Multi-parameter correlation mapping-based electric spindle collision event identification system

Abstract

The application discloses a multi-parameter correlation mapping-based electric spindle crash event identification system, which comprises a multi-parameter sensor array, an intelligent data acquisition unit, an AI correlation mapping processing unit and an execution control unit, wherein the multi-parameter sensor array comprises a three-axis MEMS vibration sensor, a PT100 / KTY84 temperature sensor, a rotating speed encoder and a tool position sensor, the three-axis MEMS vibration sensor adopts a square flat or cylindrical structure and is installed at the nose end of the electric spindle or a preset measurement point position. The multi-parameter correlation mapping-based electric spindle crash event identification system can mine the internal correlation of vibration, temperature and rotating speed parameters through an AI algorithm, for example, when a crash occurs, not only the vibration peak value mutates suddenly, but also the rotating speed drops suddenly and the temperature rises instantaneously, so that the multi-parameter collaborative determination can greatly improve the identification accuracy and reduce the misjudgment rate, the multi-parameter fusion determination can avoid the misjudgment problem of single parameter monitoring, and the accuracy of crash identification is improved.

Description

Technical Field

[0001] This invention relates to the field of electric spindle monitoring technology, specifically to an electric spindle collision event identification system based on multi-parameter correlation mapping. Background Technology

[0002] As a core component of machine tools, the operational stability of electric spindles directly affects machining accuracy and production safety. In high-speed machining scenarios, wear and tear from long-term machine tool use can lead to increased tool vibration, which can easily cause machine collisions, resulting in serious losses such as spindle damage and workpiece scrap. Traditional electric spindle monitoring often uses a single vibration sensor or analog signal transmission scheme, which has the following technical drawbacks: 1. Single-parameter monitoring is susceptible to environmental interference, resulting in a high false alarm rate in case of collisions; 2. Analog signal transmission has weak anti-interference capabilities, and data accuracy is easily attenuated; 3. Lack of multi-parameter correlation analysis makes it difficult to distinguish between normal vibration and collision impact; 4. Slow response speed (usually exceeding 1ms), making it impossible to avoid collision losses in a timely manner; 5. Complex installation, poor adaptability, and unsuitable for retrofitting old machine tools.

[0003] In existing technologies, some monitoring systems attempt to increase the number of parameter monitoring dimensions, but these suffer from problems such as data processing delays, complex fusion algorithms, and high user operation thresholds, making it difficult to meet the real-time and practical requirements of industrial scenarios. Therefore, developing a multi-parameter collaborative, fast-response, and easy-to-install collision event recognition system has significant engineering application value. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an electric spindle collision event identification system based on multi-parameter correlation mapping. This system addresses the technical problems of traditional electric spindle monitoring technologies, such as: single-parameter monitoring being susceptible to environmental interference, resulting in a high false alarm rate for collisions; weak anti-interference capability of analog signal transmission, leading to easy attenuation of data accuracy; lack of multi-parameter correlation analysis, making it difficult to distinguish between normal vibration and collision impact; slow response speed, making it impossible to avoid collision losses in a timely manner; and complex installation, poor adaptability, and unsuitability for retrofitting old machine tools.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an electric spindle collision event recognition system based on multi-parameter correlation mapping, comprising a multi-parameter sensor array, an intelligent data acquisition unit, an AI correlation mapping processing unit, and an execution control unit. The multi-parameter sensor array includes a triaxial MEMS vibration sensor, a PT100 / KTY84 temperature sensor, a speed encoder, and a tool position sensor. The triaxial MEMS vibration sensor adopts a square-flat (16×20×7mm) or cylindrical (M10×35mm) structure and is installed at the nose of the electric spindle or at a preset measuring point position to collect XYZ triaxial vibration data. The sampling frequency is up to 26,700 times / second, and the resolution is 0.061mg (±2g range). The intelligent data acquisition unit is connected to the sensor array through the PLC IO terminal, supports 5-channel temperature input of PT100 / 1000×5, 2-channel accelerometer signal and 4-20mA analog input, has IP66 protection level, can realize data storage with adjustable frequency of 0.1~5 seconds, and can cyclically record more than 30 days of operating data; The AI ​​correlation mapping processing unit has a built-in front-end signal preprocessing module and a back-end multi-parameter fusion algorithm. It achieves anti-interference processing of vibration, temperature and rotation speed parameters through digital signal transmission, and uses AI noise reduction technology to establish a multi-parameter correlation model. It automatically learns the vibration curve of the first piece and sets a dynamic threshold. The execution control unit is linked with the CNC controller. When the multi-parameter fusion criteria meet the collision conditions, an emergency stop signal is triggered within 0.1 seconds, and two-level alarm reports are output simultaneously.

