature extraction results to the intelligent fault diagnosis module. These extracted features can accurately reflect the operating status and fault characteristics of the equipment, such as changes in the intensity of specific frequency components and periodic anomalies in sound patterns.

[0040] The intelligent fault diagnosis module is based on machine learning algorithms, such as support vector machines (SVM) and convolutional neural networks (CNN). By learning and training historical equipment fault data, it establishes a mapping relationship between fault features and fault types, thereby building a fault identification and classification model. The fault identification and classification model determines whether the equipment has a fault and its specific type based on the feature extraction results, and sends the fault diagnosis information to the remote monitoring and alarm module.

[0041] If there is an equipment failure, an early warning will be issued through the remote monitoring alarm module, and maintenance personnel will inspect and repair the equipment, and can take appropriate measures in a timely and reasonable manner to prevent the failure from further deteriorating and causing equipment damage or long-term production interruption.

[0042] The optimization decision output module collects the date of the last system optimization and marks it as the adjacent table date, marks the interval between the current date and the adjacent table date as the adjacent interval, and compares the adjacent interval with the preset adjacent interval threshold. If the adjacent interval exceeds the preset adjacent interval threshold, it indicates that the system has not been optimized for a long time, and an optimization repair emergency signal is generated;

[0043] If the interval time does not exceed the preset interval time threshold, a monitoring period L1 is set based on the current date and the end date and traced back for a number of days. Preferably, L1 is fifteen days. An optimized decision coefficient is obtained by optimizing the comprehensive analysis of the repair urgency, specifically: the moment when the multi-channel audio acquisition module collects the sound signal is obtained and marked as the first moment, and the moment when the intelligent fault diagnosis module outputs the fault diagnosis information is marked as the second moment, and the interval between the corresponding first moment and the second moment is marked as the analysis duration.

[0044] Among them, the larger the value of the analysis time, the lower the efficiency of the corresponding detection and analysis process of the equipment abnormality; all analysis times in the monitoring period are obtained and the average is calculated to obtain the analysis time value, and the analysis time is compared with the preset analysis time threshold, and the proportion of analysis times that exceed the preset analysis time threshold is marked as the analysis inefficiency value;

[0045] When the intelligent fault diagnosis module determines that the equipment has a fault, the maintenance personnel obtain the inspection and maintenance information after completing the inspection and maintenance. If the equipment does not have the corresponding fault, it indicates that the corresponding diagnosis result of the intelligent fault diagnosis module is inaccurate, and the diagnostic abnormality symbol KP-1 is assigned; the number of times the diagnostic abnormality symbol KP-1 is assigned during the monitoring period is obtained and the ratio of the number of times the diagnostic abnormality symbol KP-1 is assigned to the adjacent interval time is marked as the diagnostic abnormality detection value;

[0046] The number of times during the monitoring period that maintenance personnel discovered equipment faults during routine equipment inspection and maintenance, but the intelligent fault diagnosis module failed to detect the corresponding faults, is marked as a fault identification value. The fault identification value is then calculated by comparing the fault identification value with the inspection and maintenance frequency during the monitoring period to obtain an identification detection value. The larger the values of the diagnosis detection value and the identification detection value, the worse the fault identification performance of the intelligent fault diagnosis module during the monitoring period.

[0047] The optimization decision coefficient Ghy is obtained by numerically calculating the analysis timeliness value W, the analysis inefficiency value F, the abnormality detection value T, and the abnormality identification detection value S using the formula Ghy = (q1×W+q2×F) / (q1+q2)+q3×T+q4×S. Here, q1, q2, q3, and q4 are preset proportional coefficients with values greater than zero. The larger the value of the optimization decision coefficient Ghy, the worse the overall detection and identification performance of the system for equipment faults, and the more timely the system needs to be optimized and repaired.

[0048] The optimization decision coefficient Ghy is numerically compared with the preset optimization decision coefficient threshold. If the optimization decision coefficient Ghy exceeds the preset optimization decision coefficient threshold, it indicates that the system's detection and identification performance for equipment failure is generally poor, and the system needs to be optimized and repaired in time. An optimization and repair emergency signal is generated and sent to the remote monitoring and alarm module. When the remote monitoring and alarm module receives the optimization and repair emergency signal, it issues an early warning to remind management personnel to optimize the system performance in a timely manner.

