A mold life prediction and maintenance method and intelligent management platform based on the Internet of Things

Abstract

This invention relates to the field of intelligent mold management, and in particular to an IoT-based method and intelligent management platform for mold life prediction and maintenance, comprising the following steps: Step 1: Data acquisition; Step 2: Data preprocessing; Step 3: Feature extraction; Step 4: Life prediction model training and inference; Step 5: Dynamic maintenance strategy generation; Step 6: Maintenance execution and feedback. The beneficial effects of this invention are: by employing a hybrid model integrating deep learning and mechanistic analysis, it captures both the nonlinear characteristics in the data and incorporates the fatigue damage mechanism of the mold, effectively improving the accuracy of remaining life prediction, with prediction errors controlled within 5%; by generating personalized maintenance plans based on real-time operating data, remaining life prediction values, and production plans, it avoids the drawbacks of fixed-cycle maintenance, reduces over-maintenance and under-maintenance, and lowers maintenance costs and production downtime losses.

Description

Technical Field

[0001] This invention relates to the field of intelligent mold management, and in particular to a method and intelligent management platform for predicting and maintaining mold life based on the Internet of Things. Background Technology

[0002] Molds, as core process equipment in industrial production, are widely used in industries such as automotive, electronics, and home appliances. The service life and operating condition of molds directly affect product quality, production efficiency, and production costs. Traditional mold management methods primarily rely on the experience of operators for maintenance decisions, which has the following drawbacks:

[0003] Inaccurate lifespan prediction: Relying solely on experience to judge the remaining lifespan of the mold can easily lead to over-maintenance or under-maintenance. Over-maintenance increases costs, while under-maintenance may cause sudden mold failures, resulting in production interruptions. Delayed data acquisition: Relying on manual recording of mold operation data results in poor data integrity and real-time performance, making it impossible to promptly capture abnormal mold states. Rigid maintenance strategies: Adopting a fixed-cycle maintenance model without considering the actual operating status of the mold and production plans leads to insufficient flexibility. Low management efficiency: The lack of a unified management platform results in scattered historical data and maintenance records for the mold, making it difficult to achieve full lifecycle traceability and analysis.

[0004] With the development of technologies such as the Internet of Things (IoT) and artificial intelligence (AI), some enterprises have begun to apply IoT technology to mold monitoring. However, existing solutions mostly only achieve simple data collection and display, lacking accurate lifespan prediction models and dynamic maintenance decision-making mechanisms. This fails to fully leverage the value of the data and cannot fundamentally solve the pain points of traditional management methods. Therefore, developing a method and platform capable of accurate mold lifespan prediction, dynamic maintenance decision-making, and intelligent management throughout the entire process is of significant practical importance. Summary of the Invention

[0005] To overcome the shortcomings mentioned above, the present invention provides a technical solution that can solve the above problems.

[0006] A method for predicting and maintaining the lifespan of molds based on the Internet of Things includes the following steps:

[0007] Step 1: Data Acquisition: Through various types of IoT sensors deployed in key parts such as mold cavity, guide pillars and bushings, and ejection mechanism, real-time data on temperature, vibration, pressure, wear, number of mold opening and closing times, and running time are collected during the mold operation process;

[0008] Step 2: Data preprocessing: The collected raw data is processed by removing outliers, filling in missing values, and standardizing the data to obtain high-quality, effective data;

[0009] Step 3: Feature Extraction: Extract time-domain features, frequency-domain features, and trend features related to mold life from the effective data, and construct a mold life feature vector;

[0010] Step 4: Life Prediction Model Training and Inference: Based on historical failure data, mold design parameters and material performance parameters, a hybrid prediction model integrating a long short-term memory network (LSTM) and a fatigue damage mechanism model is constructed. The model is trained and optimized using training set data. The feature vector obtained in Step 3 is input into the trained model, and the predicted value of the remaining life of the mold is output.