[0006] Preferably, the method for establishing the multi-parameter association mapping model includes: Step 1: In the first piece machining stage, the system collects the triaxial vibration peak value, temperature fluctuation range, and speed stability parameters of different tool positions (up to 64 tools) under normal working conditions, and establishes the reference parameter matrix for each tool position; Step 2: Extract the correlation coefficients of vibration-temperature and vibration-rotation speed using AI algorithms, construct a multi-parameter coupling mapping relationship, and set a first-level warning threshold (vibration 3~7g) and a second-level shutdown threshold (vibration ≥30g). Step 3: In the real-time processing stage, the collected real-time parameters are compared with the benchmark parameter matrix, and the parameter deviation and coupling feature matching degree are calculated. When any parameter exceeds the threshold and the correlation feature matching degree is ≥90%, it is determined to be a collision event.

[0007] Preferably, the intelligent data acquisition unit supports adaptation to new and old machine tools. New machine tools can be directly integrated through the CNC controller IO terminals, while old machine tools can install sensors by magnetic adsorption or drilling and locking, without the need for additional modification to the machine tool structure.

[0008] Preferably, the AI ​​association mapping processing unit adopts a "pre-processing + fusion" architecture. The pre-processing stage completes digital signal noise reduction and feature extraction, and the fusion stage realizes real-time fusion of multiple parameters through the EtherCAT communication interface. The professional version supports user-defined algorithm development, and the standard version has a built-in ready-to-use human-machine interface and IPC program.

[0009] Preferably, the determination of the collision event adopts a multi-parameter collaborative criterion: when any peak value of the triaxial vibration exceeds 30g, and is accompanied by a sudden change in rotational speed ≥10rpm, a sudden temperature fluctuation exceeding 1.17℃, or a tool position signal offset ≥0.1V, the system triggers an emergency shutdown; when the vibration parameter is in the range of 3~7g and the associated parameter shows an abnormal trend, a first-level warning is initiated.

[0010] Preferably, the execution control unit includes two DO alarm outputs, which can be directly connected to the PLCDI / DO interface of the CNC controller. The alarm signal duration is ≥100ms to ensure reliable spindle shutdown. It also supports parameter configuration and status monitoring via Wi-Fi connection to a tablet or host computer.

[0011] Preferably, the electric spindle collision event recognition system based on multi-parameter correlation mapping includes the following steps: S1: System initialization, sensor calibration and communication connection are completed, machine learning is performed through the first piece machining, and a multi-parameter benchmark model for each tool position is established. S2: Real-time acquisition of triaxial vibration data, temperature data, speed data and tool position data of the electric spindle, and transmission to the intelligent data acquisition unit via digital signals; S3: The AI ​​association mapping processing unit performs noise reduction on the collected data, extracts feature parameters such as vibration peak value, temperature change rate, and rotational speed fluctuation rate, and calculates the multi-parameter fusion confidence based on the preset association model; S4: When the fusion confidence level reaches the collision determination threshold, the AI ​​processing unit sends a trigger signal to the execution control unit. The execution control unit then links with the CNC controller to achieve an emergency stop and records the raw data for 20ms before and 70ms after the collision. S5: The system stores collision event information (including parameters such as time, vibration level, rotational speed, and temperature), and supports data export and traceability analysis.

