Punching equipment remote fault diagnosis method and system based on Internet of Things

Abstract

The invention discloses a stamping equipment remote fault diagnosis method and system based on the Internet of Things, and relates to the technical field of stamping equipment fault diagnosis, and the method comprises the steps: S1, deploying multiple types of sensing units based on an Internet of Things architecture, synchronously collecting the equipment operation data, production process data and external environment data of stamping equipment, and forming three-dimensional coupling data; according to the invention, by constructing an equipment-process-environment three-dimensional data acquisition framework, synchronously acquiring multi-dimensional coupling data, and combining with an edge calculation real-time association algorithm, accurately identifying and rejecting false fault data, the problem of high false alarm rate caused by single data in the existing diagnosis technology is effectively solved, and the diagnosis accuracy is improved. By embedding a working condition and equipment individual difference compensation factor in an adaptive feature migration algorithm, cross-working-condition and cross-equipment feature dynamic fusion is realized, the generalization ability of a diagnosis model is improved, and relying on an edge-cloud collaborative architecture, the real-time performance of data processing is ensured, and the efficient fusion of global features is also realized.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology for stamping equipment, specifically to a remote fault diagnosis method and system for stamping equipment based on the Internet of Things. Background Technology

[0002] Stamping equipment is a key industrial device that uses pressure to plastically deform or separate metallic or non-metallic materials, thereby processing them into parts of a predetermined shape. It is widely used in various industries such as automobile manufacturing, electronic component production, and machining. Its operational stability directly affects production efficiency, product quality, and production cost control, making it one of the core infrastructures ensuring continuous industrial production. Remote fault diagnosis technology for stamping equipment relies on technologies such as the Internet of Things (IoT), data transmission, and intelligent algorithms to achieve real-time monitoring, data acquisition, and fault analysis of the operating status of remote stamping equipment. Through this technology, staff can monitor equipment operation without on-site supervision, promptly identify potential faults, and develop maintenance plans. This effectively reduces unplanned downtime, lowers manual inspection costs, and prevents serious losses caused by escalating faults, playing a significant role in improving the intelligence and efficiency of industrial production.

[0003] However, existing remote fault diagnosis technologies for stamping equipment still have certain shortcomings. The data collection dimension is singular, focusing only on the equipment's own operating data and failing to link production process parameters with external environmental data. This results in a lack of comprehensive data support for fault analysis, making it difficult to effectively distinguish between real equipment faults and false faults caused by process fluctuations and environmental changes. This can easily lead to false alarms. Furthermore, the existing methods have poor adaptability across operating conditions and equipment. They do not adapt to differences in operating conditions such as stamping speed changes and material replacements, as well as individual differences between different equipment, resulting in insufficient model generalization ability and difficulty in meeting the diagnostic needs of complex production scenarios. Therefore, developing a remote fault diagnosis method and system for stamping equipment based on the Internet of Things is of great significance. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a remote fault diagnosis method and system for stamping equipment based on the Internet of Things. It can construct a three-dimensional data acquisition framework of equipment-process-environment, synchronously collect multi-dimensional coupled data, and accurately identify and eliminate false fault data by combining edge computing real-time correlation algorithms. This solves the problem of high false alarm rate caused by single data in existing diagnostic technologies. By embedding working conditions and individual equipment difference compensation factors in the adaptive feature transfer algorithm, it realizes dynamic feature fusion across working conditions and equipment, improves the generalization ability of the diagnostic model, and relies on the edge-cloud collaborative architecture to ensure the real-time performance of data processing and achieve efficient fusion of global features.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a remote fault diagnosis method for stamping equipment based on the Internet of Things, the method comprising the following steps: S1. Deploy multiple types of sensing units based on the Internet of Things architecture to synchronously collect equipment operation data, production process data and external environment data of stamping equipment to form three-dimensional coupled data; S2. Transmit the three-dimensional coupled data to the edge computing node, perform data preprocessing first, and then establish the association mapping of the three types of data through a real-time association algorithm to identify and eliminate false fault data caused by factors other than the device itself. S3. Extract fault features from the filtered valid data and use a feature fusion algorithm to complete multi-source feature fusion. S4. Transmit the fused features to the cloud diagnostic platform, and use the preset diagnostic model to identify, assess and locate the fault; S5. Transmit the diagnostic results to the remote monitoring terminal, receive the maintenance records entered by the staff and feed them back to the cloud diagnostic platform, forming a closed loop of diagnosis-maintenance-feedback.