[0011] Step 5: Dynamic maintenance strategy generation: Combining the remaining lifetime prediction, current production task priority, and maintenance resource allocation, a personalized maintenance plan is generated through a multi-objective optimization algorithm, including maintenance timing, maintenance content, and maintenance methods;

[0012] Step Six: Maintenance Execution and Feedback: Perform maintenance operations according to the maintenance plan, collect mold operation data after maintenance, and iteratively optimize the prediction model and maintenance strategy.

[0013] As a further aspect of the present invention: the IoT sensors in step one include a temperature sensor, a triaxial vibration sensor, a pressure sensor, a displacement sensor, and an infrared wear detection sensor. The sensor sampling frequency can be dynamically adjusted according to the running speed of the mold, with a range of 10-100Hz.

[0014] As a further aspect of the present invention: outlier removal in step two adopts a combination of box plot and isolated forest algorithm, missing value completion adopts interpolation algorithm based on spatiotemporal correlation, and data standardization adopts Z-score standardization method.

[0015] As a further aspect of the present invention: the time-domain features in step three include peak value, effective value, kurtosis, and skewness; the frequency-domain features are obtained through Fast Fourier Transform (FFT) and include characteristic frequency amplitude and spectral centroid; the trend features include parameter change rate and cumulative damage value.

[0016] As a further aspect of the present invention: the hybrid prediction model in step four optimizes the LSTM network through an attention mechanism to capture nonlinear dynamic features in the data, while combining a mechanism model based on Miner's linear cumulative damage theory to correct the prediction bias of the pure data-driven model. An adaptive learning rate algorithm is used during model training to improve the convergence speed.

[0017] An intelligent management platform for mold life prediction and maintenance based on the Internet of Things, characterized in that it includes:

[0018] The perception layer consists of various types of IoT sensors deployed on molds and production equipment, used to collect real-time mold operating status data and environmental data;

[0019] Network layer: including edge gateway, wireless communication module (5G, Wi-Fi, Bluetooth) and Ethernet module, used to transmit data collected by the perception layer to the platform layer, and also support downlink transmission of commands;

[0020] Platform layer: includes a data storage module, a data analysis module, a lifespan prediction module, a maintenance decision module, and a model management module; the data storage module uses a distributed database to store raw data, processed data, and historical records; the data analysis module performs data preprocessing and feature extraction; the lifespan prediction module deploys the hybrid prediction model described in claim 1; the maintenance decision module generates dynamic maintenance plans; and the model management module implements model updates, iterations, and version control.

[0021] Application layer: Includes a visual monitoring interface, maintenance task management module, alarm and early warning module, report statistics module, and user permission management module, which provide users with intuitive mold status display, maintenance task assignment, anomaly warning and data statistical analysis functions.

[0022] As a further aspect of the present invention, the perception layer also includes a sensor calibration module, which periodically performs self-calibration on the sensors to ensure data acquisition accuracy; the network layer adopts edge computing technology to complete some real-time data preprocessing and anomaly detection at the gateway, reducing data transmission latency and server load.

[0023] As a further aspect of the present invention: the application layer supports access from multiple terminals, including computer clients, mobile apps, and industrial control center displays, and also has data interfaces to connect with enterprise ERP systems and MES systems to achieve collaboration in production and mold management.

[0024] As a further aspect of the present invention: the alarm and early warning module adopts a graded early warning mechanism, which is divided into reminder level, warning level and emergency level according to the degree of deviation of the remaining life prediction value and real-time operating parameters from the threshold. The relevant personnel are notified through SMS, APP push and on-site sound and light alarm.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] 1. By adopting a hybrid model that integrates deep learning and mechanism analysis, the nonlinear characteristics in the data are captured, and the fatigue damage mechanism of the mold is combined, which effectively improves the accuracy of remaining life prediction, and the prediction error can be controlled within 5%.

[0027] 2. By generating personalized maintenance plans based on real-time operating data, remaining life predictions, and production plans, the drawbacks of fixed-cycle maintenance are avoided, over-maintenance and under-maintenance are reduced, and maintenance costs and production downtime losses are lowered.

