Method, system and equipment for monitoring state in injection mold cavity and medium

Abstract

The invention relates to the technical field of injection mold monitoring, and discloses an injection mold cavity state monitoring method, system, equipment and medium, the injection mold cavity state monitoring method comprises the following steps: multiple sensors collect mold state data in real time, and the mold state data are preprocessed and fused to generate a comprehensive data set; training a machine learning model based on historical samples, and updating model parameters in real time through an incremental learning algorithm; analyzing current data by adopting dynamic fuzzy logic reasoning, and outputting a state evaluation result; real-time data and results are uploaded to the cloud for deep analysis and distributed storage; a user feedback mechanism is integrated, and model parameters and a fuzzy inference rule base are optimized. According to the method, model parameters are updated in real time through an incremental learning algorithm, a multi-parameter nonlinear coupling relation is analyzed in combination with dynamic fuzzy logic reasoning, and based on cloud collaboration and a user feedback closed-loop mechanism, self-adaptive monitoring and continuous optimization of the injection mold state are achieved, and the anomaly detection precision and the system robustness are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of injection mold monitoring, and in particular to a method, system, equipment and medium for monitoring the state in an injection mold cavity. Background Art

[0002] With the rapid development of injection molding processes toward high-precision, multi-material, and complex structures, real-time status monitoring within the mold cavity has become a core requirement for ensuring product quality and production stability. During the injection molding process, mold conditions are dynamically affected by multiple factors, including material properties, equipment aging, and environmental fluctuations, resulting in strong nonlinear correlations between parameters. Traditional technologies struggle to effectively capture these dynamic changes and complex correlations. There is an urgent need for an intelligent monitoring method that can adapt to process drift, analyze nonlinear characteristics, and support closed-loop optimization.

[0003] Existing technologies for monitoring injection mold conditions primarily rely on single-parameter alarm mechanisms based on fixed thresholds or static machine learning models trained offline. The former relies on manually set upper and lower thresholds for single parameters like temperature and pressure to detect anomalies; the latter uses historical data to train classification models (such as support vector machines and random forests), which are then solidified and deployed locally. Furthermore, data collection and analysis are often limited to single-machine systems, relying on local storage and computing resources for basic monitoring functions.

[0004] However, existing technologies still have some shortcomings: First, due to the lack of a real-time parameter update mechanism, static models cannot adapt to the dynamic changes in the injection molding process (such as material replacement or equipment aging), resulting in a significant increase in the misjudgment rate when the working conditions shift; second, the fixed threshold method and linear model have difficulty analyzing the nonlinear coupling relationship between multiple parameters such as temperature and pressure, and the detection sensitivity of complex abnormal patterns (such as a slow drop in pressure accompanied by abnormal temperature fluctuations) is insufficient; in addition, the single-machine system architecture causes data isolation, and the collaborative analysis and process optimization between multiple devices and multiple batches lack global data support; finally, existing technologies have not established a closed-loop mechanism for user feedback and system optimization. Model iteration relies on offline manual intervention and cannot integrate real-time operating experience to achieve adaptive optimization. The root cause of these problems lies in the systematic lack of dynamics, nonlinearity, collaboration and self-optimization capabilities in existing technologies, which seriously restricts the reliability and long-term applicability of the monitoring system. Summary of the Invention

[0005] The purpose of the present invention is to provide a method, system, equipment and medium for monitoring the condition inside the injection mold cavity, which solves the problems in the prior art that static models cannot adapt to dynamic working conditions, fixed threshold methods are difficult to analyze multi-parameter nonlinear coupling relationships, single-machine systems lack the ability to collaboratively analyze multi-source data, and open-loop mechanisms cause insufficient self-optimization.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for monitoring the state of an injection mold cavity, comprising the following steps: S1. Use multiple sensors to collect the status data of the injection mold in real time, and pre-process and fuse the status data to generate a comprehensive data set; S2. Using historical normal and abnormal state samples to train a machine learning model, enabling the model to evaluate mold status and detect anomalies based on the input comprehensive data set, and updating the model in real time based on an incremental learning algorithm when new status data arrives; S3. Analyze the current state data using dynamic fuzzy logic reasoning mechanism to generate mold state evaluation results; S4, upload real-time monitoring data and results to the cloud for in-depth analysis and storage; S5. Provide a user feedback mechanism to collect operator feedback information to optimize the model and inference rules.

