Pump unit sealing fault detection method and system based on voiceprint representation

Abstract

The invention relates to the technical field of sealing detection, and discloses a pump unit sealing fault detection method and system based on voiceprint characterization, and the system comprises an acquisition unit, an analysis unit, an adjustment unit and an early warning unit. The collection unit sets a voiceprint sampling rate, and obtains sound data and vibration data of each detection point deployed in a plurality of detection points of the pump unit based on the voiceprint sampling rate, and the analysis unit extracts sound features of the target sound data and vibration features of the target vibration data. Sealing characteristics of the pump unit are calculated according to the sound characteristics and the vibration characteristics, a sealing fault prediction value of the pump unit is obtained based on a random forest model, environment temperature data of all detection points are obtained and compared with historical data, and whether the sealing fault prediction value is adjusted or not is judged according to the comparison result; the adjusting unit determines a sealing adjusting factor according to the temperature similarity; and the early warning unit gives out a corresponding alarm based on the sealing state. The device has good comprehensive analysis capability and sealing detection effect.

Description

Technical Field

[0001] The present invention relates to the technical field of seal detection, and in particular to a pump unit seal fault detection method and system based on voiceprint characterization. Background Art

[0002] In recent years, the phenomenon of oil pipeline shutdown caused by pump unit shutdown has occurred frequently. The sealing failure of the pump unit is an important factor leading to equipment performance degradation, unstable operation and even accidents. Therefore, accurate and efficient detection of the sealing status of the pump unit is an important link to ensure the safe operation of the equipment.

[0003] Traditional sealing fault detection mainly relies on a single data source, such as vibration signals. However, a single data source cannot fully reflect the sealing status of the pump unit, resulting in misjudgment or missed judgment, thereby reducing the reliability and accuracy of fault detection. In addition, the sealing status will be disturbed by the ambient temperature when detecting the sealing status, resulting in inaccurate judgment of the sealing fault. For example, the fluctuation of ambient temperature will affect the sensitivity of the sensor. The current sealing fault detection cannot be dynamically adjusted according to the ambient temperature, and cannot accurately determine the sealing status of the pump unit, and thus cannot make accurate warnings.

[0004] Therefore, it is necessary to design a pump unit seal fault detection method and system based on voiceprint characterization to solve the problems existing in the current technology. Summary of the invention

[0005] In view of this, the present invention proposes a pump unit seal fault detection method and system based on voiceprint characterization, aiming to solve the problem that the characteristics of multi-dimensional data cannot be fully utilized and the influence of ambient temperature on the detection results cannot be effectively dealt with.

[0006] In one aspect, the present invention proposes a pump unit seal fault detection system based on voiceprint characterization, comprising: Collection unit, analysis unit, adjustment unit and early warning unit; The acquisition unit is configured to set a soundprint sampling rate, obtain sound data and vibration data of each detection point deployed at a plurality of detection points of the pump unit based on the soundprint sampling rate, pre-process the sound data and the vibration data, and obtain target sound data and target vibration data respectively; The analysis unit is configured to extract the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, calculate the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, count the historical sealing characteristics of all detection points and establish a sealing data set, obtain the sealing failure prediction value of the pump unit based on the random forest model, obtain the ambient temperature data of each detection point and compare it with the historical data, and determine whether to adjust the sealing failure prediction value according to the comparison result; The adjustment unit is configured to determine a sealing adjustment factor according to the temperature similarity when it is determined that the sealing failure prediction value is to be adjusted; when the temperature similarity is greater than or equal to the similarity threshold, determine the sealing adjustment factor according to the historical data; when the temperature similarity is lower than the similarity threshold, determine a sealing temperature feature for a corresponding detection point according to the ambient temperature data, the sealing temperature feature including a first temperature feature and a second temperature feature, count the number of upstreams of the first temperature feature, count the number of downstreams of the second temperature feature, determine the sealing adjustment factor according to the upstream number and the downstream number, and adjust the sealing failure prediction value based on the sealing adjustment factor; The warning unit is configured to set a fault warning value, and compare the fault warning value with an adjusted sealing fault prediction value, derive a sealing state of the pump unit based on the comparison result, and issue a corresponding alarm based on the sealing state.

[0007] Further, when setting the voiceprint sampling rate, acquiring the sound data and vibration data of each detection point deployed at the multiple detection points of the pump unit based on the voiceprint sampling rate, and preprocessing the sound data and the vibration data, it includes: The voiceprint sampling rate includes the sound sampling rate and the vibration sampling rate; Acquire sound data of each detection point based on the sound sampling rate, and acquire vibration data of each detection point based on the vibration sampling rate; The sound data and the vibration data are preprocessed, and the preprocessing includes: noise filtering, signal segmentation and data normalization.