[0028] 3. Real-time data collection and automatic transmission are achieved through the Internet of Things, and data analysis, prediction, decision-making and task assignment are completed in combination with the intelligent management platform, reducing manual intervention and improving management efficiency;

[0029] 4. The platform records all operational data, maintenance records, and fault information of the mold from its initial use to its scrapping, enabling full lifecycle traceability and analysis, and providing data support for mold design optimization and procurement decisions;

[0030] 5. By supporting the access of multiple types of sensors and various communication methods, it can be integrated with the enterprise's existing management system, making it suitable for mold management in different industries and of different types, and has broad application prospects.

[0031] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in this embodiment or the prior art, the drawings used in the description of the embodiment or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of the mold life prediction and maintenance method.

[0034] Figure 2 This is a structural diagram of the intelligent management platform. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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.

[0036] In the description of this invention, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0037] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0038] In the embodiments of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0039] Please see Figure 1-2 ;

[0040] Example 1: Injection mold for automotive parts production

[0041] For injection molds used in automotive parts, PT100 high-precision temperature sensors (measurement range 50-200℃, accuracy ±0.1℃) are deployed on the cavity sidewalls to monitor cavity temperature changes in real time during injection molding; triaxial vibration sensors (measurement range 0-50g, frequency range 1-1000Hz) are installed at key mating parts of guide pillars and bushings to capture vibration signals during mold opening and closing; pressure sensors (measurement range 0-200MPa, accuracy ±0.5%FS) are configured at the gate to monitor melt injection pressure; laser displacement sensors (measurement range 0-100mm, accuracy ±0.01mm) are installed at the ejection mechanism to record ejection stroke and repeatability accuracy; infrared wear detection sensors are also deployed to monitor the wear on the cavity surface in real time; all sensors are connected to a cloud-based intelligent management platform via an edge gateway supporting 5G and Ethernet dual-mode communication.

[0042] Based on the operating cycle of the injection mold (approximately 15 seconds per mold opening and closing), the sensor sampling frequency is set to 50Hz. Simultaneously, data on cavity temperature, guide pillar vibration acceleration, gate pressure, ejection displacement, wear, cumulative mold opening and closing times, and ambient temperature and humidity in the production workshop are collected. The edge gateway first preprocesses the raw data: using a box plot algorithm to remove abnormal jump data caused by electromagnetic interference from the temperature sensor, and using a linear interpolation method based on spatiotemporal correlation to complete the occasional missing data from the vibration sensor. Finally, the Z-score standardization method is used to uniformly convert data of different dimensions into standard normal distribution data before transmitting it to the cloud platform.

[0043] The cloud-based data analysis module extracts multi-dimensional features from the preprocessed data: in the time domain, the peak value (0.8g), effective value (0.3g), and kurtosis (3.2) of the vibration acceleration are selected; in the frequency domain, the characteristic frequency amplitudes at 100Hz and 200Hz are extracted using Fast Fourier Transform (FFT) (0.2g and 0.15g, respectively); in the trend domain, the cavity temperature change rate (0.5℃ / cycle) and the cumulative damage value (0.35) based on the fatigue limit of the mold material (Cr12MoV) are calculated to construct a complete feature vector; the life prediction module inputs the feature vector into a trained hybrid model (an attention-optimized LSTM network and a Miner cumulative damage mechanism model are fused with a weight of 0.6:0.4) and outputs a predicted remaining life of 1200 mold opening and closing cycles for the injection mold;

[0044] The maintenance decision module, combining the current production plan (300 parts remaining to be produced) and maintenance resource status (spare parts inventory includes guide pillars and bushings, and 2 maintenance personnel are idle), establishes a multi-objective optimization function (objectives: minimize downtime losses, minimize maintenance costs, and maximize mold utilization). The optimal maintenance plan is obtained through a genetic algorithm: once the current batch of products is completed (approximately 300 mold openings and closings), maintenance operations are immediately executed, including replacing worn guide pillars and bushings, cleaning residual material in the cavity, and calibrating the positioning accuracy of the ejector mechanism. The maintenance plan is pushed to designated maintenance personnel via the platform's mobile app, and operation data is recorded in real time during maintenance (e.g., replacement of spare parts model, and a calibrated positioning error of ±0.02mm).