[0006] Furthermore, step S1, which involves acquiring three-dimensional coupled data, includes the following steps: Sensing units are deployed in the slider, motor, and die parts of the stamping equipment to collect vibration signals, temperature signals, sound signals, current signals, and visual image data as equipment operation data; Establish communication with the stamping production line control system to obtain the material parameters of the stamped parts, stamping frequency and processing accuracy requirements, as production process data; Environmental monitoring and sensing units are evenly deployed within the workshop to collect data on workshop temperature and humidity, power grid voltage fluctuations, and environmental dust concentration as external environmental data.

[0007] Furthermore, step S2, which involves identifying and removing false fault data using a real-time correlation algorithm, includes the following steps: Based on the failure mechanism of stamping equipment, we learn the association rules between failures and process parameters and environmental parameters, and construct a dynamic association mapping library. Monitor whether there are any abnormalities in the equipment operation data. When an abnormality is detected, retrieve the corresponding process data and environmental data from the associated mapping library. The abnormal data is matched and verified with the synchronously collected process data and environmental data. If the abnormality is caused by the difference in the hardness of the stamping part material, the sudden change in the power grid voltage, or the drastic fluctuation in the temperature and humidity of the workshop, it is marked as false fault data and removed.

[0008] Furthermore, step S3, which involves fault feature extraction and multi-source feature fusion, includes the following steps: Extract time-domain and frequency-domain features from equipment operation data, and extract key influencing features from production process data and external environment data to form a multi-dimensional initial feature set; An adaptive feature transfer algorithm is adopted, which embeds operating condition feature factors and equipment individual difference compensation factors to dynamically align multi-source heterogeneous features in the initial feature set. Weights are assigned according to the correlation strength between features and faults to complete feature fusion.

[0009] Furthermore, step S2, when performing data preprocessing, includes: using wavelet transform algorithm to denoise the time-series data of vibration and temperature signals in the three-dimensional coupled data; using timestamp alignment technology to adjust the multi-source data collected by different sensing units to the same time dimension to achieve time-series alignment; and performing format conversion on various types of data after denoising and alignment according to preset data format standards to complete data standardization processing.

[0010] Furthermore, in step S1, the communication with the stamping production line control system adopts the Modbus protocol, visual image data is collected by industrial cameras deployed in the equipment processing area, and vibration signals and temperature signals are collected and transmitted in real time by corresponding sensors.

[0011] Furthermore, the preset diagnostic model in step S4 is constructed in the following way: based on historical fault data, corresponding three-dimensional coupled data and fusion features, the model is trained using the support vector machine algorithm. During the training process, label data of fault type, fault location and fault severity are introduced to form a diagnostic model with classification and recognition capabilities.

[0012] Furthermore, in step S5, the diagnostic results are transmitted to the remote monitoring terminal via 5G or WiFi communication. Staff can view the diagnostic results through the terminal's display interface, and enter maintenance operation records and fault handling effect information through the terminal. The entered information is fed back to the cloud diagnostic platform through the data transmission channel.

[0013] The IoT-based remote fault diagnosis system for stamping equipment is applicable to the aforementioned IoT-based remote fault diagnosis method for stamping equipment. The system includes: a data acquisition module, an edge computing module, a data transmission module, a cloud diagnosis module, and a remote interaction module. The data acquisition module is used to simultaneously collect equipment operation data, production process data and external environment data to form three-dimensional coupled data. The edge computing module is communicatively connected to the data acquisition module and is used for data preprocessing and removal of false fault data. The data transmission module is communicatively connected to the edge computing module and the cloud diagnostic module to realize data interaction. The cloud-based diagnostic module is used for feature fusion and fault identification, assessment and location, while the remote interaction module is used for displaying diagnostic results and providing maintenance information feedback.

