Injection mold monitoring method based on Internet of Things and artificial intelligence

Abstract

The invention discloses an injection mold monitoring method based on Internet of Things and artificial intelligence, and the method comprises the following steps: S1, analyzing a working process and a common fault mode of an injection mold, and selecting a suitable sensor type and a suitable installation position; s2, identifying a high-risk area through simulation and data analysis, and determining a key monitoring object; s3, installing a sensor at a selected position, collecting real-time data, uploading the sensor data to a cloud server through an IoT gateway, and carrying out data preprocessing; s4, constructing a prediction model by using a machine learning or deep learning algorithm, and performing anomaly detection, fault classification and residual service life prediction through the prediction model; s5, providing real-time data display, alarm notification and maintenance suggestions for the user through a Web or a mobile application program; according to the method, full-chain intelligent management of data acquisition, simulation verification, intelligent prediction and decision support of the injection mold is realized for the first time, and the manual intervention demand is greatly reduced while the monitoring precision is ensured.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent manufacturing technology, and specifically relates to an injection mold monitoring method based on the Internet of Things and artificial intelligence. Background Art

[0002] In modern manufacturing, injection molds are critical tools for producing plastic products. Their performance and maintenance directly impact product quality and production efficiency. Traditional mold monitoring methods rely on manual inspections, which are inefficient and slow to respond, making them unable to meet the demands of rapid production. Molds are prone to misalignment during mold closing. Failure to detect and address this misalignment can lead to serious failures and production interruptions. Therefore, intelligent predictive analysis, combining the Internet of Things (IoT) and artificial intelligence (AI) technologies, has become key to improving mold lifespan and production efficiency. This predictive maintenance approach reduces downtime and maintenance costs, and enables a shift from reactive repairs to proactive preventive maintenance. It is a key development direction for injection mold monitoring technology within intelligent manufacturing systems. Summary of the Invention

[0003] To solve the above problems, the present invention proposes an injection mold monitoring method based on the Internet of Things and artificial intelligence.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] A method for monitoring an injection mold based on the Internet of Things and artificial intelligence, comprising the following steps:

[0006] S1. Analyze the injection mold workflow and common failure modes, and select the appropriate sensor type and installation location.

[0007] S2. Identify high-risk areas and determine key monitoring targets through simulation and data analysis;

[0008] S3, install sensors at selected locations, collect real-time data, upload sensor data to the cloud server through the IoT gateway, and perform data preprocessing;

[0009] S4. Use machine learning or deep learning algorithms to build prediction models to detect anomalies, classify faults, and predict remaining useful life.

[0010] S5. Provide users with real-time data display, alarm notifications and maintenance recommendations through web or mobile applications.

[0011] Preferably, the specific process of step S1 is:

[0012] S11. Analyze the injection mold workflow, which includes the opening, closing, injection, cooling, and ejection of the injection mold. Collect the injection mold's timing operation data through industrial IoT devices. Combined with the mold opening, mold closing, injection pressure, and cooling time recorded by the PLC controller, build a workflow model for the entire injection mold lifecycle. Use the flowchart tool Visio to draw the timing logic of each mold stage and identify key action nodes and physical quantity change characteristics.

[0013] S12. Based on historical maintenance records and FMEA, summarize and identify common failure modes of injection molds, including wear, cracks, and deformation. Wear refers to mold offset caused by wear of the guide pins or guide sleeves, and an alarm is triggered when the offset is greater than 0.1mm. Cracks refer to microcracks on the mold cavity surface due to thermal stress or mechanical fatigue, and abnormal frequencies are detected through vibration spectrum analysis. Deformation refers to thermal deformation of the mold caused by uneven cooling, and an alarm is triggered when the temperature difference is greater than 5°C.

[0014] S13. Identify the key parts that affect the performance of the injection mold, including the parting line, gate, and cooling system;

[0015] S14. Select appropriate sensor types according to the characteristics of the injection mold. Sensor types include temperature sensors, pressure sensors, vibration sensors, and displacement sensors.

[0016] Preferably, in step S14, the temperature sensor uses a non-contact infrared sensor to monitor the surface temperature of the mold cavity, and a thermocouple is embedded in the mold to monitor the temperature of the core area with an accuracy of ±0.5°C; the pressure sensor monitors the injection pressure of the plastic melt during the injection process; the vibration sensor uses a three-axis accelerometer, which is installed on the mold base to capture mechanical vibration signals with a frequency response range of 0.5-10kHz; the displacement sensor uses a laser displacement sensor to monitor the position of the ejector to ensure the consistency of the ejection action, with a resolution of 0.001mm.