[0007] Preferably, the S1 includes: The temperature, pressure, flow and humidity data of the injection mold are collected through temperature sensors, pressure sensors, flow sensors and humidity sensors; De-noising is performed on the real-time collected temperature, pressure, flow and humidity data to remove noise data caused by environmental factors; Normalize the denoised data to make the data value range uniform. The formula used is: Among them, D processed (t) is the processed data; D original (t) is the original data; μ is the data mean; σ is the data standard deviation; the processed data are aligned in time series to form a unified comprehensive data set.

[0008] Preferably, the S2 includes: Collect preprocessed historical normal state samples and abnormal state samples to build a complete training dataset; The support vector machine algorithm is used for model training. The model input is a comprehensive data set, and the output is the evaluation result of the mold status. When new status data arrives, the incremental learning algorithm is used to update the model parameters. The update formula is: Among them, θ t is the current model parameter; α is the learning rate; is the gradient of the loss function; D new For the newly arrived dataset; The cross-validation method is used to evaluate the performance of the trained model. The historical data is divided into a training set and a validation set, and the detection accuracy of the model on normal and abnormal state samples is recorded.

[0009] Preferably, the S3 includes: Through the dynamic fuzzy logic reasoning mechanism, the current state data is fuzzified to form a fuzzy set; Generate fuzzy rules, the fuzzy reasoning form is: IF(X1 IS A1)∧(X2 IS A2)THENYIS B; Among them, X1, X2 are input variables; A1, A2 are fuzzy sets; Y is the output variable; B is the output fuzzy set; IF and H are membership functions; E is the membership threshold; N is the number of rules; The mold condition evaluation results are generated based on fuzzy logic, and the output formula used is: Among them, w i is the weight of each fuzzy rule; B i is the output of the fuzzy rule; Y output is the final output variable; n is the number of fuzzy rules.

[0010] Preferably, the step S3 further includes a comparison process of the mold state and the abnormal pattern, comprising the following steps: Compare the generated mold status assessment results with the preset abnormal pattern library to identify whether there is an abnormal state; Generate alarm information based on the comparison results and record abnormal status characteristics to provide data support for subsequent analysis.

[0011] Preferably, the S4 includes: Upload real-time monitoring data and status assessment results to the cloud server via secure network protocols; Data aggregation and storage are performed in the cloud, supporting multi-source data integration and in-depth analysis.

[0012] Preferably, the S5 includes: A user-friendly feedback interface is designed to allow operators to provide real-time feedback on any alarms and monitoring results given by the system. The collected feedback information is used to update fuzzy rules and optimize the incremental learning model. The feedback information update formula is: Among them, F j is user feedback data; N1 represents the total amount of user feedback data collected; Update feedback Indicates the updated value calculated based on user feedback.

[0013] The present invention further provides an injection mold cavity state monitoring system, which uses the above-mentioned injection mold cavity state monitoring method, comprising: Sensor module, used to collect status data of injection mold in real time; A data processing module, configured to pre-process and fuse the status data, and perform mold status assessment and anomaly detection using a machine learning model; Dynamic fuzzy logic reasoning module, used to analyze current status data and generate mold status evaluation results; Cloud storage module, used to upload real-time monitoring data and assessment results to the cloud for data aggregation, storage and in-depth analysis; The user feedback module is used to collect operator feedback information and optimize the model and inference rules accordingly.

