Processing data acquisition method and system applied to numerical control machine tool

Abstract

The invention relates to the technical field of numerical control machine tool intelligent manufacturing and industrial data collection, in particular to a machining data collection method and system applied to a numerical control machine tool, and the method comprises the following steps: S1, the classification and access scheme design of the numerical control machine tool: dividing the numerical control machine tool into a class I intelligent new machine tool, a class II semi-intelligent machine tool and a class III old dumb equipment according to the digitization capability, a classification access scheme of combination of a general protocol adapter, a serial port-Ethernet converter and a non-intrusive sensor is respectively adopted to realize data access of heterogeneous and old equipment; uniform access of 10 + brand numerical control machine tools is supported, the data intercommunication rate of heterogeneous equipment reaches 100%, the problem of'information isolated island 'is solved, six types of equipment states are recognized, and the OEE calculation precision is improved to + / -2% and is improved by 5 times compared with a traditional method.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent manufacturing and industrial data acquisition of numerical control machine tools, and particularly relates to a machining data acquisition method and system applied to numerical control machine tools. BACKGROUND

[0002] The following key technical problems exist in the current machining data acquisition field of numerical control machine tools, which seriously restrict the upgrading of intelligent manufacturing:

[0003] Heterogeneous device "data island": the numerical control machine tools in the workshop are mostly from different manufacturers such as Fanuc, Siemens, Mitsubishi and Heid, and adopt different communication protocols such as FOCAS, S7, MC and HMI-Advanced, so that there is a lack of unified access scheme, data cannot be interchanged, an "information island" is formed, and global monitoring is difficult to realize;

[0004] State sensing "binary" limitation: the traditional acquisition can only obtain the binary state of the device "on / off", and cannot distinguish the fine-grained states such as "cutting operation, empty running, program debugging, waiting for material and alarm", so that the OEE (overall equipment efficiency) calculation deviation is large, and the production control precision is low;

[0005] Old machine tool acquisition problem: a large number of old machine tools (such as those without Ethernet interface and only supporting RS232 / 485 low-speed communication) that have served for more than 10 years lack digital interfaces, and the cost of adding new hardware is high (accounting for 30% to 50% of the original value of the machine tool), and the stability of the original control system is easily damaged;

[0006] Massive data processing bottleneck: the data acquisition frequency of the spindle load, feed rate and other data of the numerical control machine tool is up to 10-100Hz, and the single machine tool generates 10GB+ raw data per day, direct uploading easily leads to network bandwidth occupancy rate exceeding 80%, server storage pressure increasing dramatically, and the raw data contains noise (such as sensor fluctuation), and the data quality is low;

[0007] Production safety risk: when directly accessing the numerical control system (CNC) to acquire data, viruses and illegal write instructions are easily introduced, which may cause CNC program disorder and equipment misoperation, and seriously affect production safety;

[0008] Insufficient data value conversion: the acquired data is mostly unstructured format, which cannot be directly connected to MES (manufacturing execution system), digital twin and other upper-layer applications, and it is difficult to realize value scenarios such as machining traceability and predictive maintenance, therefore, the machining data acquisition method and system applied to numerical control machine tools are provided. SUMMARY

[0009] Therefore, the machining data acquisition method and system applied to numerical control machine tools are provided to solve or alleviate the technical problems in the prior art, and at least provide a beneficial choice.

[0010] The technical scheme of the present application is implemented as follows: a machining data acquisition method applied to a numerical control machine tool, comprising the following steps:

[0011] S1. Numerical control machine tool classification and access scheme design: according to the digitization capability, the numerical control machine tool is divided into type I intelligent new machine tool, type II semi-intelligent machine tool and type III old and obsolete “dumb equipment”, and a classification access scheme of a general protocol adapter, a serial port-Ethernet converter and a non-invasive sensor combination is adopted to realize data access of heterogeneous and old and obsolete equipment;

[0012] S2. Multi-dimensional machining data real-time acquisition: for all types of machine tools, four-dimensional data of equipment state, machining parameter, process data and environment data are uniformly acquired, time synchronization of all acquisition equipment is realized through NTP protocol, and the synchronization error is ≤50 ms, so as to ensure that the data carries an accurate time stamp;