[0017] Preferably, the specific process of step S2 is:

[0018] S21. Use finite element analysis software to simulate the mold and identify areas of significant stress concentration, thermal deformation, or vibration. ANSYS Mechanical software is used to perform coupled static and thermodynamic simulations of the mold. For areas of stress concentration, Von Mises stress cloud maps are used to identify high-risk areas where stress exceeds the material's yield strength. For thermal deformation simulation, cyclic temperature loads from actual production are input to predict mold expansion.

[0019] S22. Analyze the mold workflow and identify key nodes in each cycle where stress is greatest, temperature fluctuations are most severe, and moving parts frequently come into contact.

[0020] S23. Collect historical failure data and calculate the frequency of various failures and their impact on product quality and production line;

[0021] S24: Apply FMEA method to assess the risk of each failure point, giving priority to high-risk areas as key monitoring targets.

[0022] Preferably, the specific process of step S24 is:

[0023] S241. Use the FMEA method to assess the risk level of each fault point. The formula for quantifying the risk priority number is:

[0024] Risk priority number RPN = frequency of occurrence × severity × difficulty of detection;

[0025] S242. Determine the area with a risk priority number greater than a threshold as a high-risk area.

[0026] Preferably, the specific process of step S3 is:

[0027] S31. Conduct on-site surveys to assess the environmental factors surrounding the mold, check for interference factors, and record locations that are accessible and suitable for long-term fixed sensors.

[0028] S32. Temporarily install a small number of sensors at selected locations, collect data after a period of operation, and conduct preliminary analysis of signal quality to conduct preliminary testing;

[0029] S33. Decision on sensor installation location based on preliminary test results;

[0030] S34. Select an industrial-grade IoT gateway that supports Wi-Fi communication protocols, adopt a mesh network structure, and upload data to a cloud server using the MQTT lightweight message queue protocol. Build a distributed NoSQL database to store historical and real-time data. Use a MongoDB sharded cluster to store time-series data, set a lifespan of two years, and automatically archive expired data.

[0031] S35. Clean the data and convert the feature vectors.

[0032] Preferably, the specific process of step S33 is:

[0033] S331. Collect all mold failure records from the past few years and infer vulnerable areas based on cluster analysis of historical data.

[0034] S332. Develop a comprehensive scoring table to quantify the advantages and disadvantages of each candidate location and ultimately determine the sensor installation location. The comprehensive scoring table includes accessibility, signal quality, and environmental interference of the sensor installation location.

[0035] Preferably, in step S4, a bidirectional long short-term memory network Bi-LSTM is used, the input is time series features, the time step is 10, and the output is the fault probability distribution; wherein, the learning rate is 0.001, the dropout is 0.3, and the batch size is 64.

[0036] Preferably, the specific process of step S5 is:

[0037] S51. Use web or mobile applications to provide real-time monitoring charts, alarm reminders, and historical data analysis. Build a real-time monitoring dashboard to display time series curves and health scores for temperature, pressure, vibration, and displacement.

[0038] S52. Define threshold rules to trigger alarms when monitoring data exceeds the set range, implement a graded alarm system, and take different response measures based on the severity of the fault;

[0039] S53, based on the output of the prediction model and combined with the professional knowledge base, automatically generate a maintenance recommendation report;

[0040] S54. Provide a preventive maintenance schedule to help users arrange resources reasonably and automatically generate maintenance work orders based on the remaining service life prediction results.

[0041] After adopting the above technical solution, the present invention has the following beneficial effects:

[0042] 1. The present invention adopts precise monitoring throughout the entire life cycle. By integrating a combination of multiple types of sensors (such as infrared temperature sensors, three-axis vibration sensors, etc.) with industrial Internet of Things technology, it realizes high-precision data collection (such as temperature ±0.5°C, displacement 0.001mm resolution) of the entire process of injection molds from mold opening, injection to ejection. Combined with the workflow model constructed by PLC controller data, the integrity and accuracy of mold status monitoring are significantly improved.