[0045] After maintenance, the sensor data showed that the cavity temperature fluctuation range decreased from ±1.5℃ to ±0.8℃, the effective value of vibration acceleration decreased by 30%, and the mold operation stability was significantly improved. The platform fed back the maintenance records and performance comparison data before and after maintenance to the model management module, and dynamically adjusted the weight parameters of the hybrid prediction model (adjusted to LSTM network 0.62 and mechanism model 0.38) to further improve the prediction accuracy. Managers can monitor the mold operation status in real time through a visual interface, and the report statistics module automatically generates a monthly maintenance cost report.

[0046] Example 2: Stamping Die for Manufacturing Electronic Device Housings

[0047] For mobile phone casing stamping dies, piezoelectric pressure sensors (measurement range 0-100MPa, accuracy ±0.3% FS) are deployed at key stress-bearing parts of the punch and die to monitor contact pressure during the stamping process; a triaxial vibration sensor (measurement range 0-100g, frequency range 1-2000Hz) is installed on the die base to capture vibration signals generated by stamping impact; a displacement sensor (measurement range 0-50mm, accuracy ±0.005mm) is installed at the guide mechanism to monitor changes in guide clearance; Hall effect sensors are configured to record the number of stamping operations, and environmental sensors collect data on workshop temperature, humidity, and dust concentration; the sensors are connected to an edge gateway via a Wi-Fi 6 wireless communication module, and the gateway then connects to the cloud platform via Ethernet.

[0048] Based on the high-speed operation characteristics of the stamping die (a single stamping cycle is about 2 seconds), the sensor sampling frequency is set to 100Hz to collect punch pressure, base vibration acceleration, guide displacement, stamping number, and environmental parameters in real time. The edge gateway preprocesses the data: the isolated forest algorithm is used to detect and remove hidden outliers in the vibration data (120g peak data caused by equipment resonance), the missing data of the pressure sensor is completed by random forest interpolation, and the data is transmitted to the cloud after Z-score standardization.

[0049] The data analysis module extracts the following features: Time-domain features include peak pressure (85MPa), effective vibration acceleration (15g), and skewness (2.8); Frequency-domain features are extracted using FFT to obtain the frequency amplitudes at 500Hz and 800Hz (10g and 8g, respectively); Trend features are calculated to determine the guide clearance change rate (0.002mm / cycle) and the cumulative damage value (0.42) based on the fatigue theory of SKD11 material; The feature vectors are input into the hybrid prediction model, and the output mold remaining life prediction value is 800 stamping cycles.

[0050] Based on the current production task (an urgent order requiring 500 stamping cycles) and maintenance resources (spare parts available in the die warehouse, and one maintenance personnel available after 2 hours), the maintenance decision module generates a dynamic maintenance plan: After completing the urgent order (consuming 500 stamping cycles), a maintenance personnel will be assigned to maintain the die, including replacing worn dies, adjusting the guide mechanism clearance, cleaning dust from the die surface, and checking the torque of fastening bolts; the maintenance process will be remotely monitored through the platform's video monitoring function, and relevant parameters will be recorded after maintenance (guide clearance adjusted to 0.01mm, bolt torque reaching the preset standard of 35N*m).

[0051] Post-maintenance data showed that the dimensional tolerance pass rate of stamped parts improved from 96% to 99.5%, and the peak vibration acceleration decreased by 40%. The platform fed back the relevant data to the model management module to iteratively optimize the prediction model, keeping the subsequent remaining life prediction error within 3%. The alarm and early warning module detected a sudden increase in pressure to 98MPa (exceeding the warning threshold of 90MPa) during a stamping process, immediately triggering a warning-level alert and notifying maintenance personnel via SMS and APP. Upon inspection, it was found that the error was caused by a deviation in the feeding position, which was promptly adjusted to prevent mold damage.