[0013] S3. Edge data preprocessing and safety isolation: an industrial-grade edge gateway is deployed to perform “Kalman filtering + moving average” combined noise reduction, LZ77 algorithm compression, local caching and off-network transmission processing, and at the same time, through an industrial firewall and read-only permission configuration, safety isolation of the acquisition process and the numerical control system is realized to ensure zero production interference;

[0014] S4. Data fusion and standardization: the preprocessed standardized data is uploaded to a cloud data platform, device state, machining parameter and process data are associated and fused through “device ID + time stamp” double primary key, data is output according to OPCUA or MQTTSparkplugB industrial standard, and three types of data interfaces of real-time, historical and statistical data are provided;

[0015] S5. Data value application: based on the fused data, device fine-grained state is recognized through a decision tree algorithm and OEE is calculated, precise traceability and digital twin replication of the machining process are realized in combination with the time stamp, and predictive maintenance and process optimization are carried out by using an LSTM model and an optimization algorithm.

[0016] Further preferably, in the S1, the type I intelligent new machine tool supports Ethernet communication and a standardized protocol, a protocol analysis library of FOCAS2, S7, MCProtocol and HMI-Advanced is built in the general protocol adapter, and the numerical control system is directly connected through Ethernet; the type II semi-intelligent machine tool has an RS232 / 485 serial port, a serial port-Ethernet converter is matched with a frame analysis algorithm based on CRC check to extract spindle speed and feed amount data; the non-invasive sensor combination of the type III old and obsolete “dumb equipment” includes a Hall current sensor, a three-axis vibration sensor, a photoelectric pulse collector and a voltage detection module, which respectively acquire spindle load, vibration, feed displacement and equipment state signals.

[0017] Further preferably, in the S2, the equipment state data includes equipment startup / shutdown, running / idle / alarm / debug / standby / maintenance state, spindle start / stop, tool replacement state, with a collection frequency of 1 Hz; the machining parameter data includes spindle speed, feed rate, cutting depth, NC program number, tool number, with a collection frequency of 10 Hz; the process data includes spindle load, spindle vibration value, feed shaft displacement, machining duration, product pass rate, with a collection frequency of 50 Hz; the environment data includes machine tool working environment temperature, humidity, power grid voltage, with a collection frequency of 0.1 Hz.

[0018] Further preferably, in the S3, the noise removal rate of the edge gateway is ≥ 92%, the data compression rate is 1:5, and the local cache capacity is ≥ 7 days of data storage; the security isolation adopts a “hardware isolation + permission control” dual strategy, the hardware isolation opens only a data reading port and blocks the writing port through an industrial firewall, the permission control disables the writing instruction through the numerical control system background and configures the read-only permission of the collection equipment, and the edge gateway is installed with an industrial antivirus software to prohibit external mobile storage devices.

[0019] Further preferably, in the S4, the data association and fusion includes state-parameter association, program-tool association, and process-quality association, which respectively associates the equipment cutting state with the machining parameter, the NC program number with the tool number, and the spindle vibration value with the product pass rate to form a structured data archive; the real-time data interface is a WebSocket interface that pushes second-level real-time data, the historical data interface is a RESTful API that supports querying according to the equipment ID and time range, and the statistical data interface is a SQL query interface that outputs summary data such as daily OEE and tool service life.

[0020] Further preferably, in the S5, the fine-grained state recognition includes running cutting state, idle state, alarm state, debugging state, standby state, and maintenance state, wherein the running cutting state determination condition is spindle load > 2A + feed rate > 0 + NC program running, and the idle state determination condition is spindle load < 1A + feed rate = 0 + no alarm; the OEE calculation accuracy is ± 2%, and the process optimization finds the optimal combination of spindle speed and feed rate through a genetic algorithm to achieve machining efficiency improvement ≥ 15% and product pass rate ≥ 99.9%.

[0021] Further preferably, in the S1, in the non-invasive sensor combination of the III type old “dumb equipment”, the Hall current sensor has a range of 0-5A and an accuracy of ± 1%, the three-axis vibration sensor has a range of ± 16g and a sampling rate of 100Hz, the photoelectric pulse collector has a resolution of 1000p / r, and the voltage detection module monitors the 24V / 0V voltage signal of the operating panel indicator light; the old machine tool transformation cost of the classified access scheme is ≤ 10% of the original value of the machine tool.