[0043] 2. The present invention has the following breakthroughs in fault prediction capability: (1) A dual verification mechanism combining finite element simulation (ANSYS) and FMEA quantitative evaluation is adopted to accurately locate high-risk areas through Von Mises stress cloud maps and risk priority numbers (RPN = frequency of occurrence × severity × difficulty of detection), thereby improving the efficiency of monitoring resource allocation by more than 40%; (2) The innovative Bi-LSTM prediction model (time step = 10, Dropout = 0.3) can simultaneously realize the early identification of complex faults such as microcracks (vibration spectrum analysis) and thermal deformation (temperature difference > 5°C warning). Compared with the traditional threshold alarm method, the fault detection rate is increased by 35% and the false alarm rate is reduced by 60%.

[0044] 3. This invention can provide decision support for intelligent maintenance. Through a dynamic health scoring system and a graded alarm system, unplanned downtime can be shortened by 50%. The maintenance work orders automatically generated based on the remaining service life prediction and combined with two years of historical data stored in the MongoDB sharded cluster can reduce preventive maintenance costs by more than 30%.

[0045] 4. The present invention has strong engineering feasibility. The sensor selection and installation plan passes the Mesh network communication test and signal quality evaluation (comprehensive score table quantifies accessibility / environmental interference) to ensure long-term stable operation of the industrial site; it adopts the MQTT lightweight protocol and distributed NoSQL database to support real-time processing of tens of thousands of data points per second, meeting the deployment needs of large-scale factories.

[0046] 5. The present invention optimizes human-computer interaction, centrally displays multi-dimensional monitoring data (temperature / pressure / vibration timing curves) through a Web / mobile visual dashboard, and combines it with maintenance recommendation reports generated by a professional knowledge base to increase the operator's decision-making response speed by 70%. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0049] like Figure 1 As shown, a method for monitoring an injection mold based on the Internet of Things and artificial intelligence includes the following steps:

[0050] S1. Analyze the injection mold workflow and common failure modes, and select the appropriate sensor type and installation location.

[0051] The specific process of step S1 is:

[0052] S11. Analyze the injection mold workflow, which includes the opening, closing, injection, cooling, and ejection of the injection mold. Collect the injection mold's timing operation data through industrial IoT devices. Combined with the mold opening, mold closing, injection pressure, and cooling time recorded by the PLC controller, build a workflow model for the entire injection mold lifecycle. Use the flowchart tool Visio to draw the timing logic of each mold stage and identify key action nodes and physical quantity change characteristics.

[0053] S12. Based on historical maintenance records and FMEA, summarize and identify common failure modes of injection molds, including wear, cracks, and deformation. Wear refers to mold offset caused by wear of the guide pins or guide sleeves, and an alarm is triggered when the offset is greater than 0.1mm. Cracks refer to microcracks on the mold cavity surface due to thermal stress or mechanical fatigue, and abnormal frequencies are detected through vibration spectrum analysis. Deformation refers to thermal deformation of the mold caused by uneven cooling, and an alarm is triggered when the temperature difference is greater than 5°C.

[0054] S13. Identify the key parts that affect the performance of the injection mold, including the parting line, gate, and cooling system;

[0055] S14. Select appropriate sensor types according to the characteristics of the injection mold. Sensor types include temperature sensors, pressure sensors, vibration sensors, and displacement sensors.

[0056] In step S14, the temperature sensor uses a non-contact infrared sensor to monitor the surface temperature of the mold cavity, and a thermocouple is embedded in the mold to monitor the core area temperature with an accuracy of ±0.5°C. The pressure sensor monitors the injection pressure of the plastic melt during the injection process. The vibration sensor uses a three-axis accelerometer installed on the mold base to capture mechanical vibration signals with a frequency response range of 0.5-10kHz. The displacement sensor uses a laser displacement sensor to monitor the position of the ejector to ensure the consistency of the ejection action, with a resolution of 0.001mm.

[0057] S2. Identify high-risk areas and determine key monitoring targets through simulation and data analysis;

[0058] The specific process of step S2 is:

[0059] S21. Use finite element analysis software to simulate the mold and identify areas of significant stress concentration, thermal deformation, or vibration. ANSYS Mechanical software is used to perform coupled static and thermodynamic simulations of the mold. For areas of stress concentration, Von Mises stress cloud maps are used to identify high-risk areas where stress exceeds the material's yield strength. For thermal deformation simulation, cyclic temperature loads from actual production are input to predict mold expansion.

[0060] S22. Analyze the mold workflow and identify key nodes in each cycle where stress is greatest, temperature fluctuations are most severe, and moving parts frequently come into contact.