[0014] Furthermore, the data acquisition module includes an equipment operation sensing unit, a process parameter acquisition unit, and an environmental monitoring unit. The equipment operation sensing unit consists of a vibration sensor, a temperature sensor, a sound sensor, a current sensor, and an industrial camera. The process parameter acquisition unit is communicatively connected to the stamping production line control system. The environmental monitoring unit consists of a temperature and humidity sensor, a voltage monitor, and a dust sensor.

[0015] Compared with existing technologies, this IoT-based remote fault diagnosis method and system for stamping equipment has the following advantages: This invention constructs a three-dimensional data acquisition framework encompassing equipment, process, and environment, synchronously acquiring multi-dimensional coupled data. Combined with a real-time edge computing correlation algorithm, it accurately identifies and eliminates false fault data, effectively solving the problem of high false alarm rates in existing diagnostic technologies due to limited data. By embedding compensation factors for working conditions and individual equipment differences into an adaptive feature transfer algorithm, it achieves dynamic feature fusion across working conditions and equipment, enhancing the generalization ability of the diagnostic model and meeting the needs of complex production scenarios. Relying on an edge-cloud collaborative architecture, it ensures both real-time data processing and efficient fusion of global features, enabling timely and accurate fault location, reducing unplanned downtime, lowering maintenance costs, and providing reliable assurance for the stable operation of stamping equipment.

[0016] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0018] Figure 1 A flowchart of a remote fault diagnosis method for stamping equipment based on the Internet of Things; Figure 2 This is a flowchart of a remote fault diagnosis method for stamping equipment based on the Internet of Things. Figure 3 This is a schematic diagram of a remote fault diagnosis system for stamping equipment based on the Internet of Things. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0020] The present invention provides a remote fault diagnosis method and system for stamping equipment based on the Internet of Things, which clarifies the complete technical solution, and the core contents are as follows: The methodology section includes five key steps; see [link / reference]. Figure 1 and Figure 2 The specific details are as follows: Three-dimensional coupled data acquisition is carried out by deploying multiple types of sensing units based on the Internet of Things architecture. Vibration, temperature and other equipment operation data are collected in key parts such as the slider and motor of the stamping equipment. Process parameters are obtained by communicating with the production line control system through the Modbus protocol. Monitoring units are evenly deployed in the workshop to collect external environmental data such as temperature and humidity.

[0021] Data preprocessing and pseudo-fault removal: The 3D coupled data is transmitted to the edge computing node. After wavelet transform denoising, time alignment and standardization, a correlation mapping library is built through real-time correlation algorithm to match and verify abnormal data with process and environmental data, and to remove pseudo-fault data caused by non-equipment factors such as material differences and voltage changes.

[0022] Feature extraction and fusion: Multi-dimensional initial features are extracted from effective data, and an adaptive feature transfer algorithm is used to embed compensation factors for working conditions and individual equipment differences. Weights are assigned according to the correlation strength between features and faults to complete the fusion.

[0023] Cloud-based fault diagnosis uses historical data and support vector machine algorithms to build a diagnostic model, and inputs fused features to achieve fault identification, assessment and location.

[0024] The diagnostic feedback loop transmits diagnostic results to a remote monitoring terminal via 5G or WiFi, receives maintenance records entered by staff, and feeds them back to the cloud.

[0025] The system part corresponds to the above method and includes five main modules, see [link / reference]. Figure 3 The specific details are as follows: The data acquisition module consists of an equipment operation sensing unit, a process parameter acquisition unit, and an environmental monitoring unit, which collect three types of data to form three-dimensional coupled data. The edge computing module communicates with the data acquisition module and is responsible for data preprocessing and pseudo-fault data removal. The data transmission module connects the edge computing module and the cloud diagnostic module to achieve data interaction. The cloud diagnostic module undertakes feature fusion and fault diagnosis-related tasks. The remote interaction module is used to display diagnostic results and receive maintenance information feedback, ensuring the realization of the diagnosis-maintenance-feedback closed loop.

[0026] This technical solution addresses the problems of traditional diagnostic technologies, such as limited data, high false alarm rate, and insufficient generalization ability, through 3D data acquisition, false fault elimination, cross-scene feature fusion, and edge-cloud collaborative architecture. It enables accurate and efficient remote diagnosis of stamping equipment faults.

