Digital twinning health monitoring system and method for machine tool

Abstract

The invention discloses a machine tool digital twin health monitoring system and method, and relates to the technical field of machine tool monitoring. The machine tool digital twin health monitoring system comprises a sensor module, a data collection module, a model construction module, a fault injection module, a fault evaluation module and an output module. The invention further discloses a machine tool digital twinning health monitoring method, accurate fault diagnosis can be achieved under the condition of insufficient data by means of interpretability and visualization of the digital twinning model, analysis such as fault positioning, fault quantification and fault traceability is carried out, and the machine tool digital twinning health monitoring method is high in practicability. A numerical control machine tool digital twinborn model, real-time sensing monitoring data and a related machine learning algorithm are integrated, online and real-time fault diagnosis of the numerical control machine tool is realized, the starting rate and the working efficiency of the numerical control machine tool are improved, and the monitoring maintenance and repair cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of machine tool monitoring, and in particular to a machine tool digital twin health monitoring system and method. Background Art

[0002] In modern industrial production systems, especially in the Flexible Manufacturing System (FMS) environment where CNC machine tool technology is widely used, there are some key problems that are difficult to effectively overcome with traditional management and maintenance methods. Since CNC machine tools in FMS systems usually undertake multi-process, high-precision and high-intensity continuous operations, the complexity and harshness of their operating conditions have increased significantly, resulting in frequent common faults such as tool breakage, tool fracture and tool wear.

[0003] The invention patent with announcement number CN118192423A discloses a CNC machine tool health monitoring method based on digital twins, which includes: S1. Constructing a multi-sensor data acquisition system to collect the status data of the CNC machine tool in real time; S2. Constructing a digital twin system and establishing a data model; the digital twin system obtains the status data and matches and aligns it with the data model; S3. Deploying a CNN-LSTM model to infer and judge the current health status and milling status of the CNC machine tool, and output health prediction results; S4. Connecting to the FMS system, the FMS system obtains the current status data of the CNC machine tool and the health prediction results output by the CNN-LSTM model by calling the digital twin system to complete the health monitoring process of the CNC machine tool.

[0004] The existing CNC machine tool health monitoring method has improved the accuracy and reliability of CNC machine tool health monitoring and ensured data integrity. However, the above method ignores the problem that data characteristics change with the actual working conditions of the equipment, resulting in unevenness of extended fault samples, poor reliability, and reduced accuracy of fault diagnosis. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a machine tool digital twin health monitoring system and method to solve the existing problems.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a machine tool digital twin health monitoring system, comprising:

[0007] Sensor module, which builds a multi-sensor data acquisition system through multiple sensors to collect dynamic operation data of CNC machine tools;

[0008] The data collection module is used to collect the static data of the machine tool and normalize the dynamic data and static data uploaded by the sensor module;

[0009] Model building module, based on the digital twin system, builds the digital twin model of the corresponding CNC machine tool;

[0010] The fault injection module is used to inject the failure mode and cause of the CNC machine tool into the digital twin model of the CNC machine tool by combining the set working conditions, model physical parameters and geometric properties;

[0011] The fault assessment module is used to determine the real status of the CNC machine tool based on the status of the digital twin model of the CNC machine tool when an abnormality occurs in the CNC machine tool;

[0012] The output module is used to output the predicted value of the health status of the machine tool.

[0013] Preferably, the system further comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the computer program.

[0014] Preferably, the multiple sensors include: a cutting force sensor, a multi-channel vibration sensor, a single temperature sensor and a temperature and humidity sensor; the sensor data of each type of sensor are transmitted to the digital twin system in real time.

[0015] The machine tool digital twin health monitoring method includes the following steps:

[0016] S1: By collecting static data information of the CNC machine tool and dynamic data information during operation, the data information is normalized to obtain state data;

[0017] S2: Based on the digital twin system, a digital twin model of the corresponding CNC machine tool is established, and the collected CNC machine tool status data is transmitted to the digital twin model of the CNC machine tool so that the digital twin model of the CNC machine tool matches it;

[0018] S3: Inject the CNC machine tool fault into the CNC machine tool digital twin model. Each CNC machine tool digital twin model in the generated digital twin model library can be mapped to a certain working state of the CNC machine tool.