[0049] Furthermore, when the optimization decision output module generates an optimization and repair emergency signal, management personnel optimize and repair system performance as needed. After completing the system optimization and repair, the system verifies the system's recognition accuracy (the percentage of correctly identified faults), missed alarm rate (the percentage of actual faults that were not identified), false alarm rate (the percentage of incorrectly identified faults), and analysis efficiency (the percentage of analysis speed) for equipment faults under different noise scenarios.

[0050] If the system's recognition accuracy, missed alarm rate, false alarm rate and analysis efficiency for equipment faults in different noise scenarios meet the corresponding requirements, the optimized repair is judged to be qualified. Otherwise, the optimized repair is judged to be unqualified and the remote monitoring alarm module issues an early warning to remind management personnel to continue to take corresponding repair measures and perform secondary optimization of system performance to ensure that the system can accurately and quickly identify equipment faults after optimization, which is conducive to improving the safety of equipment operation.

[0051] Example 2: Figure 2 As shown, the difference between this embodiment and the first embodiment is that the optimization decision output module is in communication connection with the verification environment construction module, and the optimization decision output module sends the optimization repair emergency signal to the verification environment construction module. Before verifying the system performance, the verification environment construction module builds the required verification environment and performs a construction accuracy analysis on the verification environment built;

[0052] The noise decibel value, ambient temperature, ambient humidity, ambient air pressure, and ambient cleanliness of the constructed environment are collected (the higher the value of the ambient cleanliness, the lower the dust concentration in the environment). The noise decibel value is calculated by subtracting the median value of the corresponding preset noise decibel value range and taking the absolute value to obtain the noise difference value. Similarly, the temperature difference value, humidity difference value, air pressure difference value, and cleanliness difference value are obtained. Among them, the larger the value of the noise difference value, temperature difference value, humidity difference value, air pressure difference value, and cleanliness difference value, the less the simulated environmental conditions meet the verification requirements of the corresponding scenario;

[0053] Assign corresponding preset weight coefficients to the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value, and cleanliness abnormal value. It should be noted that the values of the preset weight coefficients are all positive numbers; multiply the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value, and cleanliness abnormal value by the corresponding preset weight coefficients respectively, and sum the five sets of product results to obtain the abnormality coefficient;

[0054] The construction abnormality coefficient is numerically compared with the corresponding preset construction abnormality coefficient threshold. If the construction abnormality coefficient exceeds the preset construction abnormality coefficient threshold, it indicates that the simulated environmental conditions do not meet the verification requirements under the corresponding scenario. A construction alarm signal is generated and sent to the remote monitoring alarm module for timely environmental adjustment to avoid adverse effects on the system performance verification under the corresponding scenario and ensure the accuracy of the verification results.

[0055] Example 3: Figure 2 As shown, the difference between this embodiment and the first and second embodiments is that the optimization decision output module is communicatively connected to the enhanced management necessity decision module. If the optimization repair emergency signal is not generated, the enhanced management necessity decision module analyzes the operating performance of the equipment during the monitoring period, evaluates the necessity of strengthening equipment management based on the performance, and determines whether to generate an enhanced management necessity signal.

[0056] The signal indicating the need for enhanced control is sent to the remote monitoring and alarm module. When the remote monitoring and alarm module receives the signal, it reminds the management personnel to strengthen the subsequent supervision of the equipment and reasonably adjust the subsequent supervision plan, including maintenance supervision and operation supervision, to effectively ensure the subsequent safe and stable operation of the equipment. The specific analysis process of the enhanced control necessity decision module is as follows:

[0057] All equipment faults that occurred during the monitoring period are obtained and classified. The number of occurrences of the corresponding type of fault during the monitoring period is collected and marked as the frequency coefficient. The frequency coefficient is numerically compared with the preset frequency coefficient threshold for the corresponding type of fault. If a fault type with a frequency coefficient exceeding the corresponding preset frequency coefficient threshold exists during the monitoring period, it indicates that the equipment has a high level of safety hazard and requires subsequent strengthening of equipment supervision. In this case, a necessary signal for strong control is generated.