[0014] The present invention further provides an injection mold cavity state monitoring device, and the above-mentioned injection mold cavity state monitoring method includes: The sensor assembly, including temperature sensor, pressure sensor, flow sensor and humidity sensor, monitors the status of the injection mold in real time; A data collector, used to receive the status data provided by the sensor and perform preliminary processing and integration of the data; The processing unit analyzes the collected data based on the incremental learning algorithm and fuzzy logic reasoning mechanism to realize the mold status evaluation and anomaly detection functions; Cloud computing platform, used for data storage and analysis, supporting the aggregation and in-depth mining of multi-source data; User interface, used to display monitoring results and alarm information, allowing operators to provide real-time feedback.

[0015] The present invention also provides a storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned method for monitoring the state in the cavity of an injection mold.

[0016] In summary, the present invention includes at least one of the following beneficial technical effects: 1. The real-time parameter update mechanism based on the incremental learning algorithm enables the machine learning model to continuously adapt to changes in injection molding process parameters, avoiding the performance degradation problem caused by operating condition drift in traditional static models and ensuring the long-term stable operation of the monitoring system.

[0017] 2. Through the dynamic fuzzy logic reasoning mechanism, expert experience is combined with real-time data to build an interpretable fuzzy rule base, which effectively handles the nonlinear relationship of multiple parameters such as temperature and pressure coupled in the injection molding process, and solves the limitations of traditional threshold judgment methods in complex abnormal pattern recognition.

[0018] 3. Leveraging the cloud platform's data storage and in-depth analysis capabilities, integrated management of monitoring data from multiple devices and cross-device pattern mining are achieved, providing a global perspective for process optimization. Furthermore, standardized interfaces support remote access and collaborative maintenance, enhancing the digital management of the factory.

[0019] 4. Introducing a user feedback mechanism to form a "monitoring-feedback-optimization" closed loop enables the system to dynamically adjust the model and rule base based on human experience, taking into account the advantages of both automation and human intervention, and gradually improving the accuracy of anomaly detection and the credibility of system decisions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the module architecture of the present invention; Figure 3 It is a schematic diagram of the device structure of the present invention; Figure 4 Schematic diagram of the computer device structure of the present invention. DETAILED DESCRIPTION

[0021] The following is combined with Figure 1 -Attached Figure 4 , the present invention is described in further detail.

[0022] The present invention provides a method for monitoring the intracavity status of an injection mold. By integrating a monitoring solution with dynamic model updating, fuzzy logic reasoning, cloud collaboration, and a user feedback closed loop, the method can achieve real-time monitoring and adaptive optimization of the intracavity status of an injection mold with high robustness and strong interpretability.

[0023] like Figure 1 As shown, the method for monitoring the state in the injection mold cavity may include the following steps: S1. Use multiple sensors to collect the status data of the injection mold in real time, and pre-process and fuse the status data to generate a comprehensive data set; In this embodiment, the data collection and fusion steps of the injection mold cavity condition monitoring method are implemented using the following technical solution: First, multiple sensors are placed within the injection mold cavity to collect mold condition data in real time. Preferably, the sensors include temperature sensors, pressure sensors, flow sensors, and humidity sensors, respectively used to obtain temperature, pressure, melt flow rate, and ambient humidity data during mold operation. The sensors periodically collect data at a preset sampling frequency to ensure the continuity and integrity of the time series.

[0024] The collected raw data is preprocessed to eliminate noise and unify its dimensions. Specifically, preprocessing includes denoising and normalization. During the denoising phase, a sliding window median filter algorithm is used to smooth the raw signal. This algorithm replaces the current data point with the median value of the data within the window, effectively suppressing the interference of high-frequency noise on subsequent analysis. Preferably, the length of the sliding window is dynamically adjusted based on the sensor sampling frequency to ensure a balance between noise filtering effectiveness and data real-time performance.

