Quality inspection method based on noise analysis technology of gantry crane

Abstract

The invention relates to a quality inspection method based on a noise analysis technology of a gantry crane system, and the method comprises the steps: carrying out the noise monitoring of a plurality of pieces of target equipment through a first detection method and a second detection method, and obtaining the first monitoring audio data and second monitoring audio data of each piece of target equipment; performing clustering analysis on the obtained multiple pieces of first monitoring audio data to obtain a first operation curve, and then comparing each piece of first monitoring audio data with the first operation curve to generate a preliminary judgment result of each piece of target equipment; thirdly, performing clustering analysis on the obtained multiple pieces of second monitoring audio data to obtain a second operation curve, and comparing the multiple pieces of first monitoring audio data of each piece of target equipment with the obtained second operation curve to obtain a second operation curve; and further generating a secondary judgment result of each target device based on the preliminary judgment result so as to increase the judgment accuracy of each target device.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of noise analysis, in particular to a quality inspection method based on a noise analysis technology of a portal crane. BACKGROUND

[0002] The portal crane is an important hoisting and transporting device, and undertakes a large number of cargo loading and unloading tasks in ports, stations and yards. With the development of economy and the expansion of industrial production scale, the portal crane is used more and more widely, and its safety, stability and efficiency are crucial to the production and operation of enterprises.

[0003] At present, important parts of the crane need to be detected in the production process, so that the crane has stable working performance after assembly, but the existing quality inspection method for important parts of the crane lacks detection of the parts in actual operation, so the accuracy of the detection result is poor. SUMMARY

[0004] The purpose of the application is to provide a quality inspection method based on a noise analysis technology of a portal crane to solve the problems existing in the prior art.

[0005] The above technical purpose of the application is achieved by the following technical scheme: A quality inspection method based on a noise analysis technology of a portal crane, comprising a first detection method and a second detection method, the first detection method is used to obtain a first type of data, the second detection method is used to obtain a second type of data, and the first detection method is different from the second detection method; The quality inspection method based on the noise analysis technology of the portal crane comprises: based on the first detection method, noise monitoring is performed on a plurality of target devices to obtain a plurality of first monitoring audio data; The obtained plurality of first monitoring audio data are subjected to cluster analysis to obtain a first running curve; Based on the obtained first running curve, a preliminary judgment result of each target device is generated; Based on the second detection method, noise monitoring is performed on the plurality of target devices to obtain a plurality of second monitoring audio data; The obtained plurality of second monitoring audio data are subjected to cluster analysis to obtain a plurality of second running curves; Based on the obtained plurality of second running curves, a final judgment result of each target device is generated.

[0006] Further, the noise monitoring based on the first detection method on the plurality of target devices to obtain the plurality of first monitoring audio data comprises: A first detection time interval is obtained; The noise of the same monitoring point of the plurality of target devices is monitored based on the first detection time interval to obtain first monitoring audio data of each target device.

[0007] Further, the obtained plurality of first monitoring audio data is subjected to cluster analysis to obtain a first operation curve, including: selecting one first monitoring audio data; processing the first monitoring audio data to obtain audio restoration data of the first monitoring audio data; converting the audio restoration data into a first density curve and a first energy curve; returning to the selecting one first monitoring audio data until each of the first monitoring audio data is selected once.

[0008] Further, the obtained plurality of first monitoring audio data is subjected to cluster analysis to obtain a plurality of first operation curves, further including: classifying a plurality of first density curves based on a cluster analysis method to obtain a first reference density curve; classifying a plurality of first energy curves based on a cluster analysis method to obtain a first reference energy curve.

[0009] Further, the obtained plurality of first operation curves are used to generate a preliminary judgment result of each target device, including: obtaining a first matching threshold and a second matching threshold; calculating a matching value of each first density curve with the first reference density curve, respectively, and selecting all first density curves with a matching value less than or equal to the first matching threshold, and the target devices corresponding to the selected first density curves are defined as having performance problems; calculating a matching value of each first energy curve with the first reference energy curve, respectively, and selecting first energy curves with a matching value less than or equal to the second matching threshold, and the target devices corresponding to the selected first energy curves are defined as having performance problems.