[0052] Example 3: Die-casting mold for manufacturing home appliance casings

[0053] For the die-casting mold of the washing machine inner tub, K-type thermocouple temperature sensors (measuring range 0-800℃, accuracy ±1℃) are deployed around the cavity to monitor the cavity temperature during the die-casting process; pressure sensors (measuring range 0-500MPa, accuracy ±0.5%FS) are installed at the contact point between the die-casting rod and the mold to collect die-casting pressure data; flow sensors (measuring range 0-50L / min, accuracy ±0.2L / min) are installed at the mold cooling water outlet to monitor the cooling effect; ultrasonic wear sensors are also deployed to detect the wear of the inner wall of the cavity, and a counter records the number of die-casting cycles; the sensors communicate with the edge gateway via Bluetooth, and the gateway uses a 5G network to upload the data to the cloud intelligent management platform;

[0054] The die-casting mold has a single production cycle of approximately 30 seconds. The sensor sampling frequency is set to 30Hz, simultaneously collecting cavity temperature, die-casting pressure, cooling water flow rate, wear amount, number of die-casting cycles, and workshop environmental parameters. The edge gateway preprocesses the data: it uses a combination of box plot and isolated forest algorithms to remove abnormal data of cooling water flow rate (0L / min data caused by pipe blockage), completes the missing data of the temperature sensor using time series-based interpolation, and transmits the data to the cloud after Z-score standardization.

[0055] The data analysis module extracts the following features: Time-domain features include peak die-casting pressure (420MPa), effective temperature (650℃), and kurtosis (3.5); Frequency-domain features are extracted using FFT to obtain the frequency amplitude at 300Hz and 600Hz (5MPa and 3MPa, respectively); Trend features are calculated to determine the cavity temperature change rate (5℃ / cycle), cooling water flow rate decay rate (0.5L / min*cycle), and cumulative damage value based on the fatigue limit of H13 mold steel (0.48); The feature vectors are input into the hybrid prediction model, and the predicted remaining mold life is 600 die-cast cycles.

[0056] Based on the production plan (400 die-casting cycles remaining in the regular order) and maintenance resources (cooling pipe cleaning equipment is available, and spare seals are available in the spare parts warehouse), the maintenance decision module generates a maintenance plan: After completing the current regular order (consuming 400 die-casting cycles), a comprehensive maintenance of the mold will be performed, including cleaning the cooling water system, replacing aged seals, repairing minor wear in the cavity, and calibrating the die-casting pressure control accuracy. The maintenance plan is dispatched to the maintenance team via the platform PC. After maintenance is completed, the maintenance personnel upload a maintenance report and test data (such as cooling water flow rate restored to 45L / min, die-casting pressure fluctuation range ±5MPa) to the platform.

[0057] After maintenance, mold operation data showed that: cavity cooling uniformity improved, shrinkage cavity defect rate of die castings decreased from 3% to 0.5%, and production efficiency increased by 15%; the platform fed back maintenance data to the model management module to update the training dataset of the prediction model, further improving the accuracy of the model's life prediction for this type of die casting mold; at the same time, the platform connected with the enterprise ERP system through the data interface to synchronize the mold maintenance plan to the production planning module, realizing coordinated scheduling of production and maintenance, and avoiding production disconnect caused by maintenance; the quarterly analysis report generated by the report statistics module showed that...

[0058] The circuits, electronic components, and control modules involved are all existing technologies, which can be fully implemented by those skilled in the art, and need not be elaborated upon. The scope of protection of this invention does not involve any improvement to the software and methods.

[0059] It should also be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0060] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting and maintaining the lifespan of a mold based on the Internet of Things, characterized in that: Includes the following steps: Step 1: Data Acquisition (101): Through various types of IoT sensors deployed in key parts such as mold cavity, guide pillars and bushings, and ejection mechanism, real-time data on temperature, vibration, pressure, wear, number of mold opening and closing times, and running time during mold operation are collected. Step 2: Data preprocessing (102): The collected raw data is subjected to outlier removal, missing value completion, and data standardization to obtain high-quality effective data; Step 3: Feature extraction (103): Extract time-domain features, frequency-domain features and trend features related to mold life from the effective data, and construct mold life feature vector; Step 4: Life prediction model training and inference (104): Based on historical fault data, mold design parameters and material performance parameters, a hybrid prediction model integrating long short-term memory network (LSTM) and fatigue damage mechanism model is constructed. The model is trained and optimized using training set data. The feature vector obtained in step 3 is input into the trained model, and the predicted value of the remaining life of the mold is output. Step 5: Dynamic maintenance strategy generation (105): Combining the remaining lifetime prediction, current production task priority and maintenance resource allocation, a personalized maintenance plan is generated through a multi-objective optimization algorithm, including maintenance timing, maintenance content and maintenance method; Step 6: Maintenance Execution and Feedback (106): Perform maintenance operations according to the maintenance plan, collect mold operation data after maintenance, and iteratively optimize the prediction model and maintenance strategy.