[0027] Example 1 This embodiment applies to a remote fault diagnosis scenario for stamping equipment in an automotive parts manufacturing workshop. This workshop has multiple stamping machines of different models used to process key components such as engine blocks and chassis brackets. During production, frequent switching of stamping materials and processing parameters is required. Workshop temperature, humidity, and power grid voltage are easily affected by fluctuations in the external environment. Traditional fault diagnosis methods often result in false alarms and missed alarms due to limited data and poor adaptability to operating conditions, leading to unplanned equipment downtime and impacting production progress. This embodiment utilizes an IoT-based remote fault diagnosis method and system for stamping equipment. (See [link to relevant documentation]). Figure 1 , Figure 2 and Figure 3 This enables multi-dimensional collaborative data collection, accurate fault identification, and efficient remote control, ensuring continuous and stable production.

[0028] During the data acquisition phase, multiple types of sensing units are deployed based on an IoT architecture. Vibration sensors, temperature sensors, sound sensors, and current sensors are installed on key components such as the slider, motor, and mold of each stamping machine. Industrial cameras are deployed in the processing area of ​​the equipment to collect vibration signals, temperature signals, sound signals, current signals, and visual image data in real time as equipment operation data. Communication with the stamping production line control system is established through the Modbus protocol to obtain production process data such as material parameters of stamped parts, stamping frequency, and processing accuracy requirements in real time. Temperature and humidity sensors, voltage monitors, and dust sensors are installed in the workshop according to the principle of uniform distribution to continuously collect external environmental data such as workshop temperature and humidity, power grid voltage fluctuations, and ambient dust concentration. The three types of data are collected simultaneously to form three-dimensional coupled data.

[0029] The acquired 3D coupled data is transmitted in real time to the edge computing node via the data transmission module, where it undergoes preprocessing. Wavelet transform algorithms are used to denoise the time-series data, such as vibration and temperature signals, filtering out noise caused by environmental interference. Timestamp alignment technology is used to adjust the multi-source data collected by different sensing units to the same time dimension, achieving time-series alignment. Finally, the denoised and aligned data are converted to the desired format according to a preset data format standard, completing the data standardization process.

[0030] Subsequently, false fault data is identified and eliminated through real-time correlation algorithms. Based on the fault mechanism of stamping equipment, the correlation rules between faults and process parameters and environmental parameters are learned, and a dynamic correlation mapping library is constructed. Edge computing nodes monitor equipment operation data in real time. When an abnormal vibration signal of a certain piece of equipment is detected, the process data and environmental data collected at the same time are retrieved from the correlation mapping library.

[0031] In the specific implementation of this embodiment, matching verification is performed using an association matching degree formula: ,in For correlation matching degree, The total number of process and environmental parameters involved in the matching. For the first The correlation weight of each process or environmental parameter When the abnormality occurs Real-time collected values ​​of each parameter. For the first The normal threshold center value of each parameter, , The first The historical maximum and minimum values ​​of each parameter. For the first The influence coefficient of each parameter on equipment malfunction. The frequency of false faults caused by each parameter in historical fault data is determined. The impact of each parameter on the equipment's operating status is determined based on the analysis of the stamping equipment's failure mechanism. If the verification results show that the anomaly is caused by differences in the hardness of the stamped parts, sudden changes in power grid voltage, or drastic fluctuations in workshop temperature and humidity, it is marked as false fault data and discarded, retaining only the abnormal data caused by factors inherent to the equipment itself as valid data.

[0032] The selected valid data undergoes fault feature extraction and fusion. Time-domain and frequency-domain features are extracted from equipment operation data, and key influencing features are extracted from production process data and external environment data, forming a multi-dimensional initial feature set. An adaptive feature transfer algorithm is employed, embedding operating condition feature factors and equipment individual difference compensation factors to dynamically align the multi-source heterogeneous features in the initial feature set, and assigning weights based on the correlation strength between features and faults.