[0008] Further, when extracting the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, and calculating the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, it includes: The sealing characteristics of the pump unit are derived from the following formula: ; ; ; in, Indicates the sound characteristics, represents the number of detection points, represents the normalized result of the target sound data at the i-th detection point, Represents the standardized mean value of all detection point target sound data, Represents the vibration characteristics, represents the normalized result of the target vibration data of the i-th detection point, Represents the standardized result mean of the target vibration data of all detection points, Indicates the sealing characteristics of the pump unit, and represents the weight coefficient, and + =1.

[0009] Furthermore, when the historical sealing characteristics of all detection points are counted and a sealing data set is established, the sealing failure prediction value of the pump unit is obtained based on the random forest model, including: The sealing data set is used as a model training set and a model test set. Cross-validation is used in combination with grid search to find the construction parameters of the random forest model, and a random forest model is established. The model training set is used to fit the random forest model, and the model test set is input into the random forest model for verification. When the accuracy of the sealing failure prediction value reaches a preset accuracy threshold, the sealing failure prediction value of the pump unit is obtained according to the sealing characteristics.

[0010] Further, when the ambient temperature data of each detection point is obtained and compared with the historical data, and whether to adjust the sealing failure prediction value is determined according to the comparison result, it includes: The historical data includes the historical temperature qualified mean value, the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value, and the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value correspond to each other; Compare the ambient temperature data with the qualified mean value of historical temperature, and determine whether to adjust the sealing failure prediction value according to the comparison result; When the ambient temperature data are all equal to the qualified mean value of the historical temperature, it is determined that the sealing failure prediction value is not adjusted, and the pump unit failure value is determined according to the sealing failure prediction value; When the ambient temperature data is not equal to the qualified mean value of the historical temperature, it is determined that the sealing failure prediction value should be adjusted.

[0011] Further, when it is determined that the sealing failure prediction value is adjusted, a sealing adjustment factor is determined according to the temperature similarity, and when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data, including: ; in, represents the temperature similarity, represents the number of detection points, represents the standardized result of the ambient temperature data of the i-th detection point, Represents the qualified mean value of historical temperature, Indicates the temperature of the historical detection point; when When it is greater than or equal to the similarity threshold, the mean of the historical sealing adjustment factors is determined as the sealing adjustment factor of the sealing failure prediction value.

[0012] Further, when the temperature similarity is lower than the similarity threshold, a sealing temperature feature is determined for a corresponding detection point according to the ambient temperature data, and when the sealing temperature feature includes a first temperature feature and a second temperature feature, it includes: Determine the standard temperature data corresponding to each detection point; Comparing the ambient temperature data with the standard temperature data, if the ambient temperature data is less than the standard temperature data, determining the corresponding detection point as the first temperature feature; If the ambient temperature data is greater than or equal to the standard temperature data, the corresponding detection point is determined as the second temperature feature.

[0013] Further, when determining a sealing adjustment factor according to the upstream number and the downstream number, and adjusting the sealing failure prediction value based on the sealing adjustment factor, it includes: The seal adjustment factor is given by the following formula: ; in, represents the seal adjustment factor, Indicates the number of upstream connections. Indicates the number of downstream Indicates the sum of the upstream and downstream numbers; The seal failure prediction value is proportional to the seal adjustment factor, and the pump unit failure value is obtained.

[0014] Further, when a fault warning value is set, and the fault warning value is compared with the adjusted sealing fault prediction value, the sealing state of the pump unit is obtained based on the comparison result, and a corresponding alarm is issued based on the sealing state, it includes: Presetting a first preset fault warning value and a second preset fault warning value, wherein the first preset fault warning value is greater than the second preset fault warning value; When the pump unit fault value is greater than or equal to the first preset fault warning value, the sealing state of the pump unit is judged to be a first-level abnormality and a serious alarm is issued; When the pump unit fault value is less than the first preset fault warning value and greater than the second preset fault warning value, the sealing state of the pump unit is judged to be a secondary abnormality and a moderate alarm is issued; When the pump unit fault value is less than or equal to the second preset fault warning value, the sealing state of the pump unit is judged to be a level 3 abnormality and a slight alarm is issued.

[0015] Compared with the prior art, the beneficial effect of the present invention is that by collecting sound data and vibration data from multiple detection points of the pump unit and conducting comprehensive analysis in combination with ambient temperature data, the misjudgment problem caused by a single data source is effectively avoided, the ability of comprehensive data processing is improved, and the sealing status of the pump unit is accurately derived, thereby improving the accuracy of sealing fault detection. Through the random forest model and the dynamic adjustment mechanism based on temperature similarity, automatic data processing and prediction of sealing failures can be achieved, which reduces the need for manual intervention, improves detection efficiency and reduces the risk of errors caused by human factors. By comparing the ambient temperature data with the historical data, it can automatically determine whether the sealing failure prediction value needs to be adjusted, which can avoid the problem of environmental factors affecting the detection results. When dealing with the influence of ambient temperature, the sealing adjustment factor is determined by combining the upstream number and the downstream number, which fully considers the impact of temperature changes on the operating status of the pump unit, ensuring that accurate sealing failure detection can be maintained in complex environments, and improving the adaptability of detection. Based on the comparison between the adjusted sealing failure prediction value and the preset fault warning value, the sealing status of the pump unit is generated in time and the corresponding alarm is triggered, which can detect potential sealing failures at an early stage and effectively prevent the pump unit from shutting down due to sealing failures. At the same time, it reduces the risk of missed reports and false alarms, and improves the safety, stability and operating efficiency of the pump unit.