[0019] S4: Collect the operation data of multiple CNC machine tool digital twin models representing different states of CNC machine tools, form all the operation data into sample feature data sets, and standardize the feature data sets;

[0020] S5: Use the K-Means clustering analysis method to divide the standardized feature data set and determine the boundary of each degradation state; evaluate the health status classification effect of each feature in the feature data set and select the feature subset with the highest classification effect evaluation value as the sensitive feature;

[0021] S6: Input the real-time monitoring data of the CNC machine tool into the digital twin model of the CNC machine tool, and judge the real state of the CNC machine tool according to the state of the digital twin model of the CNC machine tool, so as to output the predicted value of the health state of the machine tool.

[0022] Preferably, the static data information in step S1 includes: geometric dimensions, system structure, physical properties, working capacity and model of the CNC machine tool;

[0023] The dynamic data information in step S1 includes: working condition information, vibration information, temperature information, stiffness information, noise information and loading force information.

[0024] Preferably, in step S2, by analyzing the system structure and coupling relationship of the CNC machine tool, the system is modularized and componentized into multiple subsystems such as mechanical, electrical control and heat transfer. At the same time, according to the mechanical structure and functional characteristics of the CNC machine tool, the Modelica multi-domain unified modeling language is used to compile and describe the operating mechanism of each functional component of each subsystem of the CNC machine tool to form a model of each functional component.

[0025] Preferably, the K-Means cluster analysis in step S5 divides the characteristic signals during the operation of the equipment based on an unsupervised learning method, clarifies the boundaries of each degradation state, and adds state labels to the characteristic signals, which is more conducive to the training and learning of the classifier.

[0026] Preferably, the K-Means cluster analysis method in step S5 includes:

[0027] (1) Randomly select K objects as the initial cluster centers, and then calculate the distance between each object and each cluster center. The distance d between each object and each cluster center is:

[0028]

[0029] Where x:(x1, x2, …, x n ), y:(y1,y2,…,y n ) represent the coordinates of different objects in space;

[0030] (2) Each object is assigned to the cluster center that is closest to it. The cluster center and the object assigned to it represent a cluster. Each time a sample is assigned, the cluster center of the cluster is recalculated based on the existing objects in the cluster. This is repeated until the criterion function reaches the optimal value. The criterion function is expressed as:

[0031]

[0032] Where k is the number of clusters, C iis the center point of the ith cluster, dist(C i ,X) is the distance from the data object to its cluster center;

[0033] (3) The classification effect of the feature value is evaluated by FisherScore, specifically:

[0034]

[0035] Among them, S represents the evaluation value of the feature classification effect, c represents the number of categories, and n k represents the number of samples in the kth class, and For the mean and mean square error of the kth class of the jth feature, μ j represents the mean of all data sets for the jth feature, Represents the distance from each class to the center of all data sets. The higher the value, the easier it is for the feature to distinguish between categories. Represents the degree of discreteness of each class of data. Usually, a smaller value will produce better classification results.

[0036] Beneficial Effects

[0037] The present invention provides a machine tool digital twin health monitoring system and method. Compared with the prior art, it has the following beneficial effects:

[0038] 1. The digital twin health monitoring system and method for machine tools utilizes the interpretability and visualization of the digital twin model to achieve accurate fault diagnosis when data is insufficient, and perform fault location, fault quantification, fault tracing and other analyses. It integrates the digital twin model of CNC machine tools, real-time sensor monitoring data and related machine learning algorithms to achieve online and real-time fault diagnosis of CNC machine tools, improve the availability and work efficiency of CNC machine tools, and reduce monitoring, maintenance and repair costs.