[0025] In the normalization stage, the denoised data is mapped to a uniform numerical range to eliminate the impact of different sensor dimension differences on data fusion. The normalization formula is: Among them, D processed (t) is the processed data; D original (t) represents the raw data; μ represents the data mean; and σ represents the data standard deviation. The mean and standard deviation are obtained by statistically analyzing the mold's long-term operating data under normal operating conditions to ensure the rationality and applicability of the normalization parameters. This step converts the different sensor data into dimensionless relative values, facilitating subsequent feature extraction and standardization of model inputs.

[0026] After preprocessing, the multi-source sensor data is time-series aligned and fused. Specifically, based on the sampling timestamps of each sensor, an interpolation algorithm is used to align data with different frequencies or transmission delays to the same time base, forming a time-synchronized, integrated dataset. Preferably, the interpolation algorithm uses linear or spline interpolation to ensure smooth data alignment. This integrated dataset contains synchronized time series data for temperature, pressure, flow, and humidity parameters, providing a unified input for subsequent condition assessment and anomaly detection.

[0027] Furthermore, the data fusion step includes a mechanism for rejecting abnormal data. When a sensor's real-time data exceeds a preset acceptable range (for example, if the temperature exceeds the mold material's tolerance threshold), the system automatically marks the data as an outlier and performs local interpolation based on data from adjacent time points to prevent the abnormal data from interfering with subsequent analysis. Preferably, the preset acceptable range is dynamically adjusted based on mold design parameters and historical operating data to accommodate data fluctuations under different operating conditions.

[0028] S2. Use historical normal and abnormal state samples to train a machine learning model, enabling the model to evaluate mold status and detect anomalies based on the input comprehensive data set, and update the model in real time based on the incremental learning algorithm when new status data arrives; In this embodiment, the incremental learning and model updating steps of the injection mold cavity state monitoring method are implemented by the following technical solutions: First, an initial machine learning model is constructed based on the preprocessed comprehensive data set. The training data of the initial model includes historical normal state samples and abnormal state samples, wherein the normal state samples cover the operation data of the mold under standard process parameters (for example, temperature and pressure fluctuations are within the preset threshold range), and the abnormal state samples contain known failure mode data (such as pressure drop, temperature exceeding the limit, etc.). Preferably, the model adopts the support vector machine (SVM) algorithm, and maps low-dimensional nonlinear data to high-dimensional space through the kernel function to achieve effective classification of normal and abnormal states.

[0029] During the model training phase, the input is a multidimensional feature vector from a comprehensive dataset (e.g., time series data of temperature, pressure, flow, and humidity), and the output is a binary classification label (normal / abnormal). The model training goal is to minimize the classification error, and the model parameters are optimized through cross-validation to avoid overfitting or underfitting. Preferably, cross-validation uses a K-fold strategy, dividing the dataset into a training set and a validation set. After multiple iterations, the model parameters with the highest validation accuracy are selected as the final configuration.

[0030] When new data arrives, the model is dynamically updated using an incremental learning algorithm. The core of incremental learning is to adjust model parameters based only on new data without retraining all historical data, thereby significantly reducing computing resource consumption and improving real-time performance. Specifically, the model parameter update formula is: Among them, θ t is the current model parameter; α is the learning rate; is the gradient of the loss function; D new For the newly arrived dataset.

[0031] Through the gradient descent method, the model gradually adjusts the parameters in the direction of reducing the error, thereby adapting to changes in the mold state (such as material replacement, process adjustment, etc.).

[0032] Furthermore, the incremental learning process includes an online validation mechanism. After each model update, the system automatically extracts a portion of new data as a validation set and calculates the classification accuracy and recall of the updated model. If the performance indicators fall below a preset threshold, a model rollback or reinitialization process is triggered. Preferably, the validation mechanism uses a sliding window strategy, retaining only the most recent N batches of data for validation to reflect the model's performance under the latest operating conditions.

[0033] Through the above technical solutions, this embodiment achieves dynamic optimization and real-time adaptability of model parameters.