[0010] Further, the plurality of target devices are subjected to noise monitoring based on a second detection method to obtain a plurality of second monitoring audio data, including: obtaining a second detection time interval; monitoring the noise of the same monitoring point on each target device based on the first detection time interval to obtain a plurality of second monitoring audio data of each target device.

[0011] Further, the obtained plurality of second monitoring audio data is subjected to cluster analysis to obtain a plurality of second operation curves, including: selecting one second monitoring audio data; processing the second monitoring audio data to obtain audio restoration data of the second monitoring audio data; transforming the audio restoration data into a second density curve and a second energy curve; returning to selecting one second monitoring audio data until each of the second monitoring audio data is selected once.

[0012] Further, the obtained multiple second monitoring audio data are subjected to cluster analysis to obtain multiple second running curves, and the method further comprises: classifying the multiple second density curves based on a cluster analysis method to obtain multiple second reference density curves; classifying the multiple second energy curves based on a cluster analysis method to obtain multiple second reference energy curves.

[0013] Further, the method of generating the final judgment result of each target device based on the obtained multiple second running curves comprises: obtaining a first attenuation threshold and a second attenuation threshold; calculating a matching value of each second density curve with the second reference density curve, respectively; calculating a first matching value difference between all adjacent two second density curves based on time sequence; judging whether there is any first matching value difference greater than or equal to the first attenuation threshold; if there is any first matching value difference greater than or equal to the first attenuation threshold, the target device corresponding to the second density curve has performance problems; calculating a matching value of each second energy curve with the second reference density curve, respectively; calculating a second matching value difference between all adjacent two second energy curves based on time sequence; judging whether there is any second matching value difference greater than or equal to the second attenuation threshold; if there is any second matching value difference greater than or equal to the second attenuation threshold, the target device corresponding to the second energy curve has performance problems.

[0014] Further, the first detection time interval comprises a continuous and fixed time interval; the second detection time interval comprises multiple time intervals with the same time interval.

[0015] This application relates to a quality inspection method based on noise analysis technology for a gantry crane system. The method involves using a first detection method to monitor noise from multiple target devices, obtaining first monitoring audio data for each target device. Cluster analysis is then performed on the obtained first monitoring audio data to obtain a first operating curve. Next, by comparing each first monitoring audio data with the first operating curve, a preliminary judgment result is generated for each target device. Then, a second detection method is used to monitor noise from the multiple target devices, obtaining multiple second monitoring audio data for each target device. Cluster analysis is then performed on the obtained second monitoring audio data to obtain a second operating curve. By comparing the multiple first monitoring audio data of each target device with the obtained second operating curve, a further judgment result based on the preliminary judgment result is generated for each target device, thereby increasing the accuracy of the judgment for each target device. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a quality inspection method based on noise analysis technology for a gantry crane system, provided as an embodiment of this application. Detailed Implementation

[0017] The present invention will be further described in detail below with reference to the accompanying drawings.

[0018] Identical parts are indicated by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to the attached figures. Figure 1 In this specification, the terms "bottom surface" and "top surface," "inner" and "outer" refer to the direction toward or away from the geometry of a specific component. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this specification, "a plurality of" means two or more, unless otherwise explicitly and specifically defined by the direction of the center.

[0019] Example 1: like Figure 1 As shown, in one embodiment of this application, a first detection method and a second detection method are included. The first detection method is used to obtain a first type of data, and the second detection method is used to obtain a second type of data. The first detection method and the second detection method are different. Quality inspection methods based on noise analysis technology for gantry cranes include: S100. Based on the first detection method, noise monitoring is performed on multiple target devices to obtain multiple first monitoring audio data.

[0020] Specifically, the target device can refer to an important working component such as a motor in a crane, for example, a motor, and the plurality of target devices refer to a plurality of motors of the same type. The noise monitoring refers to monitoring the audio data of the motor in operation.