[0039] 2. The digital twin health monitoring system and method for machine tools, by integrating multi-sensor fusion technology, can capture various key parameters in the operation process of CNC machine tools in real time, fully and comprehensively, overcoming the problem of insufficient accuracy of single sensor data, and significantly improving the accuracy and reliability of health status monitoring of CNC machine tools. By introducing the K-Means clustering analysis method for data division, it is more conducive to establishing the correlation between degradation characteristics and health status, further improving the recognition and prediction accuracy of machine tool health status, reducing the occurrence rate of production accidents, improving production and processing efficiency, and ensuring production and processing accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic diagram of system connection of the present invention;

[0041] Figure 2It is a schematic diagram of the method flow of the present invention;

[0042] Figure 3 It is a schematic diagram of the K-Means clustering analysis method flow of the present invention. DETAILED DESCRIPTION

[0043] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] See also Figure 1-3 , the present invention provides three technical solutions:

[0045] First embodiment: The present invention discloses a machine tool digital twin health monitoring system, comprising:

[0046] Sensor module, which builds a multi-sensor data acquisition system through multiple sensors to collect dynamic operation data of CNC machine tools. The multiple sensors include: cutting force sensor, multi-channel vibration sensor, single temperature sensor and temperature and humidity sensor; the sensor data of various sensors are transmitted to the digital twin system in real time;

[0047] The data collection module is used to collect the static data of the machine tool and normalize the dynamic data and static data uploaded by the sensor module;

[0048] Model building module, based on the digital twin system, builds the digital twin model of the corresponding CNC machine tool;

[0049] The fault injection module is used to inject the failure mode and cause of the CNC machine tool into the digital twin model of the CNC machine tool by combining the set working conditions, model physical parameters and geometric properties;

[0050] The fault assessment module is used to determine the real status of the CNC machine tool based on the status of the digital twin model of the CNC machine tool when an abnormality occurs in the CNC machine tool;

[0051] The output module is used to output the predicted value of the health status of the machine tool.

[0052] The system further comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the computer program.

[0053] Second implementation mode: The present invention also discloses a machine tool digital twin health monitoring method, comprising the following steps:

[0054] S1: By collecting static data information of the CNC machine tool and dynamic data information during operation, the data information is normalized to obtain state data;

[0055] S2: Based on the digital twin system, a digital twin model of the corresponding CNC machine tool is established, and the collected CNC machine tool status data is transmitted to the digital twin model of the CNC machine tool to match it. By analyzing the system structure and coupling relationship of the CNC machine tool, it is modularized and componentized into multiple subsystems such as mechanical, electrical control and heat transfer. At the same time, according to the mechanical structure and functional characteristics of the CNC machine tool, the Modelica multi-domain unified modeling language is used to compile and describe the operating mechanism of each functional component of each subsystem of the CNC machine tool to form a model of each functional component;

[0056] S3: Inject the CNC machine tool fault into the CNC machine tool digital twin model. Each CNC machine tool digital twin model in the generated digital twin model library can be mapped to a certain working state of the CNC machine tool.

[0057] S4: Collect the operation data of multiple CNC machine tool digital twin models representing different states of CNC machine tools, form all the operation data into sample feature data sets, and standardize the feature data sets;

[0058] S5: Use the K-Means clustering analysis method to divide the standardized feature data set and determine the boundary of each degradation state; evaluate the health status classification effect of each feature in the feature data set and select the feature subset with the highest classification effect evaluation value as the sensitive feature;

[0059] S6: Input the real-time monitoring data of the CNC machine tool into the digital twin model of the CNC machine tool, and judge the real state of the CNC machine tool according to the state of the digital twin model of the CNC machine tool, so as to output the predicted value of the health state of the machine tool.

[0060] The machine tool digital twin health monitoring method is applied to one or more terminals, where the terminal is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASIC), programmable gate arrays (FPGA), digital processors (DSP), embedded devices, etc.

[0061] The terminal can be any electronic product that can perform human-computer interaction with a user, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), an interactive network television (IPTV), etc.