[0022] The machining data acquisition system applied to the numerical control machine tool comprises:

[0023] The device access layer comprises a general protocol adaptation module, a serial-to-ethernet conversion module and a non-intrusive sensor acquisition module, which respectively realize data access of the type I, type II and type III numerical control machine tools, the general protocol adaptation module supports 10+ manufacturer communication protocols, and the non-intrusive sensor acquisition module comprises a Hall current sensor, a three-axis vibration sensor, a photoelectric pulse collector and a voltage detection module.

[0024] The edge processing layer comprises a data preprocessing module, a security isolation module, a local cache and continuation module and a time synchronization module, is arranged in an industrial-grade edge gateway, the data preprocessing module realizes data noise reduction and compression, the security isolation module realizes industrial firewall isolation and read-only permission control, the local cache and continuation module supports offline data storage and subsequent transmission after recovery, and the time synchronization module realizes time synchronization of the acquisition device through the NTP protocol.

[0025] The data fusion layer comprises a data correlation fusion module and a data standardization module, is arranged in a cloud data platform, the data correlation fusion module correlates multi-source data based on the Spark framework according to the "device ID+time stamp", and the data standardization module outputs data according to the OPCUA or MQTTSparkplugB standard.

[0026] The application support layer comprises a state recognition and OEE calculation module, a traceability and visualization module, a predictive maintenance and process optimization module, the state recognition module recognizes the device state based on the decision tree algorithm, the traceability module supports accurate query and digital twin reproduction of the machining process, the predictive maintenance module outputs fault early warning based on the LSTM model, and the process optimization module outputs the optimal process parameters based on the genetic algorithm.

[0027] The network security layer comprises an industrial firewall module and a data security module, the industrial firewall module is arranged between the edge gateway and the numerical control system and the cloud, the data security module adopts SSL / TLS encryption transmission, AES-256 encryption storage and RBAC permission management, and guarantees the data security.

[0028] Further preferably, the universal protocol adaptation module adopts an industrial protocol converter, supports Ethernet and RS485 interfaces, and can add communication protocols through a configuration file; the serial port-Ethernet conversion module adopts a MOXAN Port5110 serial port server, supports TCP Server / Client and UDP communication modes, and is provided with a built-in CRC check algorithm; the non-invasive sensor acquisition module further comprises a JIANGHUAI USB-4711 data acquisition card, supports a 10-100Hz configurable sampling rate, and outputs analog / digital data.

[0029] Further preferably, the industrial edge gateway of the edge processing layer adopts a JIANGHUAI UNO-2484G, the CPU is an Intel Celeron J4125, the memory is greater than or equal to 8GB, and the storage is greater than or equal to 128GB SSD; the CPU of the cloud server of the data fusion layer is an Intel Xeon Gold 6330, the memory is greater than or equal to 128GB, and the storage is greater than or equal to 10TB SSD; the state identification and OEE calculation module of the application support layer is developed based on a Scikit-learn framework, and the predictive maintenance and process optimization module is developed based on a TensorFlow framework; the industrial firewall module adopts a Huawei USG6000E, and the data transmission delay is less than or equal to 100ms.

[0030] The embodiment of the present application has the following advantages due to the adoption of the above technical solutions:

[0031] Firstly, the present application supports the unified access of 10+ brands of numerical control machine tools, the data intercommunication rate of heterogeneous devices reaches 100%, solves the problem of "information island", identifies 6 types of device states, and improves the OEE calculation accuracy to ±2%, which is 5 times higher than that of the traditional method.

[0032] Secondly, the non-invasive sensor of the present application has a transformation cost of less than or equal to 10% of the original value of the machine tool, a hardware installation cost reduction of 70% compared with the traditional method, a data transmission delay of less than or equal to 100ms, a noise removal rate of greater than or equal to 92%, zero production interference (no CNC program disorder case), supports processing traceability (traceability accuracy to seconds), predictive maintenance (30% reduction in downtime), process optimization (15% efficiency improvement), and promotes the intelligent upgrading of numerical control machine tools.

[0033] The above summary is only for the purpose of the description and is not intended to limit in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present application will be readily apparent to those skilled in the art by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0035] Figure 1 Method flowchart of the present application. DETAILED DESCRIPTION

[0036] In the following, only some exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present application. Therefore, the drawings and the description are considered to be exemplary in nature rather than limiting.