[0034] S3. Analyze the current state data using a dynamic fuzzy logic reasoning mechanism to generate a mold state evaluation result; In this embodiment, the dynamic fuzzy logic reasoning step of the injection mold cavity state monitoring method is implemented by the following technical solution: First, construct a dynamic fuzzy logic reasoning mechanism based on the preprocessed comprehensive data set. The core of fuzzy logic reasoning is to convert continuous numerical sensor data into fuzzy semantic variables, and perform state evaluation through a fuzzy rule base generated by expert experience and historical data. Specifically, fuzzification processing maps physical quantities such as temperature and pressure into fuzzy sets such as "low temperature", "normal", "high temperature", "low pressure", and "high pressure", and each fuzzy set describes its semantic boundary through a membership function. Preferably, the membership function uses a Gaussian function or a triangular function to balance the accuracy of fuzzy division and computational efficiency.

[0035] The construction of fuzzy rules is based on process experience and historical data analysis results. For example, when the temperature is detected to be in the "high temperature" range and the pressure is in the "low pressure" range, the inference conclusion of "high probability of abnormal state" is triggered. The form of fuzzy rules is: IF(X1 IS A1)∧(X2 IS A2)THENYIS B; Where X1 and X2 are input variables; A1 and A2 are fuzzy sets; Y is the output variable; B is the output fuzzy set; IF and H are membership functions; E is the membership threshold; and N is the number of rules. The rule base is dynamically updated and linked to incremental learning results to ensure that the rules are adaptable to the current working conditions.

[0036] In the fuzzy inference process, the activation strength of each rule is first calculated. The activation strength is obtained by taking the minimum (MIN) or product operation of the membership values of the input variables to each fuzzy set. For example, if the current temperature has a membership of 0.8 to "high temperature" and the pressure has a membership of 0.7 to "low pressure", then the rule activation strength is w i =min(0.8,0.7)=0.7. Preferably, the calculation method of activation strength can be selected according to actual needs to balance logical rigor and flexibility.

[0037] Subsequently, the fuzzy inference results are defuzzified by weighted average method to generate the final numerical output. The output calculation formula is: Among them, w i is the weight of each fuzzy rule; B i is the output of the fuzzy rule; Y output is the final output variable; n is the number of fuzzy rules. Through this step, the fuzzy semantic conclusion is converted into a specific abnormal probability value, which facilitates subsequent threshold determination and alarm triggering.

[0038] Furthermore, the fuzzy logic reasoning mechanism supports dynamic rule base updates. When user feedback or the incremental learning model detects new abnormal patterns, the system automatically generates or adjusts fuzzy rules. For example, if multiple feedbacks show that a certain temperature-pressure combination is misjudged as normal, a new rule is added: IF (TIS specific interval) ∧ (PIS steep drop) THEN the abnormal probability IS is high. Preferably, rule updates are achieved through weight adjustment, where the weight of the old rule decreases as the number of feedbacks increases, and the weight of the new rule gradually increases, thereby achieving logical optimization without interrupting system operation.

[0039] Through the above technical solution, this embodiment combines the semantic expression advantages of fuzzy logic with its dynamic update capabilities, effectively dealing with the problems of multi-parameter coupling and complex nonlinear relationships in the injection molding process, and providing a reliable theoretical framework for accurately evaluating the mold status.

[0040] S4, upload real-time monitoring data and results to the cloud for in-depth analysis and storage; In this embodiment, the cloud-based monitoring and data sharing steps of the injection mold cavity condition monitoring method are implemented through the following technical solutions: First, the real-time collected mold condition data, the preprocessed comprehensive data set, and the condition assessment results generated by fuzzy logic reasoning are uploaded to the cloud server. Data transmission is completed through an encrypted communication protocol. Preferably, HTTPS communication based on TLS (Transport Layer Security) is adopted to ensure the integrity and confidentiality of the data during transmission. The uploaded data includes but is not limited to time series data of temperature, pressure, flow rate, humidity parameters, parameter sets after model update, and anomaly detection result labels, forming a multidimensional data stream for further processing in the cloud.