[0012] Compared with the prior art, the beneficial effects of the present invention are: (1) The electric spindle collision event recognition system based on multi-parameter correlation mapping uses AI algorithm to mine the intrinsic correlation of vibration, temperature and speed parameters. For example, when a collision occurs, not only does the vibration peak change suddenly, but it is also accompanied by a sudden drop in speed and an instantaneous increase in temperature. Multi-parameter collaborative judgment greatly improves the recognition accuracy and reduces the false judgment rate. Multi-parameter fusion judgment avoids the false judgment problem of single parameter monitoring, and the collision recognition accuracy is improved. (2) The electric spindle collision event recognition system based on multi-parameter correlation mapping has a fast response speed and an emergency shutdown response within 0.1 seconds, which is significantly shorter than the traditional system (response time > 1ms) and reduces collision losses; (3) The electric spindle collision event recognition system based on multi-parameter correlation mapping features a miniaturized sensor and magnetic base mounting design, which is compatible with both new and old machine tools, reduces retrofitting costs, and allows the installation position to be closer to the vibration source, resulting in more accurate data acquisition. (4) The electric spindle collision event recognition system based on multi-parameter correlation mapping supports independent management of 64 tools and two-level alarm settings. The parameter threshold can be customized according to the processing requirements to adapt to different processing scenarios. Detailed Implementation