[0037] The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0038] As Figure 1 shown, the embodiment of the present application provides a machining data acquisition method applied to a numerical control machine tool, comprising the following steps:

[0039] S1. Numerical control machine tool classification and access scheme design: according to the digitization ability, the numerical control machine tool is divided into type I intelligent new machine tool, type II semi-intelligent machine tool and type III old and outdated "dumb equipment", and a classification access scheme of general protocol adapter, serial-to-ethernet converter and non-invasive sensor combination is adopted respectively to realize data access of heterogeneous and old equipment;

[0040] S2. Multi-dimensional machining data real-time acquisition: for all types of machine tools, four-dimensional data of equipment state, machining parameter, process data and environment data are uniformly collected, time synchronization of all collection equipment is realized through NTP protocol, synchronization error is less than or equal to 50 ms, and accurate time stamp is ensured to be carried by data;

[0041] S3. Edge data preprocessing and safety isolation: an industrial-grade edge gateway is deployed to perform "Kalman filtering + moving average" combined noise reduction, LZ77 algorithm compression, local caching and network interruption transmission processing on the collected data, and at the same time, through an industrial firewall and read-only permission configuration, safety isolation of the collection process and the numerical control system is realized to ensure zero production interference;

[0042] S4. Data fusion and standardization: the preprocessed standardized data is uploaded to a cloud data platform, device state, machining parameter and process data are associated and fused through "device ID + time stamp" double primary key, data is output according to OPCUA or MQTTSparkplugB industrial standard, and three types of data interfaces of real-time, historical and statistical data are provided;

[0043] S5. Data value application: Based on fused data, identify the device fine-grained state through decision tree algorithm and calculate OEE, realize precise traceability and digital twin replication of processing process combined with timestamp, carry out predictive maintenance and process optimization using LSTM model and optimization algorithm.

[0044] In one embodiment, in S1, the first type of intelligent new machine tool supports Ethernet communication and standardized protocol, the general protocol adapter supports FOCAS2, S7, MCProtocol, HMI-Advanced protocol analysis library, and connects the numerical control system directly through Ethernet; the second type of semi-intelligent machine tool has RS232 / 485 serial port, and the serial port-Ethernet converter cooperates with frame analysis algorithm based on CRC check to extract spindle speed and feed data; the non-invasive sensor combination of the third type of old " dumb equipment" includes Hall current sensor, three-axis vibration sensor, photoelectric pulse collector and voltage detection module, which respectively collects spindle load, vibration, feed displacement and equipment state signal.

[0045] In one embodiment, in S2, the device state data includes device startup and shutdown, running / idle / alarm / debug / standby / maintenance state, spindle start and stop, and tool replacement state, with a collection frequency of 1 Hz; the processing parameter data includes spindle speed, feed rate, cutting depth, NC program number, and tool number, with a collection frequency of 10 Hz; the process data includes spindle load, spindle vibration value, feed shaft displacement, processing time, and product pass rate, with a collection frequency of 50 Hz; the environmental data includes machine tool working environment temperature, humidity, and power grid voltage, with a collection frequency of 0.1 Hz.

[0046] In one embodiment, in S3, the noise removal rate of the edge gateway is ≥92%, the data compression rate is 1:5, and the local cache capacity is ≥7 days of data storage; the security isolation adopts the "hardware isolation + permission control" double strategy, the hardware isolation opens only the data reading port through the industrial firewall, blocks the writing port, the permission control disables the writing instruction through the numerical control system background, configures the read-only permission of the collection device, and at the same time, the edge gateway installs industrial antivirus software and prohibits external mobile storage devices.

[0047] In one embodiment, in S4, the data correlation fusion includes state-parameter correlation, program-tool correlation, and process-quality correlation, which respectively correlate the device cutting state with the processing parameters, the NC program number with the tool number, and the spindle vibration value with the product pass rate to form a structured data file; the real-time data interface is WebSocket interface, which pushes second-level real-time data, the historical data interface is RESTful API, which supports query by device ID and time range, and the statistical data interface is SQL query interface, which outputs summary data such as daily OEE and tool service life.