[0021] S200, performing clustering analysis on the obtained plurality of first monitoring audio data to obtain a first operation curve.

[0022] S300, generating a preliminary judgment result of each target device based on the obtained first operation curve.

[0023] Specifically, the preliminary judgment result refers to the judgment result of the operation condition of the target device formed based on the data obtained by the first detection method. The operation condition of the target device can refer to the operation state of the motor in normal use of the crane, and can also refer to the operation state of the motor for quality inspection of the production.

[0024] S400, performing noise monitoring on the plurality of target devices based on a second detection method to obtain a plurality of second monitoring audio data.

[0025] S500, performing clustering analysis on the obtained plurality of second monitoring audio data to obtain a plurality of second operation curves.

[0026] S600, generating a final judgment result of each target device based on the obtained plurality of second operation curves.

[0027] In the embodiment, the first detection method is used to perform noise monitoring on the plurality of target devices, and then the first monitoring audio data of each target device is obtained. The clustering analysis is performed on the obtained plurality of first monitoring audio data, and then the first operation curve is obtained. Then, the first monitoring audio data of each target device is compared with the first operation curve, and then the preliminary judgment result of each target device is generated. Then, the second detection method is used to perform noise monitoring on the plurality of target devices, and then the plurality of second monitoring audio data of each target device is obtained. The clustering analysis is performed on the obtained plurality of second monitoring audio data, and then the second operation curve is obtained. The plurality of first monitoring audio data of each target device is compared with the obtained second operation curve, and then the further judgment result of each target device based on the preliminary judgment result is generated, and then the accuracy of the judgment of each target device is increased.

[0028] In an embodiment of the present application, the noise monitoring on the plurality of target devices based on the first detection method to obtain the plurality of first monitoring audio data comprises: S101, obtaining a first detection time interval.

[0029] Specifically, the first detection time interval includes a continuous and fixed time interval.

[0030] S102, noise monitoring is performed on the same monitoring point of the plurality of target devices based on a first detection time interval, to obtain first monitoring audio data of each target device.

[0031] In the embodiment, noise monitoring is performed on the same monitoring point of a plurality of target devices of the same type within a set first detection time interval, to obtain first monitoring audio data corresponding to each target device.

[0032] In an embodiment of the present application, the obtained plurality of first monitoring audio data are subjected to cluster analysis to obtain a first operation curve, including: S201, selecting one first monitoring audio data.

[0033] S202, processing the first monitoring audio data to obtain audio restoration data of the first monitoring audio data.

[0034] S203, converting the audio restoration data into a first density curve and a first energy curve.

[0035] S204, returning to the selecting one first monitoring audio data until each of the first monitoring audio data is selected once.

[0036] In the embodiment, the first monitoring audio data corresponding to each target device are subjected to restoration processing to obtain audio restoration data corresponding to each target device, and then the obtained audio restoration data is converted into a first density curve and a first energy curve.

[0037] In an embodiment of the present application, the obtained first monitoring audio data are subjected to cluster analysis to obtain a plurality of first operation curves, further including: S205, classifying a plurality of first density curves based on a cluster analysis method to obtain a first reference density curve.

[0038] S206, classifying a plurality of first energy curves based on a cluster analysis method to obtain a first reference energy curve.

[0039] The first operation curve is based on the obtained first operation curve to generate a preliminary judgment result of each target device, including: S301, obtaining a first matching threshold and a second matching threshold.

[0040] S302, calculating a matching value of each first density curve with the first reference density curve, respectively, and selecting all first density curves with a matching value less than or equal to the first matching threshold, and the target devices corresponding to the selected first density curves are defined as having performance problems.

[0041] S303, calculate the matching value of each first energy curve with the first reference energy curve respectively, and screen out the first energy curve with a matching value less than or equal to the second matching threshold value, and the target device corresponding to all the screened first energy curves is defined as having a performance problem.

[0042] Specifically, in a normal case, the performance difference between multiple devices of the same model will not be too large. If the performance of a certain device is too outstanding, it indicates that the performance of the device has a problem and needs to be further determined.