[0062] The terminal may also include network equipment and / or user equipment, wherein the network equipment includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud consisting of a large number of hosts or network servers based on cloud computing.

[0063] The network where the terminal is located includes but is not limited to the Internet, wide area network, metropolitan area network, local area network, virtual private network (VPN), etc.

[0064] In this embodiment, the CNC machine tool entity-related data is the factual basis for constructing the CNC machine tool digital twin model.

[0065] Static data information such as the geometric dimensions, system structure, physical properties, working capacity, model, etc. of CNC machine tools can be directly obtained through tools such as machine tool design parameters and design drawings.

[0066] As for the dynamic data information of the working conditions, vibration, temperature, stiffness, noise, loading force, etc. of the CNC machine tool entity, by analyzing the CNC machine tool structure and working environment, selecting appropriate data collection points, formulating a reasonable sensing monitoring plan, installing vibration, temperature, speed, displacement and other sensors, and collecting real-time data information during the operation of the CNC machine tool, providing complete physical data for the construction of the digital twin model of the CNC machine tool.

[0067] The third implementation method: K-Means cluster analysis divides the characteristic signals during equipment operation based on unsupervised learning methods, clarifies the boundaries of each degradation state, and adds state labels to the characteristic signals, which is more conducive to the training and learning of the classifier.

[0068] K-Means clustering analysis methods include:

[0069] (1) Randomly select K objects as the initial cluster centers, and then calculate the distance between each object and each cluster center. The distance d between each object and each cluster center is:

[0070]

[0071] Where x:(x1, x2, …, x n ), y:(y1,y2,…,yn ) represent the coordinates of different objects in space;

[0072] (2) Each object is assigned to the cluster center that is closest to it. The cluster center and the object assigned to it represent a cluster. Each time a sample is assigned, the cluster center of the cluster is recalculated based on the existing objects in the cluster. This is repeated until the criterion function reaches the optimal value. The criterion function is expressed as:

[0073]

[0074] Where k is the number of clusters, C i is the center point of the ith cluster, dist(C i ,X) is the distance from the data object to its cluster center;

[0075] (3) The classification effect of the feature value is evaluated by FisherScore, specifically:

[0076]

[0077] Among them, S represents the evaluation value of the feature classification effect, c represents the number of categories, and n k represents the number of samples in the kth class, and For the mean and mean square error of the kth class of the jth feature, μ j represents the mean of all data sets for the jth feature, Represents the distance from each class to the center of all data sets. The higher the value, the easier it is for the feature to distinguish between categories. Represents the degree of discreteness of each class of data. Usually, a smaller value will produce better classification results.

[0078] To sum up, the digital twin health monitoring system and method for machine tools of the present invention utilize the interpretability and visualization of the digital twin model to achieve accurate fault diagnosis when data is insufficient, and perform fault location, fault quantification, fault tracing and other analyses. It integrates the digital twin model of CNC machine tools, real-time sensor monitoring data and related machine learning algorithms to achieve online and real-time fault diagnosis of CNC machine tools, improve the availability and work efficiency of CNC machine tools, and reduce monitoring, maintenance and repair costs.

[0079] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited, and conventional equipment can be used.

[0080] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0081] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. The machine tool digital twin health monitoring system is characterized by: include: Sensor module, which builds a multi-sensor data acquisition system through multiple sensors to collect dynamic operation data of CNC machine tools; The data collection module is used to collect the static data of the machine tool and normalize the dynamic data and static data uploaded by the sensor module; Model building module, based on the digital twin system, builds the digital twin model of the corresponding CNC machine tool; The fault injection module is used to inject the failure mode and cause of the CNC machine tool into the digital twin model of the CNC machine tool by combining the set working conditions, model physical parameters and geometric properties; The fault assessment module is used to determine the real status of the CNC machine tool based on the status of the digital twin model of the CNC machine tool when an abnormality occurs in the CNC machine tool; The output module is used to output the predicted value of the health status of the machine tool.