[0048] In one embodiment, in S5, the fine-grained state recognition includes a running cutting state, an idle state, an alarm state, a debugging state, a waiting material state, and a maintenance state, wherein the running cutting state determination condition is spindle load > 2A + feed rate > 0 + NC program running, the idle state determination condition is spindle load < 1A + feed rate = 0 + no alarm; the OEE calculation accuracy is ± 2%, and the process optimization finds the optimal combination of spindle speed and feed rate through a genetic algorithm to achieve machining efficiency improvement ≥ 15% and product qualification rate ≥ 99.9%.

[0049] In one embodiment, in S1, in the non-invasive sensor combination of the III type old "dumb equipment", the Hall current sensor has a range of 0-5A and an accuracy of ± 1%, the three-axis vibration sensor has a range of ± 16g and a sampling rate of 100Hz, the photoelectric pulse collector has a resolution of 1000p / r, the voltage detection module monitors the 24V / 0V voltage signal of the operating panel indicator light, and the transformation cost of the old machine tool of the classification access scheme is ≤ 10% of the original value of the machine tool.

[0050] The machining data acquisition system applied to the numerical control machine tool comprises:

[0051] The device access layer comprises a general protocol adaptation module, a serial-to-ethernet conversion module, and a non-invasive sensor acquisition module, which respectively realize data access of the I type, the II type, and the III type numerical control machine tool, the general protocol adaptation module supports 10+ manufacturer communication protocols, and the non-invasive sensor acquisition module comprises a Hall current sensor, a three-axis vibration sensor, a photoelectric pulse collector, and a voltage detection module.

[0052] The edge processing layer comprises a data preprocessing module, a security isolation module, a local cache and continuation transmission module, and a time synchronization module, which are deployed on an industrial-grade edge gateway, the data preprocessing module realizes data noise reduction and compression, the security isolation module realizes industrial firewall isolation and read-only permission control, the local cache and continuation transmission module supports offline data storage and recovery for subsequent transmission, and the time synchronization module realizes time synchronization of the acquisition device through the NTP protocol.

[0053] The data fusion layer comprises a data correlation fusion module and a data standardization module, which are deployed on a cloud data platform, the data correlation fusion module correlates multi-source data based on the "device ID + timestamp" based on the Spark framework, and the data standardization module outputs data according to the OPCUA or MQTTSparkplugB standard.

[0054] The application support layer comprises a state recognition and OEE calculation module, a traceability and visualization module, a predictive maintenance and process optimization module, the state recognition module recognizes the device state based on the decision tree algorithm, the traceability module supports accurate query and digital twin reproduction of the machining process, the predictive maintenance module outputs fault warning based on the LSTM model, and the process optimization module outputs the optimal process parameters based on the genetic algorithm.

[0055] Network security layer: including industrial firewall module and data security module, the industrial firewall module is deployed between the edge gateway and the numerical control system and the cloud, the data security module adopts SSL / TLS encryption transmission, AES-256 encryption storage, RBAC permission management, and guarantees data security.

[0056] In one embodiment, the general protocol adaptation module adopts an industrial protocol converter, supports Ethernet and RS485 interface, and can add communication protocols through a configuration file; the serial-to-Ethernet conversion module adopts MOXAN Port5110 serial server, supports TCP Server / Client and UDP communication mode, and is provided with a built-in CRC check algorithm; the non-invasive sensor acquisition module further includes Advantech USB-4711 data acquisition card, supports 10-100Hz configurable sampling rate, and outputs analog / digital data.

[0057] In one embodiment, the industrial edge gateway of the edge processing layer adopts Advantech UNO-2484G, the CPU is Intel Celeron J4125, the memory is greater than or equal to 8GB, and the storage is greater than or equal to 128GB SSD; the CPU of the cloud server of the data fusion layer is Intel Xeon Gold6330, the memory is greater than or equal to 128GB, and the storage is greater than or equal to 10TB SSD; the state recognition and OEE calculation module of the application support layer is developed based on the Scikit-learn framework, the predictive maintenance and process optimization module is developed based on the TensorFlow framework; the industrial firewall module adopts Huawei USG6000E, and the data transmission delay is less than or equal to 100ms.