[0016] On the other hand, the present application also provides a pump unit seal fault detection method based on voiceprint characterization, which is used to apply the above-mentioned pump unit seal fault detection system based on voiceprint characterization, including: Setting a voiceprint sampling rate, acquiring sound data and vibration data of each detection point deployed at a plurality of detection points of the pump unit based on the voiceprint sampling rate, preprocessing the sound data and the vibration data, and obtaining target sound data and target vibration data respectively; Extract the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, calculate the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, count the historical sealing characteristics of all detection points and establish a sealing data set, obtain the sealing failure prediction value of the pump unit based on the random forest model, obtain the ambient temperature data of each detection point and compare it with the historical data, and determine whether to adjust the sealing failure prediction value according to the comparison result; When it is determined that the sealing failure prediction value is to be adjusted, a sealing adjustment factor is determined according to the temperature similarity; when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data; when the temperature similarity is lower than the similarity threshold, a sealing temperature feature is determined for a corresponding detection point according to the ambient temperature data, the sealing temperature feature includes a first temperature feature and a second temperature feature, the number of upstreams of the first temperature feature is counted, the number of downstreams of the second temperature feature is counted, a sealing adjustment factor is determined according to the upstream number and the downstream number, and the sealing failure prediction value is adjusted based on the sealing adjustment factor; A fault warning value is set, and the fault warning value is compared with the adjusted sealing fault prediction value, the sealing state of the pump unit is obtained based on the comparison result, and a corresponding alarm is issued based on the sealing state.

[0017] It can be understood that the above-mentioned pump unit seal fault detection method and system based on voiceprint characterization have the same beneficial effects, which will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 A functional block diagram of a pump unit sealing fault detection system based on voiceprint characterization provided by an embodiment of the present invention; Figure 2 A flow chart of a pump unit sealing fault detection method based on voiceprint characterization provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, in the absence of conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0020] In some embodiments of the present application, see Figure 1 As shown, a pump unit sealing fault detection system based on voiceprint characterization includes: Collection unit, analysis unit, adjustment unit and early warning unit; The acquisition unit is configured to set a voiceprint sampling rate, obtain sound data and vibration data of each detection point deployed at a plurality of detection points of the pump unit based on the voiceprint sampling rate, pre-process the sound data and vibration data, and obtain target sound data and target vibration data respectively; The analysis unit is configured to extract the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, calculate the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, count the historical sealing characteristics of all the detection points and establish a sealing data set, obtain the sealing failure prediction value of the pump unit based on the random forest model, obtain the ambient temperature data of each detection point and compare it with the historical data, and determine whether to adjust the sealing failure prediction value according to the comparison result; The adjustment unit is configured to determine a sealing adjustment factor according to the temperature similarity when it is determined that the sealing failure prediction value is to be adjusted; when the temperature similarity is greater than or equal to the similarity threshold, determine the sealing adjustment factor according to the historical data; when the temperature similarity is lower than the similarity threshold, determine the sealing temperature feature for the corresponding detection point according to the ambient temperature data, the sealing temperature feature includes a first temperature feature and a second temperature feature, count the number of uplinks of the first temperature feature, count the number of downlinks of the second temperature feature, determine the sealing adjustment factor according to the number of uplinks and the number of downlinks, and adjust the sealing failure prediction value based on the sealing adjustment factor; The early warning unit is configured to set a fault early warning value, and compare the fault early warning value with the adjusted sealing fault prediction value, derive the sealing status of the pump unit based on the comparison result, and issue a corresponding alarm based on the sealing status.