[0033] In the specific implementation of this embodiment, the fused feature values ​​are calculated using the feature fusion formula: ,in For the final fused feature values, This represents the total number of features in the initial feature set. For the first The fusion weights of the initial features, For the first The normalized values ​​of the initial features, These are the coefficients of the operating condition characteristic factor. For the first The working condition adaptation value corresponding to each feature For individual equipment differences, compensation coefficient For the first Device difference adaptation value corresponding to each feature. Determined based on the recognition contribution rate of each feature in historical fault diagnosis. , Based on training and optimization using historical feature migration data across operating conditions and equipment, the determination method is determined. , The parameters are calculated based on current operating conditions, equipment factory parameters, and service life. This achieves efficient fusion of multi-source features, enhancing the feature's ability to characterize faults.

[0034] The fused features are transmitted to the cloud-based diagnostic platform via the data transmission module. The platform's preset diagnostic model is built using a support vector machine algorithm, based on historical fault data, corresponding 3D coupled data, and fused features. During training, label data for fault type, fault location, and fault severity are introduced. In this specific implementation, the model parameters are optimized using a classification loss formula. ,in The classification loss value for the model, This represents the total number of training samples. For the first The true label value of each training sample. For the model to the first The predicted class probability value for each training sample. This formula has no additional weights or coefficients; the loss is calculated directly from the deviation between the true label and the predicted probability of the training sample. During model parameter optimization, the goal is to minimize... To achieve the goal, the internal weight matrix of the model is iteratively adjusted.

[0035] After receiving and fusing features, the model quickly identifies, assesses, and locates faults, clarifying the fault type, specific location, and severity. The diagnostic results are transmitted to the workshop's remote monitoring terminal via 5G communication. Staff can view the results in real-time through the terminal's display interface. For identified mold wear faults, maintenance personnel are dispatched with specialized tools to the site for mold replacement and adjustment. After repair, maintenance operation records and fault handling effectiveness information are entered into the terminal. This information is then fed back to the cloud-based diagnostic platform via data transmission, forming a closed loop of diagnosis, maintenance, and feedback, providing data support for subsequent model optimization.

[0036] In summary, this embodiment constructs a three-dimensional data acquisition framework encompassing equipment, process, and environment to comprehensively acquire multi-dimensional data on the operation of stamping equipment and its surrounding environment. Combined with real-time data processing and false fault elimination via edge computing nodes, it effectively avoids false alarms caused by single data points. An adaptive feature transfer algorithm enables dynamic feature fusion across operating conditions and equipment, enhancing the diagnostic model's generalization ability in complex production scenarios. Relying on an edge-cloud collaborative architecture, it ensures both real-time data processing and efficient global feature fusion. The entire diagnostic process requires no on-site personnel, enabling accurate fault location and rapid response, significantly reducing unplanned equipment downtime and lowering manual inspection and maintenance costs. This provides reliable technical support for the efficient and stable operation of automotive parts production workshops. Furthermore, the resulting diagnostic-maintenance-feedback closed loop continuously optimizes the diagnostic model's performance, adapting to long-term changes in operating conditions and equipment aging during production.

[0037] Example 2 This embodiment applies to a precision stamping production base for electronic components. This base has multiple stamping production lines of different specifications, focusing on the mass production of high-precision components such as micro-connectors and chip pins. During production, the stamping equipment needs to operate continuously under high-frequency, high-precision conditions, and the processed materials are mostly thin metal sheets, placing extremely high demands on equipment operational stability and process adaptability. Simultaneously, the base's workshops are widely distributed, with natural differences in temperature, humidity, and dust concentration in different areas. The power grid voltage is prone to slight fluctuations due to the simultaneous operation of multiple production lines. Traditional diagnostic methods struggle to address the complex fault causes in multi-equipment collaborative production scenarios, often resulting in delayed fault location and insufficient cross-equipment diagnostic adaptability. This embodiment builds upon the aforementioned embodiments, referring to... Figure 1 , Figure 2 and Figure 3 It enhances the collaborative diagnostic capabilities of multiple production lines and the adaptability to precision working conditions. By optimizing the data acquisition layout and model application logic, it enables accurate and rapid diagnosis of faults in precision stamping equipment.