[0041] On the cloud server side, a hierarchical data storage architecture is established. Raw data, preprocessed data, and analysis results are stored by data type and timestamp. Preferably, a time series database is used to store real-time sensor data, indexed by time as the primary key, to support efficient time range queries and aggregate calculations. A structured database stores model parameters, fuzzy rules, and user feedback information for fast retrieval and updating. The storage solution optimizes storage efficiency through data partitioning strategies (such as dividing data tables by hour or date), reduces the amount of data in a single table, and improves query response speed.

[0042] Furthermore, the cloud platform conducts in-depth analysis of the uploaded data, including abnormal pattern mining and multi-source data association analysis. Abnormal pattern mining is based on the comparison of historical data and real-time data. It identifies potential abnormal pattern clusters through clustering algorithms (such as DBSCAN) and dynamically adds newly discovered patterns to the abnormal pattern library. Preferably, the neighborhood radius parameter of the clustering algorithm is adaptively adjusted according to the data distribution density to avoid subjective bias caused by manual intervention. Multi-source data association analysis reveals the coupling relationship between parameters such as temperature and pressure by constructing a correlation matrix between parameters (such as the Pearson correlation coefficient matrix), providing data support for fuzzy rule optimization.

[0043] In addition, the cloud platform provides a data sharing interface that supports multi-device access. Designed based on a RESTful API, the interface allows authorized users or systems to access monitoring data or analysis reports for a specific time period through standardized requests. Preferably, interface access permissions are managed via the OAuth 2.0 protocol to ensure data security and operational traceability.

[0044] The cloud architecture is designed with both real-time and scalability in mind, and can adapt to data processing needs in large-scale distributed deployment scenarios.

[0045] S5. Provide a user feedback mechanism to collect operator feedback information to optimize the model and inference rules; In this embodiment, the user feedback mechanism of the injection mold cavity state monitoring method is implemented by the following technical solutions: First, a user-friendly interactive interface is designed to receive real-time feedback from operators on the system output. The interface is integrated into the local management terminal or remote mobile device to display real-time monitoring data, status assessment results and abnormal alarm information in a visual form. Preferably, the interface provides click-through feedback options (such as "Confirm Alarm" and "Mark False Alarm" buttons) to allow operators to quickly verify or correct abnormal events determined by the system. Feedback information includes but is not limited to the timestamp of the abnormal event, the correction label and additional annotation text, forming a structured feedback data set.

[0046] Collected user feedback data is transmitted to the data processing module via an encrypted channel and stored in association with real-time monitoring data. This feedback data is mapped to the corresponding model inference results and raw sensor data to ensure traceability for subsequent analysis. Preferably, data association is based on a dual index of timestamp and event ID to avoid data misalignment due to timing deviations.

[0047] Furthermore, feedback data is used to dynamically optimize the incremental learning model parameters and fuzzy logic rule base. Specifically, model optimization is achieved by adjusting the weights of the incremental learning algorithm's loss function. For example, if a certain type of abnormal event is repeatedly marked as a false positive, the weight of the corresponding data in the loss function is reduced, making the model less sensitive to this type of false positive pattern when updating. The weight adjustment formula is: Among them, F j is user feedback data; N1 represents the total amount of user feedback data collected; Update feedback Indicates the updated value calculated based on the user's feedback information; Through this mechanism, user feedback is directly involved in the model training process, ensuring the consistency of system decisions with human experience.

[0048] In terms of fuzzy logic rule optimization, feedback data triggers dynamic adjustments to the rule base. For example, if an abnormal event corresponding to a fuzzy rule is frequently flagged as a false positive, the system automatically reduces the activation weight of that rule or directly deletes redundant rules. Preferably, rule weight adjustments are based on the statistical distribution of feedback data, using a sliding window strategy to calculate the cumulative impact of recent feedback, thus avoiding fluctuations in the rule base caused by occasional feedback.

