Method and system for monitoring running state of photogravure press

Abstract

The invention relates to the field of photogravure press monitoring, in particular to a photogravure press running state monitoring method and system. The method comprises the following steps: acquiring multi-dimensional monitoring data of the photogravure press at each sampling moment; calculating the data concentration degree of the target dimension; taking other dimensions except the target dimension as contrast dimensions, calculating irrelevance between the target dimension and any contrast dimension, traversing to obtain irrelevance between the target dimension and each contrast dimension, and taking a mean value of all irrelevance as independence of the target dimension; and calculating importance, taking the importance as a correction factor to correct the abnormal degree obtained by the HBOS algorithm, obtaining the abnormal degree of the target dimension, integrating the abnormal degrees of all dimensions to obtain the final abnormal degree, and completing state monitoring. According to the technical scheme, the precision and efficiency of the monitoring result can be improved, and guarantee is provided for stable operation of the photogravure press.

Description

Technical Field

[0001] The present invention relates to the field of gravure printing machine monitoring, and more particularly to a method and system for monitoring the operating status of a gravure printing machine. Background Art

[0002] Gravure printing presses are widely used in the printing industry, playing a particularly important role in producing various printed products, such as packaging materials, newspapers, and books. They engrave grooves on the surface of a printing plate and transfer images and text using ink to paper or other media. Due to the complex and high-precision workflow of gravure printing presses, print quality and machine operational stability directly impact production efficiency and the quality of the final product.

[0003] Currently, while some high-end printing presses are equipped with certain sensors and monitoring systems, such as temperature, pressure, and speed sensors, to monitor specific mechanical or electrical parameters, these monitoring systems typically only provide early warnings for single issues and lack a comprehensive assessment of the equipment's overall operating status. Furthermore, these traditional monitoring systems often fail to promptly identify potential equipment anomalies, resulting in failures being discovered only after they occur, impacting production efficiency and product quality.

[0004] HBOS (Histogram-based Outlier Detection) is a histogram-based anomaly detection algorithm used in unsupervised anomaly detection tasks. Its core concept is to identify outliers that differ significantly from the majority of data points by analyzing the density of the data distribution. However, gravure printing presses typically require monitoring of data in multiple dimensions during operation. When processing multidimensional data, the HBOS anomaly data detection algorithm fails to effectively consider the differences in the neutrality of data sets in different dimensions and the independence between dimensions, resulting in inaccurate monitoring results for gravure printing presses. Summary of the Invention

[0005] In order to solve the problem of inaccurate monitoring results of the operation of the gravure printing machine, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention discloses a method for monitoring the operating status of a gravure printing machine, comprising: obtaining multi-dimensional monitoring data of the gravure printing machine at each sampling moment; taking any dimension as a target dimension and calculating the data concentration of the target dimension; taking dimensions other than the target dimension as reference dimensions, calculating the irrelevance between the target dimension and any reference dimension, traversing to obtain the irrelevance between the target dimension and each reference dimension, and taking the mean of all irrelevances as the independence of the target dimension; normalizing the product of the data concentration and the independence as the importance of the target dimension, using the importance as a correction factor to correct the degree of abnormality obtained by the HBOS algorithm, obtaining the abnormality of the target dimension, integrating the abnormalities of all dimensions to obtain the final abnormality, and completing status monitoring.

[0007] Preferably, the data concentration satisfies the relationship: , Represents the target dimension The data concentration, Represents the target dimension The mean of the data, Represents the target dimension The standard deviation of the data, Represents the target dimension The number of data in Indicates quantity.

[0008] Preferably, the data concentration includes: Calculate the number of data points within the preset distance range for any data in the target dimension and normalize them, and take the mean of all normalized data points as the data concentration of the target dimension.

[0009] Preferably, the irrelevance satisfies the relationship: , Represents the target dimension and comparison dimension The irrelevance of Represents the target dimension The data is Time comparison dimension The data series, Represents the target dimension The data is Time comparison dimension The data series, Indicates similarity.

[0010] Preferably, the similarity is Pearson correlation coefficient or KL divergence.

[0011] Preferably, the irrelevance includes: obtaining the irrelevance between the target dimension and any dimension through hypothesis testing.

[0012] Preferably, the final abnormality degree satisfies the relationship: , represents the final abnormality, Represents the target dimension The importance of Represents the target dimension obtained by the HBOS algorithm degree of abnormality.

[0013] Preferably, the multi-dimensional monitoring data includes temperature, humidity, vibration, pressure, displacement and current.

[0014] Preferably, the completion status monitoring includes: generating and sending an alarm signal in response to the final abnormality degree being greater than a preset abnormality threshold.

[0015] In a second aspect, the present invention discloses a gravure printing machine operating status monitoring system, comprising: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, any one of the gravure printing machine operating status monitoring methods described above is implemented.