[0021] Specifically, the acquisition unit sets a voiceprint sampling rate and sets a suitable voiceprint sampling rate. The acquisition unit can effectively capture the sound data and vibration data of each detection point deployed at multiple detection points of the pump unit through sound sensors and vibration sensors. Since there will be signal interference between the collected data, the sound data and vibration data need to be preprocessed, and finally the target sound data and target vibration data that can represent the pump unit are obtained. The target sound data and target vibration data reflect the real-time status of the pump unit. Sound features are extracted from the target sound data, and vibration features are extracted from the target vibration data. Sound features and vibration features are two important features that constitute sealing features. The historical sealing features of all detection points are counted. The sealing data set is collected and established. The sealing data set represents the sealing characteristics of the pump unit at different times. The sealing data set is used to train the random forest model. The random forest model is a machine learning algorithm. By training the random forest model, the sealing failure prediction value of the pump unit can be obtained according to the input sealing characteristics. The prediction accuracy of the random forest model depends on the quality of the sealing characteristics and the training effect of the model. Simply using the model for prediction cannot comprehensively consider the impact of environmental factors on the sealing failure prediction value. The ambient temperature data of each detection point is obtained through the temperature sensor, and compared with the historical data to determine whether the sealing failure prediction value needs to be adjusted, thereby improving the flexibility and prediction accuracy of the system. When it is determined that the sealing failure prediction value needs to be adjusted, the sealing adjustment factor is determined according to the calculated temperature similarity. The sealing adjustment factor is a parameter that dynamically adjusts the sealing characteristics based on the ambient temperature, which improves the accuracy and reliability of the system prediction. A fault warning value is set, and the adjusted sealing failure prediction value is compared with the fault warning value. According to the comparison result, the sealing status of the pump unit is obtained and the corresponding alarm is issued, which provides processing time for maintenance personnel to make decisions, avoids the risk of major accidents caused by sealing failure of the pump unit, and is conducive to the safe management of the pump unit.

[0022] It is understandable that the sealing characteristics of the pump unit are calculated based on the sound characteristics and vibration characteristics, and analyzed in combination with the random forest model. The random forest model has good robustness in dealing with complex nonlinear problems, which improves the stability of the prediction. The introduction of ambient temperature data dynamically adjusts the sealing failure prediction value, allowing the system to adapt to different operating environments, thereby improving the reliability of the detection results. The system's real-time detection and automated analysis do not require frequent manual inspections, which reduces the operation and maintenance costs of the pump unit and ensures the detection of the sealing status of the pump unit, thereby reducing the risk of sealing failures.

[0023] In some embodiments of the present application, when setting the voiceprint sampling rate, acquiring the sound data and vibration data of each detection point deployed at multiple detection points of the pump unit based on the voiceprint sampling rate, and preprocessing the sound data and vibration data, it includes: The voiceprint sampling rate includes the sound sampling rate and the vibration sampling rate; Acquire sound data of each detection point based on the sound sampling rate, and acquire vibration data of each detection point based on the vibration sampling rate; The sound data and vibration data are preprocessed, and the preprocessing includes: noise filtering, signal segmentation and data normalization.

[0024] In some embodiments of the present application, when extracting the sound features of the target sound data and the vibration features of the target vibration data, and calculating the sealing features of the pump unit according to the sound features and the vibration features, it includes: The sealing characteristics of the pump unit are given by the following formula: ; ; ; in, Indicates the sound characteristics, represents the number of detection points, represents the normalized result of the target sound data at the i-th detection point, Represents the standardized mean value of all detection point target sound data, Represents the vibration characteristics, represents the normalized result of the target vibration data of the i-th detection point, Represents the standardized result mean of the target vibration data of all detection points, Indicates the sealing characteristics of the pump unit, and represents the weight coefficient, and + =1.

[0025] Specifically, the sound sampling rate is preferably 15kHz-30kHz, and the vibration sampling rate is preferably 5kHz-20kHz. The sound sampling rate and vibration sampling rate can be set according to specific needs. Noise filtering can filter out environmental noise, and signal segmentation ensures that the sound data and vibration data have time continuity during feature extraction. Data normalization unifies the dimensions of the data. The accuracy of the sound features and vibration features is improved through noise filtering, signal segmentation and data normalization. The standardized result mean of the target sound data and the standardized result mean of the target vibration data ensure that the calculation of the sealing features is carried out under the same dimension. The sealing features of the pump unit are calculated based on the sound features and vibration features, which reduces the reliance on manual experience, reduces the impact of human errors on the sealing state of the pump unit, and improves the automation and reliability of the system.

[0026] In some embodiments of the present application, when the historical sealing features of all detection points are counted and a sealing data set is established, and a sealing failure prediction value of the pump unit is obtained based on a random forest model, it includes: The sealing data set is used as the model training set and model test set. Cross-validation is used in combination with grid search to find the construction parameters of the random forest model. The random forest model is established. The model training set is used to fit the random forest model. The model test set is input into the random forest model for verification. When the accuracy of the sealing failure prediction value reaches the preset accuracy threshold, the sealing failure prediction value of the pump unit is obtained according to the sealing characteristics.

[0027] Specifically, the sealing data set records the sealing characteristics of the pump unit in different time periods. The sealing data set is divided into a model training set and a model test set. 80%-90% of the data is used as the model training set, and the rest is used as the model test set to ensure that both the model training set and the model test set contain data on a variety of sealing characteristics, thereby improving the generalization ability of the random forest model. Cross-validation is used in combination with grid search to find the construction parameters of the random forest model. Cross-validation trains the model multiple times to verify its stability, while grid search searches for the construction parameters of the random forest model in the parameters, such as the number of trees. Secondly, the model training set is used to fit The random forest model reduces the risk of overfitting by integrating multiple decision trees and averaging the prediction results of multiple trees, thereby improving the stability of the random forest model. The model test set is input into the trained random forest model to calculate the prediction accuracy of the random forest model. The accuracy reflects the performance of the random forest model on unknown data. When the random forest model reaches the preset accuracy threshold, the current sealing characteristics are input into the random forest model to obtain the sealing failure prediction value of the pump unit, which improves the system's ability to efficiently and accurately predict sealing failures of pump units, thereby timely discovering potential problems and enhancing safety management of pump units.