[0049] Furthermore, the feedback mechanism supports the addition of custom rules. Operators can enter new anomaly determination criteria (such as a specific temperature-pressure combination) through the interface. The system converts these into fuzzy rules and adds them to the rule base. The initial weights of the newly added rules are dynamically set based on the confidence level of the feedback data, and their effectiveness is subsequently verified through incremental learning.

[0050] Through the above technical solution, this embodiment deeply integrates human experience with the automated monitoring system to form a closed-loop optimization mechanism, which not only improves the system's adaptability to complex working conditions, but also enhances the accuracy of anomaly detection through continuous learning, providing a guarantee for the long-term stable operation of the injection molding process.

[0051] The injection mold cavity state monitoring system described below and the injection mold cavity state monitoring method described above can refer to each other.

[0052] Please see the attached Figure 2 The present invention further provides an injection mold cavity state monitoring system, which uses the above-mentioned injection mold cavity state monitoring method, including: Sensor module, used to collect status data of injection mold in real time; Data processing module, used to pre-process and fuse status data, and use machine learning models to perform mold status assessment and anomaly detection; Dynamic fuzzy logic reasoning module, used to analyze current status data and generate mold status evaluation results; Cloud storage module, used to upload real-time monitoring data and assessment results to the cloud for data aggregation, storage and in-depth analysis; The user feedback module is used to collect operator feedback information and optimize the model and inference rules accordingly.

[0053] The system of this embodiment can be used to execute the above method embodiments, and its principles and technical effects are similar, so they will not be repeated here.

[0054] The following description of an injection mold cavity state monitoring device and the above description of an injection mold cavity state monitoring method may refer to each other.

[0055] Please see the attached Figure 3 The present invention further provides an injection mold cavity state monitoring device, using the above-mentioned injection mold cavity state monitoring method, comprising: The sensor assembly, including temperature sensor, pressure sensor, flow sensor and humidity sensor, monitors the status of the injection mold in real time; Data collector, used to receive status data provided by sensors and perform preliminary processing and fusion of the data; The processing unit analyzes the collected data based on the incremental learning algorithm and fuzzy logic reasoning mechanism to realize the mold status evaluation and anomaly detection functions; Cloud computing platform, used for data storage and analysis, supporting the aggregation and in-depth mining of multi-source data; User interface, used to display monitoring results and alarm information, allowing operators to provide real-time feedback.

[0056] The device of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, so they will not be repeated here.

[0057] The storage medium described below and the method for monitoring the state in the injection mold cavity described above can refer to each other.

[0058] Please see the attached Figure 4 The present invention also provides a storage medium having a computer program stored thereon, which implements the above-mentioned method for monitoring the state in the injection mold cavity when executed by a processor.

[0059] The storage medium of this embodiment can be used to execute the above method embodiment, and its principles and technical effects are similar, which will not be described here in detail.

[0060] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for monitoring the state of an injection mold cavity, characterized in that: The following steps are involved: S1. Use multiple sensors to collect the status data of the injection mold in real time, and pre-process and fuse the status data to generate a comprehensive data set; S2. Using historical normal and abnormal state samples to train a machine learning model, enabling the model to evaluate mold status and detect anomalies based on the input comprehensive data set, and updating the model in real time based on an incremental learning algorithm when new status data arrives; S3. Analyze the current state data using dynamic fuzzy logic reasoning mechanism to generate mold state evaluation results; S4, upload real-time monitoring data and results to the cloud for in-depth analysis and storage; S5. Provide a user feedback mechanism to collect operator feedback information to optimize the model and inference rules.

2. The method for monitoring the state of an injection mold cavity according to claim 1, characterized in that: Said S1 comprises: The temperature, pressure, flow and humidity data of the injection mold are collected through temperature sensors, pressure sensors, flow sensors and humidity sensors; De-noising is performed on the real-time collected temperature, pressure, flow and humidity data to remove noise data caused by environmental factors; Normalize the denoised data to make the data value range uniform. The formula used is: Among them, D processed (t) is the processed data; D original (t) is the original data; μ is the data mean; σ is the data standard deviation; the processed data are aligned in time series to form a unified comprehensive data set.