[0043] In this embodiment, the obtained multiple first density curves are classified by cluster analysis to obtain the first reference density curve, and then the obtained multiple first energy curves are classified by cluster analysis to obtain the first reference energy curve.

[0044] The first matching threshold value and the second matching threshold value are obtained; the matching value of each first density curve with the first reference density curve is calculated, and the obtained matching value is compared with the first matching threshold value. If the matching value is less than or equal to the first matching threshold value, it is determined that the target device corresponding to the first density curve has a performance problem.

[0045] The matching value of each first energy curve with the first reference energy curve is calculated, and the obtained matching value is compared with the second matching threshold value. If the matching value is less than or equal to the second matching threshold value, it is determined that the target device corresponding to the first energy curve has a performance problem.

[0046] In an embodiment of the present application, the noise of the multiple target devices is monitored based on the second detection method to obtain multiple second monitoring audio data, which includes: S401, obtaining a second detection time interval.

[0047] Specifically, the second detection time interval includes multiple time intervals with the same time interval.

[0048] S402, based on the second detection time interval, the noise of the same monitoring point on each target device is monitored to obtain multiple second monitoring audio data of each target device.

[0049] In this embodiment, the noise of the same monitoring point on each target device is monitored at multiple unit time points to obtain multiple second monitoring audio data of the same monitoring point of each target device in the time dimension, and one second monitoring audio data corresponds to the monitoring audio data of one time interval.

[0050] In an embodiment of the present application, the obtained multiple second monitoring audio data are all subjected to cluster analysis to obtain multiple second running curves, which includes: S501, selecting a second monitoring audio data.

[0051] S502, processing the second monitoring audio data to obtain audio restoration data of the second monitoring audio data.

[0052] S503, converting the audio restoration data into a second density curve and a second energy curve.

[0053] S504, returning to the selection of the second monitoring audio data until each of the second monitoring audio data is selected once.

[0054] In this embodiment, the second monitoring audio data corresponding to each target device is first processed to obtain audio restoration data corresponding to each target device, and then the obtained audio restoration data is converted into a second density curve and a second energy curve.

[0055] In an embodiment of the present application, the obtained plurality of second monitoring audio data is subjected to cluster analysis to obtain a plurality of second running curves, further comprising: S505, classifying a plurality of second density curves based on cluster analysis to obtain a plurality of second reference density curves.

[0056] S506, classifying a plurality of second energy curves based on cluster analysis to obtain a plurality of second reference energy curves.

[0057] The final judgment result of each target device is generated based on the obtained plurality of second running curves, comprising: S601, obtaining a first decay threshold and a second decay threshold.

[0058] S602, calculating the matching value of each second density curve with the second reference density curve, respectively.

[0059] S603, calculating the first matching value difference between all adjacent two second density curves based on time sequence.

[0060] S604, determining whether there is any first matching value difference greater than or equal to the first decay threshold.

[0061] S605, if there is any first matching value difference greater than or equal to the first decay threshold, the target device corresponding to the second density curve has performance problems.

[0062] S606, calculating the matching value of each second energy curve with the second reference density curve, respectively.

[0063] S607, calculating the second matching value difference between all adjacent two second energy curves based on time sequence.

[0064] S608, judging whether there is any one second matching value difference greater than or equal to the second attenuation threshold.

[0065] S609, if there is any one second matching value difference greater than or equal to the second attenuation threshold, the target device corresponding to the second energy curve has a performance problem.

[0066] Specifically, the same monitoring point of the same target device is monitored at different times, and then the obtained multiple audio data monitoring data are analyzed to analyze the performance attenuation speed of the target device. Comparing the obtained performance attenuation speed with the set standard attenuation speed can distinguish whether the device has a performance problem.

[0067] In this embodiment, taking one target device as an example, the multiple second density curves obtained for the target device are sorted in time sequence, then the matching value of each second density curve with the second reference density curve is calculated, then the difference between the matching values of adjacent two second density curves is calculated, and then multiple matching value differences are obtained. If any one of the obtained multiple matching value differences is greater than or equal to the first attenuation threshold, it is determined that the target device has a performance problem, for example, the performance attenuation is too fast.