[0028] In some embodiments of the present application, when the ambient temperature data of each detection point is obtained and compared with the historical data, and whether to adjust the sealing failure prediction value according to the comparison result is determined, it includes: The historical data includes the historical temperature qualified mean value, the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value, and the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value correspond to each other; Compare the ambient temperature data with the qualified mean value of historical temperature, and determine whether to adjust the sealing failure prediction value based on the comparison result; When the ambient temperature data are all equal to the qualified mean value of the historical temperature, it is determined that the seal failure prediction value is not adjusted, and the pump unit failure value is determined according to the seal failure prediction value; When the ambient temperature data is not equal to the qualified mean value of the historical temperature, it is determined that the sealing failure prediction value should be adjusted.

[0029] Specifically, the historical temperature qualified mean can be adjusted according to the temperature environment of the pump unit. For example, when the pump unit is in a 70°C environment for a long time, the historical temperature qualified mean can be set to 65-75°C to ensure that it is within the allowable temperature error range of the pump unit. The ambient temperature will have a certain impact on the seal of the pump unit. By introducing the historical temperature qualified mean as the judgment basis, the accuracy and automation of the adjustment of the sealing failure prediction value are improved, the interference and error of human judgment are reduced, and the system's adaptability to different environmental conditions is enhanced.

[0030] In some embodiments of the present application, when it is determined that the sealing failure prediction value is adjusted, the sealing adjustment factor is determined according to the temperature similarity, and when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data, including: ; in, represents the temperature similarity, represents the number of detection points, represents the standardized result of the ambient temperature data of the i-th detection point, Represents the qualified mean value of historical temperature, Indicates the temperature of the historical detection point; when When it is greater than or equal to the similarity threshold, the mean of the historical sealing adjustment factors is determined as the sealing adjustment factor of the sealing failure prediction value.

[0031] Specifically, the degree of match between ambient temperature data and historical adjustments is determined through a similarity threshold. When historical data with a high temperature similarity is found, these data can be used to determine the sealing adjustment factor. By comprehensively utilizing historical data, rich reference data is provided for the determination of the sealing adjustment factor, thereby improving the system's automation level and ensuring the reliability of the adjustment of the sealing failure prediction value.

[0032] In some embodiments of the present application, when the temperature similarity is lower than the similarity threshold, the sealing temperature feature is determined for the corresponding detection point according to the ambient temperature data, and when the sealing temperature feature includes the first temperature feature and the second temperature feature, it includes: Determine the standard temperature data corresponding to each detection point; The ambient temperature data is compared with the standard temperature data, and if the ambient temperature data is less than the standard temperature data, the corresponding detection point is determined as the first temperature feature; If the ambient temperature data is greater than or equal to the standard temperature data, the corresponding detection point is determined as the second temperature feature.

[0033] In some embodiments of the present application, when a sealing adjustment factor is determined according to the number of uplinks and the number of downlinks, and the sealing failure prediction value is adjusted based on the sealing adjustment factor, the method includes: The seal adjustment factor is given by the following formula: ; in, represents the seal adjustment factor, Indicates the number of upstream connections. Indicates the number of downstream Indicates the sum of the upstream and downstream numbers; The seal failure prediction value is proportional to the seal adjustment factor, and the pump unit failure value is obtained.

[0034] Specifically, the standard temperature data corresponding to each detection point is the allowable working temperature environment of the pump unit. The sealing adjustment factor is determined by the upstream number and the downstream number, and then the dynamic adjustment of the sealing failure prediction value can be realized. The upstream number and the downstream number reflect the distribution of the standard temperature data. The sealing failure prediction value is adjusted according to the sealing adjustment factor. Assuming that the sealing failure prediction value is A and the sealing adjustment factor is B, the adjusted sealing failure prediction value is determined to be the pump unit failure value. The pump unit failure value is equal to A*B, which improves the flexibility of dynamic adjustment of the system.

[0035] In some embodiments of the present application, when a fault warning value is set, and the fault warning value is compared with the adjusted sealing fault prediction value, the sealing state of the pump unit is obtained based on the comparison result, and a corresponding alarm is issued based on the sealing state, it includes: Presetting a first preset fault warning value and a second preset fault warning value, wherein the first preset fault warning value is greater than the second preset fault warning value; When the pump unit fault value is greater than or equal to the first preset fault warning value, the sealing state of the pump unit is judged to be a first-level abnormality and a serious alarm is issued; When the pump unit fault value is less than the first preset fault warning value and greater than the second preset fault warning value, the sealing state of the pump unit is judged to be a secondary abnormality and a moderate alarm is issued; When the pump unit fault value is less than or equal to the second preset fault warning value, the sealing state of the pump unit is judged to be a level 3 abnormality and a slight alarm is issued.