3. The method for monitoring the state of an injection mold cavity according to claim 1, characterized in that: The S2 includes: Collect preprocessed historical normal state samples and abnormal state samples to build a complete training dataset; The support vector machine algorithm is used for model training. The model input is a comprehensive data set, and the output is the evaluation result of the mold status. When new status data arrives, the incremental learning algorithm is used to update the model parameters. The update formula is: Among them, θ t is the current model parameter; α is the learning rate; is the gradient of the loss function; D new For the newly arrived dataset; The cross-validation method is used to evaluate the performance of the trained model. The historical data is divided into a training set and a validation set, and the detection accuracy of the model on normal and abnormal state samples is recorded.

4. The method for monitoring the state of an injection mold cavity according to claim 1, wherein: The S3 includes: Through the dynamic fuzzy logic reasoning mechanism, the current state data is fuzzified to form a fuzzy set; Generate fuzzy rules, the fuzzy reasoning form is: IF(X1 IS A1)∧(X2 IS A2)THENYIS B; Among them, X1, X2 are input variables; A1, A2 are fuzzy sets; Y is the output variable; B is the output fuzzy set; IF and H are membership functions; E is the membership threshold; N is the number of rules; The mold condition evaluation results are generated based on fuzzy logic, and the output formula used is: Among them, w i is the weight of each fuzzy rule; B i is the output of the fuzzy rule; Y output is the final output variable; n is the number of fuzzy rules.

5. The method for monitoring the state of an injection mold cavity according to claim 4, characterized in that: The step S3 further includes a comparison process between the mold state and the abnormal pattern, including the following steps: Compare the generated mold status assessment results with the preset abnormal pattern library to identify whether there is an abnormal state; Generate alarm information based on the comparison results and record abnormal status characteristics to provide data support for subsequent analysis.

6. The method for monitoring the state of an injection mold cavity according to claim 1, characterized in that: The S4 includes: Upload real-time monitoring data and status assessment results to the cloud server via secure network protocols; Data aggregation and storage are performed in the cloud, supporting multi-source data integration and in-depth analysis.

7. The method for monitoring the state of an injection mold cavity according to claim 1, characterized in that: The S5 includes: Design a user-friendly feedback interface to allow operators to provide real-time feedback on any alarms and monitoring results given by the system; The collected feedback information is used to update the fuzzy rules and optimize the incremental learning model. The feedback information update formula is: Among them, F j is user feedback data; N1 represents the total amount of user feedback data collected; Update feedback Indicates the updated value calculated based on user feedback.

8. An injection mold cavity status monitoring system, characterized in that: A method for monitoring the state of an injection mold cavity according to any one of claims 1 to 7, comprising: Sensor module, used to collect status data of injection mold in real time; A data processing module, configured to pre-process and fuse the status data, and perform mold status assessment and anomaly detection using a machine learning model; Dynamic fuzzy logic reasoning module, used to analyze current status data and generate mold status evaluation results; Cloud storage module, used to upload real-time monitoring data and assessment results to the cloud for data aggregation, storage and in-depth analysis; The user feedback module is used to collect operator feedback information and optimize the model and inference rules accordingly.

9. An injection mold cavity status monitoring device, characterized in that: A method for monitoring the state of an injection mold cavity according to any one of claims 1 to 7, comprising: The sensor assembly, including temperature sensor, pressure sensor, flow sensor and humidity sensor, monitors the status of the injection mold in real time; A data collector, used to receive the status data provided by the sensor and perform preliminary processing and integration of the data; The processing unit analyzes the collected data based on the incremental learning algorithm and fuzzy logic reasoning mechanism to realize the mold status evaluation and anomaly detection functions; Cloud computing platform, used for data storage and analysis, supporting the aggregation and in-depth mining of multi-source data; User interface, used to display monitoring results and alarm information, allowing operators to provide real-time feedback.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images