[0058] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of various changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A machining data acquisition method applied to a numerical control machine tool, characterized in that: Comprise the following steps: S1. Numerical control machine tool classification and access scheme design: according to the digital ability, the numerical control machine tool is divided into I type intelligent new machine tool, II type semi-intelligent machine tool and III type old " dumb equipment", respectively adopting general protocol adapter, serial port-Ethernet converter and non-invasive sensor combination classification access scheme, realizing data access of heterogeneous and old equipment; S2. Multi-dimensional processing data real-time acquisition: for all types of machine tools, four-dimensional data of equipment state, processing parameters, process data and environment data are uniformly collected, time synchronization of all collection equipment is realized through NTP protocol, synchronization error is less than or equal to 50ms, and accurate time stamp is ensured to be carried by data; S3. Edge data preprocessing and safety isolation: industrial edge gateway is deployed, and the collected data is subjected to "Kalman filtering + moving average" combined noise reduction, LZ77 algorithm compression, local cache and offline transmission processing, and at the same time, through industrial firewall and read-only permission configuration, the safety isolation of collection process and numerical control system is realized, and zero production interference is ensured; S4. Data fusion and standardization: the preprocessed standardized data is uploaded to the cloud data platform, the equipment state, processing parameters and process data are associated and fused through "equipment ID + time stamp" double primary key, the data is output according to OPCUA or MQTTSparkplugB industrial standard, and three types of data interfaces of real-time, historical and statistical data are provided; S5. Data value application: based on the fused data, the equipment fine-grained state is identified through decision tree algorithm and OEE is calculated, the processing process is accurately traced and digital twin is reproduced by combining time stamp, and LSTM model and optimization algorithm are used to carry out predictive maintenance and process optimization.

2. The machining data acquisition method for a numerical control machine tool according to claim 1, characterized in that: In the S1, the I type intelligent new machine tool supports Ethernet communication and standardization protocol, the general protocol adapter supports FOCAS2, S7, MCProtocol and HMI-Advanced protocol analysis library built-in, and directly connects the numerical control system through Ethernet; The II type semi-intelligent machine tool has RS232 / 485 serial port, the serial port-Ethernet converter cooperates with frame analysis algorithm based on CRC check to extract spindle speed and feed amount data; The non-invasive sensor combination of the III type old " dumb equipment" includes Hall current sensor, three-axis vibration sensor, photoelectric pulse collector and voltage detection module, which respectively collect spindle load, vibration, feed displacement and equipment state signal.

3. The machining data acquisition method for a numerical control machine tool according to claim 1, characterized in that: In the S2, the equipment state data includes equipment startup and shutdown, running / idle / alarm / debug / standby / maintenance state, spindle start and stop, tool replacement state, and the collection frequency is 1Hz; The processing parameter data includes spindle speed, feed rate, cutting depth, NC program number and tool number, and the collection frequency is 10Hz; The process data includes spindle load, spindle vibration value, feed shaft displacement, processing time and product qualification rate, and the collection frequency is 50Hz; The environment data includes machine tool working environment temperature, humidity and power grid voltage, and the collection frequency is 0.1Hz.

4. The machining data collection method for a CNC machine tool according to claim 1, characterized in that: In the S3, the noise removal rate of the edge gateway is greater than or equal to 92%, the data compression rate is 1:5, and the local cache capacity is greater than or equal to 7 days of data storage; the security isolation adopts a double strategy of "hardware isolation + permission control", the hardware isolation only opens the data reading port and blocks the writing port through the industrial firewall, the permission control disables the writing instruction through the numerical control system background and configures the read-only permission of the acquisition device, and the edge gateway is installed with industrial antivirus software to prohibit external mobile storage devices.

5. The method for collecting machining data applied to a numerical control machine tool according to claim 1, characterized in that: In the S4, the data correlation fusion includes state-parameter correlation, program-tool correlation and process-quality correlation, respectively correlates the equipment cutting state with the machining parameter, the NC program number with the tool number and the spindle vibration value with the product qualification rate, and forms a structured data file; the real-time data interface is a WebSocket interface, which pushes second-level real-time data, the historical data interface is a RESTful API, which supports query according to the equipment ID and time range, and the statistical data interface is a SQL query interface, which outputs summary data such as daily OEE and tool service life.