[0036] Specifically, by setting the first preset fault warning value and the second preset fault warning value, the sealing status of the pump unit can be judged according to the pump unit fault value and a corresponding alarm can be issued. The hierarchical alarm mechanism can effectively remind the current sealing status of the pump unit and improve the response efficiency of sealing fault handling. It avoids missed reports and false alarms caused by human detection, reduces the risk of ignoring sealing failures, and thus reduces the probability of accidents.

[0037] To sum up, the beneficial effect of the present invention is that by collecting sound data and vibration data from multiple detection points of the pump unit and conducting comprehensive analysis in combination with ambient temperature data, the misjudgment problem caused by a single data source is effectively avoided, the ability of comprehensive data processing is improved, and the sealing status of the pump unit is accurately derived, thereby improving the accuracy of sealing fault detection. Through the random forest model and the dynamic adjustment mechanism based on temperature similarity, automatic data processing and prediction of sealing failures can be achieved, which reduces the need for manual intervention, improves detection efficiency and reduces the risk of errors caused by human factors. By comparing the ambient temperature data with the historical data, it can automatically determine whether the sealing failure prediction value needs to be adjusted, which can avoid the problem of environmental factors affecting the detection results. When dealing with the influence of ambient temperature, the sealing adjustment factor is determined by combining the upstream number and the downstream number, which fully considers the impact of temperature changes on the operating status of the pump unit, ensuring that accurate sealing failure detection can be maintained in complex environments, and improving the adaptability of detection. Based on the comparison between the adjusted sealing failure prediction value and the preset fault warning value, the sealing status of the pump unit is generated in time and the corresponding alarm is triggered, which can detect potential sealing failures at an early stage and effectively prevent the pump unit from shutting down due to sealing failures. At the same time, it reduces the risk of missed reports and false alarms, and improves the safety, stability and operating efficiency of the pump unit.

[0038] In another preferred embodiment based on the above embodiment, refer to Figure 2 As shown, this embodiment provides a pump unit sealing fault detection method based on voiceprint characterization, which is used to apply the above-mentioned pump unit sealing fault detection system based on voiceprint characterization, including: S100: setting a voiceprint sampling rate, acquiring sound data and vibration data of each detection point deployed at multiple detection points of the pump unit based on the voiceprint sampling rate, preprocessing the sound data and vibration data, and obtaining target sound data and target vibration data respectively; S200: extracting the sound features of the target sound data and the vibration features of the target vibration data, calculating the sealing features of the pump unit according to the sound features and the vibration features, counting the historical sealing features of all the detection points and establishing a sealing data set, obtaining the sealing failure prediction value of the pump unit based on the random forest model, obtaining the ambient temperature data of each detection point and comparing it with the historical data, and judging whether to adjust the sealing failure prediction value according to the comparison result; S300: when it is determined that the sealing failure prediction value is to be adjusted, a sealing adjustment factor is determined according to the temperature similarity; when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data; when the temperature similarity is lower than the similarity threshold, a sealing temperature feature is determined for a corresponding detection point according to the ambient temperature data, the sealing temperature feature includes a first temperature feature and a second temperature feature, the number of uplinks of the first temperature feature is counted, the number of downlinks of the second temperature feature is counted, the sealing adjustment factor is determined according to the uplink number and the downlink number, and the sealing failure prediction value is adjusted based on the sealing adjustment factor; S400: Setting a fault warning value, and comparing the fault warning value with the adjusted sealing fault prediction value, obtaining the sealing status of the pump unit based on the comparison result, and issuing a corresponding alarm based on the sealing status.

[0039] Specifically, in step S100, the voiceprint sampling rate is set, and sound data and vibration data are obtained from each detection point of the pump unit according to the voiceprint sampling rate. The sound data comes from the sound sensor deployed at each detection point, and the vibration data comes from the vibration sensor deployed at each detection point. The sound data and vibration data are preprocessed to obtain target sound data and target vibration data, respectively, which improves the quality of the preprocessed data and provides data support for subsequent analysis. In step S200, the sound features and vibration features of the target sound data and target vibration data are extracted respectively. By analyzing these features, the sealing features of the pump unit are calculated. The historical sealing features of all detection points are counted, a sealing data set is established, and the sealing fault value of the pump unit is obtained by training a random forest model. The machine learning model is used to improve the accuracy of the prediction. The ambient temperature data of each detection point is obtained. The ambient temperature data comes from the temperature sensor deployed at each detection point. The ambient temperature data is compared with the historical data to determine whether to adjust the sealing failure prediction value to improve the accuracy of the prediction result. In step S300, when it is determined to adjust the sealing failure prediction value, the sealing adjustment factor is determined according to the temperature similarity, which improves the flexibility and adaptability of the adjustment. When the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data. When the temperature similarity is lower than the similarity threshold, the sealing temperature characteristics of the corresponding detection point are determined according to the ambient temperature data. The sealing temperature characteristics include the first temperature characteristics and the second temperature characteristics. The sealing adjustment factor is determined by counting the uplink and downlink numbers of these two characteristics. The sealing failure prediction value is adjusted according to the sealing adjustment factor to improve the accuracy and reliability of the prediction. In step S400, the fault warning value is set, the fault warning value is compared with the adjusted sealing failure prediction value, the sealing state of the pump unit is determined, and a corresponding alarm signal is issued according to the sealing state. According to the corresponding alarm signal, the extent of the sealing failure can be judged to ensure that the sealing failure can be handled in time, avoiding further damage to the pump unit and effectively improving the operating safety of the pump unit.