[0038] The data acquisition phase builds upon the aforementioned three-dimensional coupled data acquisition framework, with optimized deployments tailored to precision stamping scenarios. Miniature vibration and temperature sensors are added to key precision components of each stamping machine, such as the mold cavity and feeding mechanism. These sensors, combined with existing sensors, collect high-frequency vibration signals and instantaneous temperature changes, providing data on equipment operation. Distributed communication is established with the control systems of multiple production lines via the Modbus protocol, synchronously acquiring production process data such as the material thickness of stamped parts, stamping pressure parameters, and processing cycle time from each production line, enabling parallel data acquisition from multiple production lines. Environmental monitoring units are deployed in different areas of the workshop according to their distance from the equipment. Nearby data collects local temperature, humidity, and dust concentration data around the equipment, while distant data collects overall fluctuations in the regional power grid from power grid voltage monitors. This creates external environmental data that considers both local and global factors. The simultaneous acquisition of these three types of data forms more targeted three-dimensional coupled data.

[0039] The collected 3D coupled data from multiple production lines is categorized and transmitted to an edge computing node cluster via a data transmission module. Each node corresponds to a data processing task for one production line. The data preprocessing stage retains the wavelet transform denoising, timestamp alignment, and standardization processes described in the previous embodiment. For the high-frequency characteristics of precision stamping data, the filtering threshold of the denoising algorithm is optimized to improve the ability to retain transient abnormal signals. Subsequently, a pseudo-fault data removal process is initiated. Based on the fault mechanism of precision stamping equipment, association rules related to precision machining, such as material thickness deviation and stamping pressure fluctuation, are added to the dynamic association mapping library.

[0040] When an abnormality is detected in the current signal of a piece of equipment on a certain production line, the process data of that production line during the same period and the environmental data of the corresponding area are retrieved. In the specific implementation of this embodiment, the verification is performed using the correlation matching degree formula: If the anomaly is caused by differences in the uniformity of the thin metal sheet material, excessive dust concentration in the local environment, or slight fluctuations in the regional power grid voltage, it will be marked as false fault data and removed to ensure that only valid data related to the precision components of the equipment itself is retained.

[0041] The fault feature extraction and fusion process adds a feature screening step to the previous embodiments. Features such as time-domain peak values ​​and frequency-domain harmonics of equipment operating data, parameter deviation features of production process data, and local fluctuation features of external environmental data are extracted from valid data to form a multi-dimensional initial feature set. Redundant features are then eliminated through feature correlation analysis.

[0042] An adaptive feature transfer algorithm is employed, embedding condition feature factors optimized for precision operating conditions and equipment individual difference compensation factors. The selected features are dynamically aligned, and weights are assigned based on the correlation strength between the features and precision faults. In the specific implementation of this embodiment, the fused feature value is calculated using a feature fusion formula. This enables the accurate fusion of multi-source features, enhancing the ability to characterize micro-fault features.

[0043] The fused features are aggregated to the cloud-based diagnostic platform via a data transmission module. This platform's diagnostic model, based on the support vector machine algorithm described in the previous embodiment, incorporates a multi-task learning mechanism to construct customized sub-diagnostic models for different production line equipment models and processing techniques. In the specific implementation of this embodiment, during model training, the parameters of each sub-model are optimized using the classification loss formula: After receiving the fused features, the platform automatically matches the corresponding sub-diagnostic model according to the production line to which the data belongs, and quickly completes fault identification, assessment and location, clarifying the fault type of precision components, such as mold micro-cracks, feeding mechanism jamming and other fault locations and impact ranges.

[0044] The diagnostic results are transmitted via WiFi to the central monitoring terminal at the base and the on-site operation terminals on each production line. Staff, combining the fault details displayed on the terminals, dispatch specialized maintenance personnel with precision testing tools to conduct targeted repairs. After repairs are completed, the fault handling process, component replacement information, and production recovery status are recorded on the terminals. This information is fed back to the cloud-based diagnostic platform, providing data support for model optimization and creating a fault case library across multiple production lines, serving as a reference for handling similar faults in the future.