6. The method for collecting machining data applied to a numerical control machine tool according to claim 1, characterized in that: In the S5, the fine-grained state recognition includes running cutting state, idle state, alarm state, debugging state, waiting state and maintenance state, wherein the running cutting state determination condition is that the spindle load is greater than 2A, the feed rate is greater than 0 and the NC program is running, and the idle state determination condition is that the spindle load is less than 1A, the feed rate is 0 and there is no alarm; the OEE calculation accuracy is ±2%, and the process optimization finds the optimal combination of spindle speed and feed rate through a genetic algorithm, so that the machining efficiency is improved by greater than or equal to 15% and the product qualification rate is greater than or equal to 99.9%.

7. The method for collecting machining data applied to a numerical control machine tool according to claim 1, characterized in that: In the S1, in the non-invasive sensor combination of the III type old " dumb equipment", the Hall current sensor has a range of 0-5A and an accuracy of ±1%, the three-axis vibration sensor has a range of ±16g and a sampling rate of 100Hz, the photoelectric pulse collector has a resolution of 1000p / r, and the voltage detection module monitors the 24V / 0V voltage signal of the operating panel indicator light; the old machine tool transformation cost of the classification access scheme is less than or equal to 10% of the original value of the machine tool.

8. A processing data collection system applied to a numerical control machine tool, matched with the processing data collection method applied to the numerical control machine tool as claimed in any one of claims 1-7, characterized in that: It comprises: a device access layer, including a general protocol adaptation module, a serial-to-ethernet conversion module and a non-invasive sensor acquisition module, which respectively realize data access of the I type, II type and III type numerical control machine tools, the general protocol adaptation module supports 10+ manufacturer communication protocols, and the non-invasive sensor acquisition module includes a Hall current sensor, a three-axis vibration sensor, a photoelectric pulse collector and a voltage detection module; an edge processing layer, including a data preprocessing module, a security isolation module, a local cache and a continuation module and a time synchronization module, which are deployed on an industrial-grade edge gateway, the data preprocessing module realizes data denoising and compression, the security isolation module realizes industrial firewall isolation and read-only permission control, the local cache and the continuation module support offline data storage and recovery for subsequent transmission, and the time synchronization module realizes time synchronization of the acquisition device through the NTP protocol; The data fusion layer includes a data correlation fusion module and a data standardization module, is deployed in a cloud data platform, the data correlation fusion module correlates multi-source data based on a Spark framework according to "device ID + timestamp", and the data standardization module outputs data according to OPCUA or MQTTSparkplugB standards; The application support layer includes a state recognition and OEE calculation module, a traceability and visualization module, a predictive maintenance and process optimization module, the state recognition module recognizes device states based on a decision tree algorithm, the traceability module supports accurate query and digital twin reproduction of a processing process, the predictive maintenance module outputs fault warnings based on an LSTM model, and the process optimization module outputs optimal process parameters based on a genetic algorithm; The network security layer includes an industrial firewall module and a data security module, the industrial firewall module is deployed between an edge gateway and a numerical control system and a cloud, and the data security module uses SSL / TLS encryption transmission, AES-256 encryption storage and RBAC permission management to ensure data security.

9. The machining data acquisition system for use in a numerically controlled machine tool according to claim 8, characterized in that: The general protocol adaptation module uses an industrial protocol converter, supports Ethernet and RS485 interfaces, and can add communication protocols through a configuration file; the serial-to-Ethernet conversion module uses a MOXAN Port5110 serial server, supports TCP Server / Client and UDP communication modes, and has a built-in CRC check algorithm; the non-intrusive sensor acquisition module also includes an Advantech USB-4711 data acquisition card, supports a configurable sampling rate of 10-100Hz, and outputs analog / digital data.

10. The machining data acquisition system for use in a CNC machine tool according to claim 8, characterized in that: The industrial edge gateway of the edge processing layer uses an Advantech UNO-2484G, the CPU is an Intel Celeron J4125, the memory is greater than or equal to 8GB, and the storage is greater than or equal to 128GB SSD; the CPU of the cloud server of the data fusion layer is an Intel Xeon Gold 6330, the memory is greater than or equal to 128GB, and the storage is greater than or equal to 10TB SSD; the state recognition and OEE calculation module of the application support layer is developed based on a Scikit-learn framework, and the predictive maintenance and process optimization module is developed based on a TensorFlow framework; the industrial firewall module uses a Huawei USG6000E, and the data transmission delay is less than or equal to 100ms.

Images