[0040] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0041] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0042] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0043] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0044] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A pump unit seal fault detection system based on voiceprint characterization, characterized in that: include: Collection unit, analysis unit, adjustment unit and early warning unit; The acquisition unit is configured to set a soundprint sampling rate, obtain sound data and vibration data of each detection point deployed at a plurality of detection points of the pump unit based on the soundprint sampling rate, pre-process the sound data and the vibration data, and obtain target sound data and target vibration data respectively; The analysis unit is configured to extract the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, calculate the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, count the historical sealing characteristics of all detection points and establish a sealing data set, obtain the sealing failure prediction value of the pump unit based on the random forest model, obtain the ambient temperature data of each detection point and compare it with the historical data, and determine whether to adjust the sealing failure prediction value according to the comparison result; The adjustment unit is configured to determine a sealing adjustment factor according to the temperature similarity when it is determined that the sealing failure prediction value is to be adjusted; when the temperature similarity is greater than or equal to the similarity threshold, determine the sealing adjustment factor according to the historical data; when the temperature similarity is lower than the similarity threshold, determine a sealing temperature feature for a corresponding detection point according to the ambient temperature data, the sealing temperature feature including a first temperature feature and a second temperature feature, count the number of upstreams of the first temperature feature, count the number of downstreams of the second temperature feature, determine the sealing adjustment factor according to the upstream number and the downstream number, and adjust the sealing failure prediction value based on the sealing adjustment factor; The warning unit is configured to set a fault warning value, and compare the fault warning value with an adjusted sealing fault prediction value, derive a sealing state of the pump unit based on the comparison result, and issue a corresponding alarm based on the sealing state.

2. The pump unit seal fault detection system based on voiceprint characterization according to claim 1 is characterized in that: When setting the voiceprint sampling rate, acquiring the sound data and vibration data of each detection point deployed at the multiple detection points of the pump unit based on the voiceprint sampling rate, and preprocessing the sound data and the vibration data, it includes: The voiceprint sampling rate includes the sound sampling rate and the vibration sampling rate; Acquire sound data of each detection point based on the sound sampling rate, and acquire vibration data of each detection point based on the vibration sampling rate; The sound data and the vibration data are preprocessed, and the preprocessing includes: noise filtering, signal segmentation and data normalization.

3. The pump unit seal fault detection system based on voiceprint characterization according to claim 2 is characterized in that: When extracting the sound feature of the target sound data and the vibration feature of the target vibration data, and calculating the sealing feature of the pump unit according to the sound feature and the vibration feature, it includes: The sealing characteristics of the pump unit are derived from the following formula: ； ； ； in, Indicates the sound characteristics, represents the number of detection points, represents the normalized result of the target sound data at the i-th detection point, Represents the standardized mean value of all detection point target sound data, Represents the vibration characteristics, represents the normalized result of the target vibration data of the i-th detection point, Represents the standardized result mean of the target vibration data of all detection points, Indicates the sealing characteristics of the pump unit, and represents the weight coefficient, and + =1.

4. The pump unit seal fault detection system based on voiceprint characterization according to claim 3 is characterized in that: When the historical sealing characteristics of all detection points are counted and a sealing data set is established, and the sealing failure prediction value of the pump unit is obtained based on the random forest model, it includes: The sealing data set is used as a model training set and a model test set. Cross-validation is used in combination with grid search to find the construction parameters of the random forest model, and a random forest model is established. The model training set is used to fit the random forest model, and the model test set is input into the random forest model for verification. When the accuracy of the sealing failure prediction value reaches a preset accuracy threshold, the sealing failure prediction value of the pump unit is obtained according to the sealing characteristics.