[0016] Beneficial effects of the present invention: This invention calculates the data concentration and independence of each dimension and adjusts the anomaly level based on an importance correction factor, ensuring more accurate and meaningful monitoring results. By comprehensively considering the characteristics of each dimension, it avoids the potential bias introduced by a single dimension, thereby providing more reliable condition monitoring. Ultimately, with integrated anomaly analysis, the system automatically issues an alarm when anomalies exceed a preset threshold, rapidly responding and implementing measures to avoid potential damage to the equipment or production interruption, effectively improving equipment reliability and production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The present invention is a flowchart of a method for monitoring the operating status of a gravure printing machine according to an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, but not all of the embodiments.

[0019] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Reference Figure 1A method for monitoring the operating status of a gravure printing machine includes steps S1 to S4, which are described in detail below.

[0021] S1: Obtain multi-dimensional monitoring data of the gravure printing machine at each sampling moment.

[0022] In one embodiment, a large number of tiny MEMS sensor nodes are deployed in key components and prone to failure within a gravure printing press, such as the plate cylinder, impression cylinder, transmission gears, and motors. These sensor nodes integrate multiple sensors, including temperature, humidity, vibration, pressure, displacement, and current, enabling real-time, comprehensive monitoring of the status of each key component of the printing press. This approach collects multi-dimensional monitoring data at each sampling moment, accurately reflecting the physical changes occurring during equipment operation. This data provides a solid foundation for fault prediction, performance optimization, and early fault diagnosis, helping to identify potential problems in the long-term operation of the equipment. By combining and analyzing multi-dimensional data, the accuracy of fault monitoring can be improved, ensuring efficient and stable operation of the gravure printing press, extending its service life, and reducing downtime, thereby improving production efficiency and lowering operating costs.

[0023] S2: Take any dimension as the target dimension and calculate the data concentration of the target dimension.

[0024] In one embodiment, the data concentration satisfies the relationship: , Represents the target dimension The data concentration, Represents the target dimension The mean of the data, Represents the target dimension The standard deviation of the data, Represents the target dimension The number of data in Indicates quantity.

[0025] By limiting the data distribution to within three standard deviations of the mean, we eliminate the influence of outliers and ensure the stability and representativeness of the dataset. This concentration measurement helps identify regularities and abnormal fluctuations in the dataset, providing a more accurate basis for subsequent data analysis.

[0026] In one embodiment, the data concentration includes: calculating the number of data points within a preset distance range for any data in the target dimension and normalizing them, and taking the average of all normalized data points as the data concentration of the target dimension.

[0027] The more concentrated the data is along a given dimension, the more accurate the HBOS algorithm's detection results are, as the algorithm relies on the distribution characteristics of each dimension to assess the anomaly of a data point. When data is highly concentrated along a particular dimension, the data distribution exhibits less fluctuation, making it easier to distinguish outliers because they significantly deviate from the main data distribution pattern. In this case, the HBOS algorithm can more effectively model the data by constructing a histogram, accurately assessing the degree of anomaly for each data point. Therefore, when data concentration is high, the difference between outliers and normal data is more pronounced, enabling the algorithm to more accurately identify anomalies and avoid false positives and missed detections. However, when data distribution is more dispersed or there is significant noise, the HBOS algorithm may be affected, resulting in less sensitive or accurate detection of anomalies. Therefore, the degree of data concentration directly affects the algorithm's detection performance. In cases of high concentration, the HBOS algorithm can more clearly identify anomalies, thereby improving the accuracy of anomaly detection.

[0028] S3: Take the dimensions other than the target dimension as control dimensions, calculate the irrelevance between the target dimension and any control dimension, traverse to obtain the irrelevance between the target dimension and each control dimension, and take the mean of all irrelevance as the independence of the target dimension.

[0029] It should be noted that if the data in two dimensions are independent, it means that the data values in one dimension are not affected by the data values in the other dimension. In this case, the changes in each dimension can independently reflect the characteristics of the dimension itself, without being correlated or offset by changes in other dimensions. For example, in multidimensional data analysis, if two dimensions are independent, the data fluctuations in one dimension will not have a direct impact on the fluctuations in the other dimension. Therefore, the data distribution of each dimension can be analyzed separately, and the detection of outliers can only focus on the characteristics of each dimension without considering the interdependence between dimensions. This independence assumption usually simplifies the construction of the model, allowing us to treat each dimension independently and avoid the complex joint distribution modeling problem.

[0030] In one embodiment, the irrelevance satisfies the relationship: , Represents the target dimension and comparison dimension The irrelevance of Represents the target dimension The data is Time comparison dimension The data series, Represents the target dimension The data is Time comparison dimension The data series, Indicates similarity.

[0031] Here, the similarity is the Pearson correlation coefficient or the KL divergence.