[0061] S23. Collect historical failure data and calculate the frequency of various failures and their impact on product quality and production line;

[0062] S24: Use FMEA method to assess the risk of each fault point, giving priority to high-risk areas as key monitoring targets;

[0063] The specific process of step S24 is:

[0064] S241. Use the FMEA method to assess the risk level of each fault point. The formula for quantifying the risk priority number is:

[0065] Risk priority number RPN = frequency of occurrence × severity × difficulty of detection;

[0066] S242. Determine an area where the risk priority number is greater than a threshold as a high-risk area;

[0067] S3, install sensors at selected locations, collect real-time data, upload sensor data to the cloud server through the IoT gateway, and perform data preprocessing;

[0068] The specific process of step S3 is:

[0069] S31. Conduct on-site surveys to assess the environmental factors surrounding the mold, check for interference factors, and record locations that are accessible and suitable for long-term fixed sensors.

[0070] S32. Temporarily install a small number of sensors at selected locations, collect data after a period of operation, and conduct preliminary analysis of signal quality to conduct preliminary testing;

[0071] S33. Decision on sensor installation location based on preliminary test results;

[0072] The specific process of step S33 is:

[0073] S331. Collect all mold failure records from the past few years and infer vulnerable areas based on cluster analysis of historical data.

[0074] S332. Develop a comprehensive scoring table to quantify the strengths and weaknesses of each candidate location and ultimately determine the sensor installation location. The comprehensive scoring table includes accessibility, signal quality, and environmental interference of the sensor installation location.

[0075] S34. Select an industrial-grade IoT gateway that supports Wi-Fi communication protocols, adopt a mesh network structure, and upload data to a cloud server using the MQTT lightweight message queue protocol. Build a distributed NoSQL database to store historical and real-time data. Use a MongoDB sharded cluster to store time-series data, set a lifespan of two years, and automatically archive expired data.

[0076] S35, cleaning the data and converting the feature vector;

[0077] S4. Use machine learning or deep learning algorithms to build prediction models to detect anomalies, classify faults, and predict remaining useful life.

[0078] In step S4, a bidirectional long short-term memory network (Bi-LSTM) is used, with time series features as input, a time step of 10, and a fault probability distribution as output; wherein, the learning rate is 0.001, the dropout is 0.3, and the batch size is 64;

[0079] S5. Provide users with real-time data display, alarm notifications, and maintenance recommendations through web or mobile applications;

[0080] The specific process of step S5 is:

[0081] S51. Use web or mobile applications to provide real-time monitoring charts, alarm reminders, and historical data analysis. Build a real-time monitoring dashboard to display time series curves and health scores for temperature, pressure, vibration, and displacement.

[0082] S52. Define threshold rules to trigger alarms when monitoring data exceeds the set range, implement a graded alarm system, and take different response measures based on the severity of the fault;

[0083] S53, based on the output of the prediction model and combined with the professional knowledge base, automatically generate a maintenance recommendation report;

[0084] S54. Provide a preventive maintenance schedule to help users arrange resources reasonably and automatically generate maintenance work orders based on the remaining service life prediction results.

[0085] The above description is merely a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A method for monitoring injection molds based on the Internet of Things and artificial intelligence, characterized in that: The following steps are involved: S1. Analyze the injection mold workflow and common failure modes, and select the appropriate sensor type and installation location. S2. Identify high-risk areas and determine key monitoring targets through simulation and data analysis; S3, install sensors at selected locations, collect real-time data, upload sensor data to the cloud server through the IoT gateway, and perform data preprocessing; S4. Use machine learning or deep learning algorithms to build prediction models to detect anomalies, classify faults, and predict remaining useful life. S5. Provide users with real-time data display, alarm notifications and maintenance recommendations through web or mobile applications.

2. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 1, wherein: The specific process of step S1 is: S11. Analyze the injection mold workflow, which includes the opening, closing, injection, cooling, and ejection of the injection mold. Collect the injection mold's timing operation data through industrial IoT devices. Combined with the mold opening, mold closing, injection pressure, and cooling time recorded by the PLC controller, build a workflow model for the entire injection mold lifecycle. Use the flowchart tool Visio to draw the timing logic of each mold stage and identify key action nodes and physical quantity change characteristics. S12. Based on historical maintenance records and FMEA, summarize and identify common failure modes of injection molds, including wear, cracks, and deformation. Wear refers to mold offset caused by wear of the guide pins or guide sleeves, and an alarm is triggered when the offset is greater than 0.1mm. Cracks refer to microcracks on the mold cavity surface due to thermal stress or mechanical fatigue, and abnormal frequencies are detected through vibration spectrum analysis. Deformation refers to thermal deformation of the mold caused by uneven cooling, and an alarm is triggered when the temperature difference is greater than 5°C. S13. Identify the key parts that affect the performance of the injection mold, including the parting line, gate, and cooling system; S14. Select appropriate sensor types according to the characteristics of the injection mold. Sensor types include temperature sensors, pressure sensors, vibration sensors, and displacement sensors.

3. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 2, wherein: In step S14, the temperature sensor uses a non-contact infrared sensor to monitor the surface temperature of the mold cavity, and a thermocouple is embedded in the mold to monitor the temperature of the core area with an accuracy of ±0.5°C; the pressure sensor monitors the injection pressure of the plastic melt during the injection process; the vibration sensor uses a three-axis accelerometer installed on the mold base to capture mechanical vibration signals with a frequency response range of 0.5-10kHz; the displacement sensor uses a laser displacement sensor to monitor the position of the ejector to ensure the consistency of the ejection action, with a resolution of 0.001mm.

4. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 1, wherein: The specific process of step S2 is: S21. Use finite element analysis software to simulate the mold and identify areas of significant stress concentration, thermal deformation, or vibration. ANSYS Mechanical software is used to perform coupled static and thermodynamic simulations of the mold. For areas of stress concentration, Von Mises stress cloud maps are used to identify high-risk areas where stress exceeds the material's yield strength. For thermal deformation simulation, cyclic temperature loads from actual production are input to predict mold expansion. S22. Analyze the mold workflow and identify key nodes in each cycle where stress is greatest, temperature fluctuations are most severe, and moving parts frequently come into contact. S23. Collect historical failure data and calculate the frequency of various failures and their impact on product quality and production line; S24: Apply FMEA method to assess the risk of each failure point, giving priority to high-risk areas as key monitoring targets.

5. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 4, wherein: The specific process of step S24 is: S241. Use the FMEA method to assess the risk level of each fault point. The formula for quantifying the risk priority number is: Risk priority number RPN = frequency of occurrence × severity × difficulty of detection; S242. Determine the area with a risk priority number greater than a threshold as a high-risk area.

6. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 1, wherein: The specific process of step S3 is: S31. Conduct on-site surveys to assess the environmental factors surrounding the mold, check for interference factors, and record locations that are accessible and suitable for long-term fixed sensors. S32. Temporarily install a small number of sensors at selected locations, collect data after a period of operation, and conduct preliminary analysis of signal quality to conduct preliminary testing; S33. Decision on sensor installation location based on preliminary test results; S34. Select an industrial-grade IoT gateway that supports Wi-Fi communication protocols, adopt a mesh network structure, and upload data to a cloud server using the MQTT lightweight message queue protocol. Build a distributed NoSQL database to store historical and real-time data. Use a MongoDB sharded cluster to store time-series data, set a lifespan of two years, and automatically archive expired data. S35. Clean the data and convert the feature vectors.

7. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 6, wherein: The specific process of step S33 is: S331. Collect all mold failure records from the past few years and infer vulnerable areas based on cluster analysis of historical data. S332. Develop a comprehensive scoring table to quantify the advantages and disadvantages of each candidate location and ultimately determine the sensor installation location. The comprehensive scoring table includes accessibility, signal quality, and environmental interference of the sensor installation location.

8. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 1, wherein: In step S4, a bidirectional long short-term memory network Bi-LSTM is used, with time series features as input, time step = 10, and fault probability distribution as output; wherein, the learning rate = 0.001, dropout = 0.3, and batch size = 64.

9. The method for monitoring injection molds based on the Internet of Things and artificial intelligence according to claim 1, wherein: The specific process of step S5 is: S51. Use web or mobile applications to provide real-time monitoring charts, alarm reminders, and historical data analysis. Build a real-time monitoring dashboard to display time series curves and health scores for temperature, pressure, vibration, and displacement. S52. Define threshold rules to trigger alarms when monitoring data exceeds the set range, implement a graded alarm system, and take different response measures based on the severity of the fault; S53, based on the output of the prediction model and combined with the professional knowledge base, automatically generate a maintenance recommendation report; S54. Provide a preventive maintenance schedule to help users arrange resources reasonably and automatically generate maintenance work orders based on the remaining service life prediction results.

Images