5. The pump unit seal fault detection system based on voiceprint characterization according to claim 4 is characterized in that: When the ambient temperature data of each detection point is obtained and compared with the historical data, and whether to adjust the sealing failure prediction value according to the comparison result is determined, it includes: The historical data includes the historical temperature qualified mean value, the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value, and the historical detection point temperature, the historical sealing adjustment factor and the historical sealing failure prediction value correspond to each other; Compare the ambient temperature data with the qualified mean value of historical temperature, and determine whether to adjust the sealing failure prediction value according to the comparison result; When the ambient temperature data are all equal to the qualified mean value of the historical temperature, it is determined that the sealing failure prediction value is not adjusted, and the pump unit failure value is determined according to the sealing failure prediction value; When the ambient temperature data is not equal to the qualified mean value of the historical temperature, it is determined that the sealing failure prediction value should be adjusted.

6. The pump unit seal fault detection system based on voiceprint characterization according to claim 5 is characterized in that: When it is determined that the sealing failure prediction value is to be adjusted, a sealing adjustment factor is determined according to the temperature similarity, and when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data, including: ； in, represents the temperature similarity, represents the number of detection points, represents the standardized result of the ambient temperature data of the i-th detection point, Represents the qualified mean value of historical temperature, Indicates the temperature of the historical detection point; when When it is greater than or equal to the similarity threshold, the mean value of the historical sealing adjustment factors is determined as the sealing adjustment factor of the sealing failure prediction value.

7. The pump unit seal fault detection system based on voiceprint characterization according to claim 6 is characterized in that: When the temperature similarity is lower than the similarity threshold, determining the sealing temperature feature for the corresponding detection point according to the ambient temperature data, and when the sealing temperature feature includes the first temperature feature and the second temperature feature, comprising: Determine the standard temperature data corresponding to each detection point; Comparing the ambient temperature data with the standard temperature data, if the ambient temperature data is less than the standard temperature data, determining the corresponding detection point as the first temperature feature; If the ambient temperature data is greater than or equal to the standard temperature data, the corresponding detection point is determined as the second temperature feature.

8. The pump unit seal fault detection system based on voiceprint characterization according to claim 7 is characterized in that: When a sealing adjustment factor is determined according to the upstream number and the downstream number, and the sealing failure prediction value is adjusted based on the sealing adjustment factor, the method includes: The seal adjustment factor is given by the following formula: ； in, represents the seal adjustment factor, Indicates the number of upstream connections. Indicates the number of downstream Indicates the sum of the upstream and downstream numbers; The seal failure prediction value is proportional to the seal adjustment factor, and the pump unit failure value is obtained.

9. The pump unit seal fault detection system based on voiceprint characterization according to claim 8 is characterized in that: When a fault warning value is set, and the fault warning value is compared with the adjusted sealing fault prediction value, the sealing state of the pump unit is obtained based on the comparison result, and a corresponding alarm is issued based on the sealing state, it includes: Presetting a first preset fault warning value and a second preset fault warning value, wherein the first preset fault warning value is greater than the second preset fault warning value; When the pump unit fault value is greater than or equal to the first preset fault warning value, the sealing state of the pump unit is judged to be a first-level abnormality and a serious alarm is issued; When the pump unit fault value is less than the first preset fault warning value and greater than the second preset fault warning value, the sealing state of the pump unit is judged to be a secondary abnormality and a moderate alarm is issued; When the pump unit fault value is less than or equal to the second preset fault warning value, the sealing state of the pump unit is judged to be a level 3 abnormality and a slight alarm is issued.

10. A pump unit seal fault detection method based on voiceprint characterization, using the pump unit seal fault detection system based on voiceprint characterization as described in any one of claims 1 to 9, characterized in that: include: Setting a soundprint sampling rate, acquiring sound data and vibration data of each detection point deployed at a plurality of detection points of the pump unit based on the soundprint sampling rate, preprocessing the sound data and the vibration data, and obtaining target sound data and target vibration data respectively; Extract the sound characteristics of the target sound data and the vibration characteristics of the target vibration data, calculate the sealing characteristics of the pump unit according to the sound characteristics and the vibration characteristics, count the historical sealing characteristics of all detection points and establish a sealing data set, obtain the sealing failure prediction value of the pump unit based on the random forest model, obtain the ambient temperature data of each detection point and compare it with the historical data, and determine whether to adjust the sealing failure prediction value according to the comparison result; When it is determined that the sealing failure prediction value is to be adjusted, a sealing adjustment factor is determined according to the temperature similarity; when the temperature similarity is greater than or equal to the similarity threshold, the sealing adjustment factor is determined according to the historical data; when the temperature similarity is lower than the similarity threshold, a sealing temperature feature is determined for a corresponding detection point according to the ambient temperature data, the sealing temperature feature includes a first temperature feature and a second temperature feature, the number of upstreams of the first temperature feature is counted, the number of downstreams of the second temperature feature is counted, a sealing adjustment factor is determined according to the upstream number and the downstream number, and the sealing failure prediction value is adjusted based on the sealing adjustment factor; A fault warning value is set, and the fault warning value is compared with the adjusted sealing fault prediction value, the sealing state of the pump unit is obtained based on the comparison result, and a corresponding alarm is issued based on the sealing state.

Images