2. The machine tool digital twin health monitoring system according to claim 1, characterized in that: The system further comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor executes the computer program.

3. The machine tool digital twin health monitoring system according to claim 1, characterized in that: The multiple sensors include: cutting force sensors, multi-channel vibration sensors, single temperature sensors and temperature and humidity sensors; the sensor data of various sensors are transmitted to the digital twin system in real time.

4. A machine tool digital twin health monitoring method, characterized in that: The following steps are involved: S1: By collecting static data information of the CNC machine tool and dynamic data information during operation, the data information is normalized to obtain state data; S2: Based on the digital twin system, a digital twin model of the corresponding CNC machine tool is established, and the collected CNC machine tool status data is transmitted to the digital twin model of the CNC machine tool so that the digital twin model of the CNC machine tool matches it; S3: Inject the CNC machine tool fault into the CNC machine tool digital twin model. Each CNC machine tool digital twin model in the generated digital twin model library can be mapped to a certain working state of the CNC machine tool. S4: Collect the operation data of multiple CNC machine tool digital twin models representing different states of CNC machine tools, form all the operation data into sample feature data sets, and standardize the feature data sets; S5: Use the K-Means clustering analysis method to divide the standardized feature data set and determine the boundary of each degradation state; evaluate the health status classification effect of each feature in the feature data set and select the feature subset with the highest classification effect evaluation value as the sensitive feature; S6: Input the real-time monitoring data of the CNC machine tool into the digital twin model of the CNC machine tool, and judge the real state of the CNC machine tool according to the state of the digital twin model of the CNC machine tool, so as to output the predicted value of the health state of the machine tool.

5. The machine tool digital twin health monitoring method according to claim 4, characterized in that: The static data information in step S1 includes: geometric dimensions, system structure, physical properties, working capacity and model of the CNC machine tool; The dynamic data information in step S1 includes: working condition information, vibration information, temperature information, stiffness information, noise information and loading force information.

6. The machine tool digital twin health monitoring method according to claim 4, characterized in that: In the step S2, by analyzing the system structure and coupling relationship of the CNC machine tool, it is modularized and componentized into multiple subsystems such as mechanical, electrical control and heat transfer. At the same time, according to the mechanical structure and functional characteristics of the CNC machine tool, the Modelica multi-domain unified modeling language is used to compile and describe the operating mechanism of each functional component of each subsystem of the CNC machine tool to form a model of each functional component.

7. The machine tool digital twin health monitoring method according to claim 4, characterized in that: The K-Means cluster analysis in step S5 divides the characteristic signals during the operation of the equipment based on an unsupervised learning method, clarifies the boundaries of each degradation state, and adds state labels to the characteristic signals, which is more conducive to the training and learning of the classifier.

8. The machine tool digital twin health monitoring method according to claim 4, characterized in that: The K-Means cluster analysis method in step S5 includes: (1) Randomly select K objects as the initial cluster centers, and then calculate the distance between each object and each cluster center. The distance d between each object and each cluster center is: Where x:(x1, x2, …, x n ), y:(y1,y2,…,y n ) represent the coordinates of different objects in space; (2) Each object is assigned to the cluster center that is closest to it. The cluster center and the object assigned to it represent a cluster. Each time a sample is assigned, the cluster center of the cluster is recalculated based on the existing objects in the cluster. This is repeated until the criterion function reaches the optimal value. The criterion function is expressed as: Where k is the number of clusters, C i is the center point of the ith cluster, dist(C i ,X) is the distance from the data object to its cluster center; (3) The classification effect of the feature value is evaluated by FisherScore, specifically: Among them, S represents the evaluation value of the feature classification effect, c represents the number of categories, and n k represents the number of samples in the kth class, and For the mean and mean square error of the kth class of the jth feature, μ j represents the mean of all data sets for the jth feature, Represents the distance from each class to the center of all data sets. The higher the value, the easier it is for the feature to distinguish between categories. Represents the degree of discreteness of each class of data. Usually, a smaller value will produce better classification results.

Images