2. The method for predicting and maintaining mold life based on the Internet of Things according to claim 1, characterized in that: The IoT sensors in step one include temperature sensors, triaxial vibration sensors, pressure sensors, displacement sensors, and infrared wear detection sensors. The sensor sampling frequency can be dynamically adjusted according to the mold running speed, ranging from 10 to 100 Hz.

3. The method for predicting and maintaining mold life based on the Internet of Things according to claim 1, characterized in that: In step two, outlier removal uses a combination of box plot and isolated forest algorithms, missing value completion uses an interpolation algorithm based on spatiotemporal correlation, and data standardization uses the Z-score standardization method.

4. The method for predicting and maintaining mold life based on the Internet of Things according to claim 1, characterized in that: The time-domain features in step three include peak value, RMS value, kurtosis, and skewness. The frequency-domain features are obtained through Fast Fourier Transform (FFT) and include characteristic frequency amplitude and spectral centroid. The trend features include parameter change rate and cumulative damage value.

5. The method for predicting and maintaining mold life based on the Internet of Things according to claim 1, characterized in that: In step four, the hybrid prediction model optimizes the LSTM network through an attention mechanism to capture nonlinear dynamic features in the data. At the same time, it combines a mechanism model based on Miner's linear cumulative damage theory to correct the prediction bias of the pure data-driven model. An adaptive learning rate algorithm is used during model training to improve the convergence speed.

6. An intelligent management platform for the IoT-based mold life prediction and maintenance method according to claims 1-5, characterized in that: include: Perception layer (201): Composed of various types of IoT sensors deployed on molds and production equipment, used to collect mold operating status data and environmental data in real time; Network layer (202): includes edge gateway, wireless communication module (5G, Wi-Fi, Bluetooth) and Ethernet module, used to transmit data collected by the perception layer to the platform layer, and also supports downlink transmission of commands; Platform layer (203): includes a data storage module, a data analysis module, a lifespan prediction module, a maintenance decision module, and a model management module; the data storage module uses a distributed database to store raw data, processed data, and historical records; the data analysis module implements data preprocessing and feature extraction; the lifespan prediction module deploys the hybrid prediction model described in claim 1; the maintenance decision module generates dynamic maintenance plans; and the model management module implements model updates, iterations, and version control. Application layer (204): includes a visual monitoring interface, maintenance task management module, alarm and early warning module, report statistics module and user permission management module, which are used to provide users with intuitive mold status display, maintenance task assignment, anomaly warning and data statistical analysis functions.

7. The intelligent management platform according to claim 6, characterized in that: The perception layer (201) also includes a sensor calibration module, which periodically performs self-calibration on the sensors to ensure data acquisition accuracy; the network layer (202) adopts edge computing technology to complete some real-time data preprocessing and anomaly detection at the gateway, reducing data transmission latency and server load.

8. The intelligent management platform according to claim 6, characterized in that: The application layer (204) supports access from multiple terminals, including computer clients, mobile apps and industrial control center displays. It also has data interfaces to connect with enterprise ERP systems and MES systems to achieve collaboration in production and mold management.

9. The intelligent management platform according to claim 6, characterized in that: The alarm and early warning module adopts a graded early warning mechanism, which is divided into reminder level, warning level and emergency level according to the degree of deviation of the remaining life prediction value and real-time operating parameters from the threshold. It notifies relevant personnel through SMS, APP push and on-site audible and visual alarms.

Images