[0032] For example, randomly select all data with a data value of s and the corresponding sampling time in the target dimension, and obtain the data sequence of the control dimension at the same sampling time in any reference dimension. Then, use this method to obtain all data with a data value of t and the corresponding sampling time in the target dimension, and obtain the data sequence of the control dimension at the same sampling time in any reference dimension. Calculate the similarity of the two data sequences. The greater the similarity, the greater the irrelevance, indicating that no matter what value is taken in the target dimension, it does not affect the data in the reference dimension.

[0033] Traverse to obtain the irrelevance between the target dimension and each control dimension, and take the mean of all irrelevance as the independence of the target dimension.

[0034] In one embodiment, the irrelevance includes: obtaining the irrelevance between the target dimension and any dimension through hypothesis testing. Hypothesis testing is a well-known technique for those skilled in the art and will not be described in detail here.

[0035] S4: Normalize the product of data concentration and independence as the importance of the target dimension, use the importance as a correction factor to correct the degree of abnormality obtained by the HBOS algorithm, obtain the abnormality of the target dimension, integrate the abnormality of all dimensions to obtain the final abnormality, and complete the status monitoring.

[0036] In one embodiment, the importance satisfies the relationship: , Represents the target dimension The importance of Represents the target dimension The data concentration, Represents the target dimension independence.

[0037] According to the HBOS algorithm, the degree of abnormality corresponding to the monitoring data of each dimension can be obtained. The importance is used as a correction factor to correct the degree of abnormality to obtain the abnormality of the target dimension. The abnormality of all dimensions is integrated to obtain the final abnormality. The final abnormality satisfies the relationship: , represents the final abnormality, Represents the target dimension The importance of Represents the target dimension obtained by the HBOS algorithm degree of abnormality.

[0038] When the final abnormality of the monitored data for a certain dimension exceeds the preset abnormality threshold, it indicates that the data in that dimension may deviate from the normal pattern, indicating a potential fault or abnormality. To enable timely response, an alarm signal is automatically generated and transmitted based on this abnormality exceeding the threshold. This alarm signal not only identifies the abnormality but can also trigger further actions, such as activating backup systems, notifying maintenance personnel, or conducting troubleshooting. The generation and transmission of alarm signals not only facilitates timely response but also provides data support for subsequent analysis of the cause of the abnormality.

[0039] The system includes a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a method for monitoring the operating status of an intaglio printing machine according to the first aspect of the present invention is implemented.

[0040] The system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and therefore will not be described in detail here.

[0041] It should be noted that those skilled in the art may make various modifications and improvements without departing from the scope of the present invention, and these modifications and improvements fall within the scope of protection of the present invention. Therefore, the scope of protection of the patent for this invention shall be based on the appended claims.

Claims

1. A method for monitoring the operating status of a gravure printing machine, characterized in that: include: Obtain multi-dimensional monitoring data of the gravure printing machine at each sampling moment; Take any dimension as the target dimension and calculate the data concentration of the target dimension; Take the other dimensions except the target dimension as the control dimension, calculate the irrelevance between the target dimension and any control dimension, traverse to obtain the irrelevance between the target dimension and each control dimension, and take the mean of all irrelevance as the independence of the target dimension; The product of data concentration and independence is normalized as the importance of the target dimension. The importance is used as a correction factor to correct the degree of abnormality obtained by the HBOS algorithm to obtain the abnormality of the target dimension. The abnormality of all dimensions is integrated to obtain the final abnormality to complete the status monitoring.

2. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The data concentration satisfies the relationship: , Represents the target dimension The data concentration, Represents the target dimension The mean of the data, Represents the target dimension The standard deviation of the data, Represents the target dimension The number of data in Indicates quantity.

3. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The data concentration includes: Calculate the number of data points within the preset distance range for any data in the target dimension and normalize them, and take the mean of all normalized data points as the data concentration of the target dimension.

4. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The irrelevance satisfies the relationship: , Represents the target dimension and comparison dimension The irrelevance of Represents the target dimension The data is Time comparison dimension The data series, Represents the target dimension The data is Time comparison dimension The data series, Indicates similarity.

5. The method for monitoring the operating status of a gravure printing machine according to claim 4, wherein: The similarity is the Pearson correlation coefficient or the KL divergence.

6. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The irrelevance includes: The non-correlation between the target dimension and any dimension is obtained through hypothesis testing.

7. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The final abnormality degree satisfies the relationship: , represents the final abnormality, Represents the target dimension The importance of Represents the target dimension obtained by the HBOS algorithm degree of abnormality.

8. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The multi-dimensional monitoring data includes temperature, humidity, vibration, pressure, displacement and current.

9. The method for monitoring the operating status of a gravure printing machine according to claim 1, wherein: The completion status monitoring includes: In response to the final abnormality degree being greater than a preset abnormality threshold, an alarm signal is generated and sent.

10. A gravure printing machine operating status monitoring system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for monitoring the operating status of a gravure printing machine according to any one of claims 1 to 9 is implemented.

Images