[0013] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] This invention provides a technical solution: an electric spindle collision event recognition system based on multi-parameter correlation mapping, comprising a multi-parameter sensor array, an intelligent data acquisition unit, an AI correlation mapping processing unit, and an execution control unit. The multi-parameter sensor array includes a triaxial MEMS vibration sensor, a PT100 / KTY84 temperature sensor, a speed encoder, and a tool position sensor. The triaxial MEMS vibration sensor adopts a square-flat (16×20×7mm) or cylindrical (M10×35mm) structure and is installed at the nose of the electric spindle or at a preset measuring point position to collect XYZ triaxial vibration data. The sampling frequency is up to 26,700 times / second, and the resolution is 0.061mg (±2g range). Specifically, the multi-parameter sensor array is responsible for comprehensively collecting the operating status parameters of the electric spindle. Among them, the triaxial MEMS vibration sensor adopts a miniaturized design and can be installed on the nose of the electric spindle (old machine tools) or in a pre-set mounting slot (new machine tools) by magnetic adsorption or adhesive, avoiding damage to the machine tool structure and ensuring that the sensor is in close contact with the spindle surface. The PT100 / KTY84 temperature sensor is installed in the spindle bearing housing and motor housing to monitor the temperature of key parts of the spindle with an accuracy of 0.1℃. The speed encoder is connected to the spindle output end, and the tool position sensor is connected to the turret signal interface. The speed encoder and the tool position sensor respectively collect the spindle speed (error ±0.2rpm) and the tool positioning signal (error ±0.1V), providing a data basis for multi-parameter correlation. The intelligent data acquisition unit connects to the sensor array via PLC IO terminals, supports 5-channel temperature input from PT100 / 1000, 2-channel accelerometer signals, and 4-20mA analog input. It has an IP66 protection rating and can store data at an adjustable frequency of 0.1~5 seconds, and cyclically record more than 30 days of operating data. Specifically, the intelligent data acquisition unit is installed at the rear of the electric spindle. As the core of data transmission and temporary storage, it is powered by a 24V power supply and supports wired and wireless (Wi-Fi) communication. It can be directly connected to the CNC controller IO terminals of new and old machine tools. The unit has a built-in microcomputer, eliminating the need for additional human-machine interface and IPC hardware, thus reducing system deployment costs. It also has a collision data storage function, which can store more than 1,000 collision event records. The AI ​​correlation mapping processing unit has a built-in front-end signal preprocessing module and a back-end multi-parameter fusion algorithm. It achieves anti-interference processing of vibration, temperature and speed parameters through digital signal transmission, and uses AI noise reduction technology to establish a multi-parameter correlation model. It automatically learns the vibration curve of the first piece and sets a dynamic threshold. Specifically, the AI ​​correlation mapping processing unit, as the core processing module of the system, adopts an innovative AI front-end and back-end division architecture. The front-end module is responsible for digital signal noise reduction and parameter preprocessing, solving the problem that traditional analog signals are easily interfered with. The back-end module establishes a coupled mapping relationship between vibration, temperature and rotation speed through a multi-parameter correlation algorithm, automatically learns the normal operating parameter range under different tool positions and different processing conditions, dynamically adjusts the judgment threshold, and realizes independent vibration management and two-level alarm settings for 64 tools. The execution control unit and CNC controller work together. When the multi-parameter fusion criteria meet the collision conditions, an emergency stop signal is triggered within 0.1 seconds, and two-level alarm reports are output simultaneously. Specifically, the execution control unit and the CNC controller work together in real time. When the system determines that a collision event has occurred, it sends an emergency stop signal within 0.1 seconds (the actual response speed depends on the processing speed of the CNC controller) and triggers an audible and visual alarm through the DO output interface. This unit supports user-defined stop thresholds, taking into account both machine tool protection and machining flexibility. Furthermore, methods for establishing multi-parameter association mapping models include: Step 1: In the first piece machining stage, the system collects the triaxial vibration peak value, temperature fluctuation range, and speed stability parameters of different tool positions (up to 64 tools) under normal working conditions, and establishes the reference parameter matrix for each tool position; Step 2: Extract the correlation coefficients of vibration-temperature and vibration-rotation speed using AI algorithms, construct a multi-parameter coupling mapping relationship, and set a first-level warning threshold (vibration 3~7g) and a second-level shutdown threshold (vibration ≥30g). Step 3: In the real-time processing stage, the collected real-time parameters are compared with the benchmark parameter matrix, and the parameter deviation and coupling feature matching degree are calculated. When any parameter exceeds the threshold and the correlation feature matching degree is ≥90%, it is determined to be a collision event. Furthermore, the intelligent data acquisition unit supports adaptation to both new and old machine tools. New machine tools can be directly integrated through the CNC controller IO terminals, while old machine tools can install sensors by magnetic adsorption or drilling and locking, without the need for additional modifications to the machine tool structure. Furthermore, the AI ​​association mapping processing unit adopts a "pre-processing + post-fusion" architecture. The pre-processing stage completes digital signal noise reduction and feature extraction, while the post-fusion stage achieves real-time fusion of multiple parameters through the EtherCAT communication interface. The professional version supports user-defined algorithm development, while the standard version has a built-in ready-to-use human-machine interface and IPC program. Furthermore, the determination of collision events adopts a multi-parameter collaborative criterion: when any peak value of triaxial vibration exceeds 30g, and is accompanied by a sudden change in rotational speed ≥10rpm, a sudden temperature fluctuation exceeding 1.17℃, or a tool position signal offset ≥0.1V, the system triggers an emergency shutdown; when the vibration parameters are in the range of 3~7g and the associated parameters show an abnormal trend, a first-level warning is activated. Furthermore, the execution control unit includes DO×2 alarm outputs, which can be directly connected to the PLC DI / DO interface of the CNC controller. The alarm signal duration is ≥100ms to ensure reliable spindle shutdown. It also supports parameter configuration and status monitoring via Wi-Fi connection to a tablet or host computer. Furthermore, the electric spindle collision event recognition system based on multi-parameter correlation mapping includes the following steps: S1: System initialization, sensor calibration and communication connection are completed, machine learning is performed through the first piece machining, and a multi-parameter benchmark model for each tool position is established. S2: Real-time acquisition of triaxial vibration data, temperature data, speed data and tool position data of the electric spindle, and transmission to the intelligent data acquisition unit via digital signals; S3: The AI ​​association mapping processing unit performs noise reduction on the collected data, extracts feature parameters such as vibration peak value, temperature change rate, and rotational speed fluctuation rate, and calculates the multi-parameter fusion confidence based on the preset association model; S4: When the fusion confidence level reaches the collision determination threshold, the AI ​​processing unit sends a trigger signal to the execution control unit. The execution control unit then links with the CNC controller to achieve an emergency stop and records the raw data for 20ms before and 70ms after the collision. S5: The system stores collision event information (including parameters such as time, vibration level, rotational speed, and temperature), and supports data export and traceability analysis; Workflow: Benchmark model establishment: During the first piece machining, the system automatically collects machining data of 64 tools, including the triaxial vibration curves of each tool position at different speeds, temperature fluctuation range (≤±1.17℃), and speed stability parameters, and establishes a benchmark model of tool position-parameter-threshold correlation. Real-time monitoring and data processing: During normal processing, the sensor array collects multi-parameter data at a frequency of up to 26,700 times / second, which is transmitted to the intelligent acquisition unit through digital signals. The AI ​​front-end module performs noise reduction processing and extracts characteristic parameters such as vibration peak and temperature change rate. Correlation mapping and collision determination: The AI ​​back-end module correlates and compares real-time feature parameters with the benchmark model and calculates the multi-parameter fusion confidence. When the vibration peak value is ≥30g, and the speed change is ≥10rpm and the temperature fluctuation exceeds 1.17℃, the fusion confidence is ≥90%, and the system determines it to be a collision event. Response execution: The execution control unit immediately sends an emergency stop signal to the CNC controller, stopping the spindle within 0.1 seconds. At the same time, it outputs a first-level alarm (indicator light) and a second-level alarm (audible prompt), and records key data such as collision time, vibration level, and speed. Performance testing: Collision recognition accuracy: The system accurately recorded the collision events in six simulated collision tests of different intensities, with an extreme collision error of ≤1.67g and a position signal recognition error of ≤0.1V. Response speed: After the vibration signal reaches the 30g threshold, the average system response time is 0.08 seconds, which meets the emergency stop requirement within 0.1 seconds; Anti-interference capability: In the complex electromagnetic environment of the workshop, the anti-interference capability of digital signal transmission is 80% higher than that of traditional analog signals, and the vibration data measurement error is ≤0.19g; Processing adaptability: After continuous operation at 10,000 rpm, the system can be stably monitored for more than 2 hours, with a temperature measurement error of ≤1.17℃ and a speed measurement error of ≤0.2 rpm; Any content not described in detail in this specification is prior art known to those skilled in the art.