[0058] If there is no fault type whose frequency coefficient exceeds the corresponding preset frequency coefficient threshold during the monitoring period, then the corresponding preset impact coefficient is assigned to it based on the degree of harm of the corresponding type of fault to the equipment. The values of all preset impact coefficients are greater than zero, and the higher the degree of harm of the corresponding type of fault to the equipment, the larger the value of the preset impact coefficient corresponding to it;

[0059] The frequency coefficient of the corresponding type of fault is calculated by ratio with the corresponding preset frequency coefficient threshold, and the product of the ratio result and the corresponding preset influence coefficient is marked as the fault hazard value; the fault hazard values of all types of faults occurring during the monitoring period are summed up to obtain the fault hazard value, and the fault hazard value is numerically compared with the preset fault hazard threshold. If the fault hazard value exceeds the preset fault hazard threshold, it indicates that the equipment safety hazard level is high and equipment supervision needs to be strengthened in the future, then a necessary signal for strong control is generated.

[0060] Furthermore, if the fault danger meter value does not exceed the preset fault danger meter threshold, the total duration of equipment suspension due to the fault during the monitoring period is collected and marked as the operation impact value, the total amount of loss caused by the equipment failure during the monitoring period is collected and marked as the fault loss value, and the ratio of the fault danger meter value to the corresponding preset fault danger meter threshold is marked as the fault detection value;

[0061] The necessary coefficient for strengthening the pipe, Qey, is calculated by numerically calculating the fault detection value B, the operational impact value M, and the fault loss value Y using the formula Qey = (uk1 × B + uk2 × M + uk3 × Y) / 3. Here, uk1, uk2, and uk3 are preset proportional coefficients with values greater than zero. A larger value for the necessary coefficient for strengthening the pipe, Qey, indicates that the overall performance of the equipment during the monitoring period is worse, and that subsequent equipment supervision needs to be strengthened.

[0062] The necessary coefficient for strong control, Qey, is numerically compared with the preset threshold value for the necessary coefficient for strong control. If the necessary coefficient for strong control, Qey, exceeds the threshold value, it indicates that the overall performance of the equipment during the monitoring period is poor and that equipment supervision needs to be strengthened in the future. In this case, a high necessary signal for strong control is generated.

[0063] The working principle of the present invention is as follows: when in use, a multi-channel audio acquisition module is used to capture various sound signals generated by the device during operation and perform preliminary amplification and filtering processing; the adaptive noise suppression module uses an adaptive filtering algorithm based on deep learning to perform real-time noise suppression on the collected sound signals to improve the clarity of the fault feature signal; the fault feature extraction module uses audio signal processing technology to extract features of the sound signal after noise suppression processing; the intelligent fault diagnosis module determines whether there is a fault in the equipment and its specific type based on the feature extraction results; if there is an equipment fault, an early warning is issued through the remote monitoring alarm module, and corresponding processing measures can be taken in a timely and reasonable manner to avoid further deterioration of the fault, resulting in equipment damage or long-term production interruption; and the optimization decision output module is used to analyze the urgency of system repair optimization, optimize and repair the system performance when generating the optimization repair emergency signal, and verify the system's recognition performance for equipment faults in different noise scenarios after completing the system optimization and repair, to ensure that the system can accurately and quickly identify equipment faults and further improve the safety of equipment operation.

[0064] The above formulas are all dimensionless and calculated by taking their numerical values. The formula is a formula for the latest real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made based on the contents of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that technicians in the relevant technical field can well understand and use the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A high-precision detection system for equipment failure in a strong noise environment, characterized by: It includes a multi-channel audio acquisition module, an adaptive noise suppression module, a fault feature extraction module, an intelligent fault diagnosis module, and a remote monitoring and alarm module. The multi-channel audio acquisition module uses an array of multiple high-sensitivity, wide-bandwidth microphones to capture various sound signals generated by the equipment during operation, and performs preliminary amplification and filtering through the signal conditioning circuit. The adaptive noise suppression module uses an adaptive filtering algorithm based on deep learning to automatically adjust the filtering parameters according to changes in the noise environment, and performs real-time noise suppression on the collected sound signals to remove background noise; The fault feature extraction module uses audio signal processing technology to extract features from the sound signal after noise suppression processing, and sends the feature extraction results to the intelligent fault diagnosis module; The intelligent fault diagnosis module is used to build a fault identification and classification model. The fault identification and classification model determines whether there is a fault in the equipment and its specific type based on the feature extraction results, and sends the fault diagnosis information to the remote monitoring and alarm module. If there is an equipment fault, the remote monitoring and alarm module will issue an early warning and maintenance personnel will inspect and repair the equipment.