[0068] The multiple second energy curves obtained for the target device are sorted in time sequence, then the matching value of each second energy curve with the second reference energy curve is calculated, then the difference between the matching values of adjacent two second energy curves is calculated, and then multiple matching value differences are obtained. If any one of the obtained multiple matching value differences is greater than or equal to the second attenuation threshold, it is determined that the target device has a performance problem, for example, the performance attenuation is too fast.

[0069] Specifically, by combining and analyzing the initial performance difference of the target device and the running performance attenuation difference of the target device, the state of the target device can be more accurately judged or inspected.

[0070] Embodiment 2: The multi-sensor data fusion mechanism is introduced, the dynamic adaptive threshold system is constructed, and the device degradation trend prediction function is added to further improve the quality inspection accuracy and engineering applicability. The specific implementation steps are as follows: Multi-sensor cooperative monitoring and data fusion can solve the problem that a single audio sensor is easily disturbed by environmental noise and cannot fully reflect the mechanical state of the device.

[0071] In the same monitoring point of the target device (such as the hoisting motor of the crane), three types of sensors are synchronously deployed: a high-precision microphone with a sampling rate of ≥48 kHz: to collect the original audio signal, which is the data source of the first detection method and the second detection method; a three-axis vibration sensor with a frequency range of 0.5 Hz-10 kHz, used to monitor the mechanical vibration energy of the device; and an infrared thermal imager with a resolution of 640x480, used to capture the temperature rise distribution of the device in real time.

[0072] It should be noted here that, because there are many collectors, data synchronization and fusion need to be controlled: The hardware timestamp alignment technology is adopted, with a time accuracy of ±1 ms, to ensure the spatio-temporal synchronization of the three types of data and generate a fusion data packet for each target device; the envelope spectrum energy entropy of the vibration signal and the Mel-frequency cepstral coefficient (MFCC) of the audio signal are jointly reduced (PCA algorithm) to generate a composite feature vector as the new input for clustering analysis; the limitations of audio analysis are compensated by vibration and temperature data to reduce the misjudgment rate caused by environmental interference.

[0073] Traditional K-means clustering is sensitive to noise data outliers and requires a pre-set class number K; the robust clustering algorithm is designed: DBSCAN (Density-Based Spatial Clustering) is used instead of K-means: automatically identifies outliers, such as sudden impact noise, and does not force them into any cluster; generates clusters based on density, suitable for non-spherical data distribution.

[0074] Adaptive clustering parameters: set the neighborhood radius ε to 1.5 times the median of the sample spacing, and the minimum point number MinPts=√N, where N is the number of samples.

[0075] Curve similarity measurement optimization: use dynamic time warping (DTW) algorithm to calculate the density / energy curve matching value, solve the alignment problem of different lengths at runtime; thereby improve the clustering robustness of non-stationary noise, and the accuracy rate of benchmark curve generation is improved by 18%.

[0076] The technical features of the above-described embodiments can be combined in any manner, and the execution order of the method steps is not limited. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described, but as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present disclosure.

[0077] The above-described embodiments are merely illustrative of several embodiments of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the scope of the patent. It should be noted that for those skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the present application should be subject to the appended claims.

Claims

1. A quality inspection method based on a noise analysis technique of a portal crane system, comprising a first detection method and a second detection method, the first detection method is used to obtain a first type of data, the second detection method is used to obtain a second type of data, the first detection method is different from the second detection method, characterized in that, The quality inspection method of the noise analysis technology based on the portal crane system comprises the following steps: Noise of the target devices is monitored based on the first detection method to obtain a plurality of first monitoring audio data; The obtained plurality of first monitoring audio data are subjected to cluster analysis to obtain a first operation curve; A preliminary judgment result of each target device is generated based on the obtained first operation curve; Noise of the target devices is monitored based on the second detection method to obtain a plurality of second monitoring audio data; The obtained plurality of second monitoring audio data are subjected to cluster analysis to obtain a plurality of second operation curves; A final judgment result of each target device is generated based on the obtained plurality of second operation curves.

2. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 1, characterized in that, The first detection time interval comprises a continuous and fixed time interval, and the noise of the target devices is monitored based on the first detection method to obtain a plurality of first monitoring audio data, which comprises the following steps: A first detection time interval is obtained; Noise of the same monitoring point of the target devices is monitored based on the first detection time interval to obtain the first monitoring audio data of each target device.

3. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 2, characterized in that, The obtained plurality of first monitoring audio data are subjected to cluster analysis to obtain a first operation curve, which comprises the following steps: One first monitoring audio data is selected; The first monitoring audio data is processed to obtain audio restoration data of the first monitoring audio data; The audio restoration data is converted into a first density curve and a first energy curve; The selection of one first monitoring audio data is returned until each first monitoring audio data is selected once.

4. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 3, characterized in that, The obtained plurality of first monitoring audio data are subjected to cluster analysis to obtain a first operation curve, which further comprises the following steps: The plurality of first density curves are classified based on the cluster analysis method to obtain a first reference density curve; The plurality of first energy curves are classified based on the cluster analysis method to obtain a first reference energy curve.

5. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 4, characterized in that, The preliminary judgment result of each target device is generated based on the obtained first operation curve, which comprises the following steps: First and second matching thresholds are obtained; Matching values of each first density curve with the first reference density curve are calculated, and all first density curves with matching values less than or equal to the first matching threshold are screened out, and the target devices corresponding to the screened out first density curves are defined as having performance problems; Matching values of each first energy curve with the first reference energy curve are calculated, and first energy curves with matching values less than or equal to the second matching threshold are screened out, and the target devices corresponding to the screened out first energy curves are defined as having performance problems.

6. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 5, characterized in that, The second detection time interval comprises a plurality of time intervals with the same time interval, and the noise of the target devices is monitored based on the second detection method to obtain a plurality of second monitoring audio data, which comprises the following steps: A second detection time interval is obtained; Noise of the same monitoring point on each target device is monitored based on the second detection time interval to obtain a plurality of second monitoring audio data of each target device.

7. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 6, characterized in that, The obtained plurality of second monitoring audio data are subjected to cluster analysis to obtain a plurality of second operation curves, which comprises the following steps: One second monitoring audio data is selected; processing the second monitoring audio data to obtain audio restoration data of the second monitoring audio data; transforming the audio restoration data into a second density curve and a second energy curve; returning to selecting one second monitoring audio data until each of the second monitoring audio data is selected once.

8. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 7, characterized in that, The clustering analysis on the obtained multiple second monitoring audio data to obtain multiple second running curves further comprises: classifying the multiple second density curves based on the clustering analysis method to obtain multiple second reference density curves; classifying the multiple second energy curves based on the clustering analysis method to obtain multiple second reference energy curves.

9. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 8, characterized in that, The final judgment result of each target device based on the obtained multiple second running curves comprises: obtaining a first attenuation threshold and a second attenuation threshold; calculating the matching value of each second density curve with the second reference density curve respectively; calculating the first matching value difference between all adjacent two second density curves based on the time sequence; judging whether there is any first matching value difference greater than or equal to the first attenuation threshold; if there is any first matching value difference greater than or equal to the first attenuation threshold, the target device corresponding to the second density curve has performance problems; calculating the matching value of each second energy curve with the second reference density curve respectively; calculating the second matching value difference between all adjacent two second energy curves based on the time sequence; judging whether there is any second matching value difference greater than or equal to the second attenuation threshold; if there is any second matching value difference greater than or equal to the second attenuation threshold, the target device corresponding to the second energy curve has performance problems.

10. The quality inspection method based on the noise analysis technique of the gantry crane system according to claim 9, characterized in that, The first detection time interval comprises a continuous and fixed time interval; The second detection time interval comprises multiple time intervals with the same time interval.

Images