[0015] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An electric spindle collision event recognition system based on multi-parameter correlation mapping, comprising a multi-parameter sensor array, an intelligent data acquisition unit, an AI correlation mapping processing unit, and an execution control unit, characterized in that: The multi-parameter sensor array includes a triaxial MEMS vibration sensor, a PT100 / KTY84 temperature sensor, a speed encoder, and a tool position sensor. The triaxial MEMS vibration sensor adopts a square-flat (16×20×7mm) or cylindrical (M10×35mm) structure and is installed at the nose of the electric spindle or at a preset measuring point position to collect XYZ triaxial vibration data. The sampling frequency is up to 26,700 times / second, and the resolution is 0.061mg (±2g range). The intelligent data acquisition unit is connected to the sensor array through the PLC IO terminal, supports 5-channel temperature input of PT100 / 1000×5, 2-channel accelerometer signal and 4-20mA analog input, has IP66 protection level, can realize data storage with adjustable frequency of 0.1~5 seconds, and can cyclically record more than 30 days of operating data. The AI ​​correlation mapping processing unit has a built-in front-end signal preprocessing module and a back-end multi-parameter fusion algorithm. It achieves anti-interference processing of vibration, temperature and rotation speed parameters through digital signal transmission, and uses AI noise reduction technology to establish a multi-parameter correlation model. It automatically learns the vibration curve of the first piece and sets a dynamic threshold. The execution control unit is linked with the CNC controller. When the multi-parameter fusion criteria meet the collision conditions, an emergency stop signal is triggered within 0.1 seconds, and a two-level alarm report is output at the same time.

2. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to claim 1, characterized in that: The method for establishing the multi-parameter association mapping model includes: Step 1: In the first piece machining stage, the system collects the triaxial vibration peak value, temperature fluctuation range, and speed stability parameters of different tool positions (up to 64 tools) under normal working conditions, and establishes the reference parameter matrix for each tool position; Step 2: Extract the correlation coefficients of vibration-temperature and vibration-rotation speed using AI algorithms, construct a multi-parameter coupling mapping relationship, and set a first-level warning threshold (vibration 3~7g) and a second-level shutdown threshold (vibration ≥30g). Step 3: In the real-time processing stage, the collected real-time parameters are compared with the benchmark parameter matrix, and the parameter deviation and coupling feature matching degree are calculated. When any parameter exceeds the threshold and the correlation feature matching degree is ≥90%, it is determined to be a collision event.

3. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to claim 1, characterized in that: The intelligent data acquisition unit supports adaptation to both new and old machine tools. New machine tools can be directly integrated through the CNC controller IO terminals, while old machine tools can install sensors by magnetic adsorption or drilling and locking, without the need for additional modifications to the machine tool structure.

4. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to claim 1, characterized in that: The AI ​​association mapping processing unit adopts a "pre-processing + fusion" architecture. The pre-processing stage completes digital signal noise reduction and feature extraction, while the fusion stage achieves real-time fusion of multiple parameters through the EtherCAT communication interface. The professional version supports user-defined algorithm development, while the standard version has a built-in ready-to-use human-machine interface and IPC program.

5. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to claim 1, characterized in that: The determination of the collision event adopts a multi-parameter collaborative criterion: when any peak value of the triaxial vibration exceeds 30g, and is accompanied by a sudden change in rotational speed ≥10rpm, a sudden temperature fluctuation exceeding 1.17℃, or a tool position signal offset ≥0.1V, the system triggers an emergency shutdown; when the vibration parameter is in the range of 3~7g and the associated parameter shows an abnormal trend, a first-level warning is initiated.

6. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to claim 1, characterized in that: The execution control unit includes two DO alarm outputs, which can be directly connected to the PLC DI / DO interface of the CNC controller. The alarm signal duration is ≥100ms to ensure reliable spindle shutdown. It also supports parameter configuration and status monitoring via Wi-Fi connection to a tablet or host computer.

7. The electric spindle collision event recognition system based on multi-parameter correlation mapping according to any one of claims 1-6, characterized in that: Includes the following steps: S1: System initialization, sensor calibration and communication connection are completed, machine learning is performed through the first piece machining, and a multi-parameter benchmark model for each tool position is established. S2: Real-time acquisition of triaxial vibration data, temperature data, speed data and tool position data of the electric spindle, and transmission to the intelligent data acquisition unit via digital signals; S3: The AI ​​association mapping processing unit performs noise reduction on the collected data, extracts feature parameters such as vibration peak value, temperature change rate, and rotational speed fluctuation rate, and calculates the multi-parameter fusion confidence based on the preset association model; S4: When the fusion confidence level reaches the collision determination threshold, the AI ​​processing unit sends a trigger signal to the execution control unit. The execution control unit then links with the CNC controller to achieve an emergency stop and records the raw data for 20ms before and 70ms after the collision. S5: The system stores collision event information (including parameters such as time, vibration level, rotational speed, and temperature), and supports data export and traceability analysis.