2. The high-precision detection system for equipment failure in a strong noise environment according to claim 1 is characterized in that: The remote monitoring alarm module is connected to the optimization decision output module in communication. The optimization decision output module compares the interval time with a preset interval time threshold. If the interval time exceeds the preset interval time threshold, an optimization repair emergency signal is generated. If the interval time does not exceed the preset interval time threshold, the optimization decision coefficient is obtained through comprehensive analysis of the optimization repair urgency. If the optimization decision coefficient exceeds the preset optimization decision coefficient threshold, an optimization repair emergency signal is generated and sent to the remote monitoring alarm module.

3. The high-precision detection system for equipment failure in a strong noise environment according to claim 2 is characterized in that: The specific analysis process of optimizing the comprehensive analysis of repair urgency is as follows: All analysis times during the monitoring period are obtained and their average is calculated to obtain the analysis time efficiency value, and the proportion of analysis times that exceed the preset analysis time threshold is marked as the analysis inefficiency value; the optimized decision coefficient is obtained by numerically calculating the analysis time efficiency value, analysis inefficiency value, diagnosis value and recognition value.

4. The high-precision detection system for equipment failure in a strong noise environment according to claim 2 is characterized in that: When the optimization decision output module generates an optimization repair emergency signal, the management personnel optimize and repair the system performance as needed. After completing the system optimization and repair, the system verifies the recognition accuracy, missed alarm rate, false alarm rate and analysis efficiency of equipment failures under different noise scenarios and determines whether the optimization and repair are qualified. If the optimization and repair are judged to be unqualified, the remote monitoring alarm module will issue an early warning.

5. The high-precision detection system for equipment failure in a strong noise environment according to claim 4 is characterized in that: The optimization decision output module communicates with the verification environment construction module. The optimization decision output module sends the optimization repair emergency signal to the verification environment construction module. Before verifying the system performance, the verification environment construction module builds the required verification environment and performs a construction accuracy analysis on the constructed verification environment to obtain a construction abnormality coefficient. If the construction abnormality coefficient exceeds the preset construction abnormality coefficient threshold, a construction alarm signal is generated and the construction alarm signal is sent to the remote monitoring alarm module.

6. The high-precision detection system for equipment failure in a strong noise environment according to claim 5 is characterized in that: The specific analysis process of the construction accuracy analysis is as follows: assign corresponding preset weight coefficients to the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value and cleanliness abnormal value of the constructed environment, multiply the noise abnormal value, temperature abnormal value, humidity abnormal value, air pressure abnormal value and cleanliness abnormal value with the corresponding preset weight coefficients respectively, and sum up the five groups of product results to obtain the construction anomaly coefficient.

7. The high-precision detection system for equipment failure in a strong noise environment according to claim 2 is characterized in that: The optimization decision output module is communicated with the strong management necessity decision module. If the optimization repair emergency signal is not generated, the strong management necessity decision module will analyze the operating performance of the equipment during the monitoring period, evaluate the degree of necessity of strengthening equipment management based on this, and determine whether to generate a strong management high necessity signal. The strong management high necessity signal is sent to the remote monitoring alarm module. When the remote monitoring alarm module receives the strong management high necessity signal, it reminds the management personnel to strengthen the subsequent supervision of the equipment.

8. The high-precision detection system for equipment failure in a strong noise environment according to claim 7 is characterized in that: The specific analysis process of the forced control necessity decision module is as follows: all equipment faults occurring during the monitoring period are obtained and classified. If a fault type with a frequency coefficient exceeding the corresponding preset frequency coefficient threshold exists during the monitoring period, a forced control necessity signal is generated; If there is no fault type whose frequency coefficient exceeds the corresponding preset frequency coefficient threshold during the monitoring period, the fault hazard values of all types of faults occurring during the monitoring period are summed up to obtain the fault hazard value. If the fault hazard value exceeds the preset fault hazard threshold, a strong control high necessary signal is generated.

9. The high-precision detection system for equipment failure in a strong noise environment according to claim 8, characterized in that: If the fault danger meter value does not exceed the preset fault danger meter threshold, the necessary coefficient for strong pipe is obtained by numerically calculating the fault detection value, operation impact value and fault damage value. If the necessary coefficient for strong pipe exceeds the preset necessary coefficient threshold, a high necessary signal for strong pipe is generated.

Images