[0045] In summary, based on the aforementioned embodiments, this embodiment optimizes the data acquisition layout and processing logic to meet the high-frequency, high-precision requirements of precision stamping of electronic components. By deploying environmental monitoring units in a layered manner, supplementing relevant rules for precision machining, adding feature filtering steps, and constructing a multi-task sub-diagnostic model, the targeting and accuracy of fault diagnosis are improved. The multi-production line parallel data processing mode adapts to the large-scale operation needs of the production base, effectively solving the problems of difficult identification of micro-faults in precision stamping equipment and poor cross-production line diagnostic adaptability. The diagnostic process achieves rapid fault location and accurate processing, significantly reducing the scrap rate of precision parts and minimizing production downtime caused by equipment failures. Simultaneously, the resulting multi-production line fault case library provides valuable data support for the long-term operation and maintenance of the base, further improving the intelligence and efficiency of electronic component stamping production and providing a feasible technical solution for equipment fault diagnosis in the precision manufacturing industry.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A remote fault diagnosis method for a stamping apparatus based on an Internet of Things, characterized by, The method comprises the following steps: S1, based on the Internet of Things architecture, deploying multiple types of sensing units, synchronously collecting device operation data, production process data and external environment data of the stamping equipment, forming three-dimensional coupled data; S2, the three-dimensional coupled data is transmitted to the edge computing node, first data preprocessing is carried out, then the real-time correlation algorithm is used to establish the correlation mapping of the three types of data, and the pseudo-fault data caused by non-device itself factors is identified and removed; S3, extracting fault features from the screened effective data, and completing multi-source feature fusion by using feature fusion algorithm; S4, transmitting the fused features to the cloud diagnosis platform, and identifying, evaluating and positioning the fault through the preset diagnosis model; S5, transmitting the diagnosis result to the remote monitoring terminal, receiving the maintenance record input by the staff and feeding back to the cloud diagnosis platform, forming a diagnosis-maintenance-feedback closed loop.

2. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized in that, When the step S1 collects three-dimensional coupled data, the following steps are included: Deploy sensing units at the slide block, motor and die parts of the stamping equipment, collect vibration signals, temperature signals, sound signals, current signals and visual image data as device operation data; Establish communication with the stamping production line control system to obtain stamping part material parameters, stamping frequency and machining precision requirements as production process data; Uniformly deploy environmental monitoring sensing units in the workshop to collect workshop temperature and humidity, power grid voltage fluctuation and environmental dust concentration as external environment data.

3. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized in that, When the step S2 identifies and removes pseudo-fault data by using real-time correlation algorithm, the following steps are included: Based on the fault mechanism of stamping equipment, learn the correlation rules of fault, process parameters and environmental parameters, and build a dynamic correlation mapping library; Monitor whether the device operation data is abnormal, and when an abnormality is detected, retrieve the corresponding process data and environmental data from the correlation mapping library; Match and verify the abnormal data with the synchronously collected process data and environmental data, if the abnormality is caused by the hardness difference of stamping parts, power grid voltage mutation or sharp fluctuation of workshop temperature and humidity, mark it as pseudo-fault data and remove it.

4. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized in that, When the step S3 extracts fault features and fuses multi-source features, the following steps are included: Extract time domain features and frequency domain features from device operation data, extract key influence features from production process data and external environment data, form a multi-dimensional initial feature set; Use adaptive feature migration algorithm to embed working condition feature factor and device individual difference compensation factor, dynamically align the multi-source heterogeneous features in the initial feature set, assign weights according to the correlation strength of features and faults, and complete feature fusion.

5. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized in that, When the step S2 preprocesses data, the following steps are included: use wavelet transform algorithm to denoise the time series data of vibration signals and temperature signals in three-dimensional coupled data, use timestamp alignment technology to adjust the multi-source data collected by different sensing units to the same time dimension, realize time alignment, convert the format of each type of data after denoising and alignment according to the preset data format standard, and complete data standardization processing.

6. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized by, The communication in the step S1 with the stamping production line control system adopts a Modbus protocol, visual image data is collected by an industrial camera deployed in the equipment processing area, and vibration signals and temperature signals are collected and transmitted in real time by corresponding sensors.

7. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized by, The preset diagnosis model in the step S4 is constructed by the following method: based on historical fault data, corresponding three-dimensional coupling data and fusion features, a support vector machine algorithm is used for model training, label data of fault type, fault position and fault severity is introduced during the training process, and a diagnosis model with classification and identification capabilities is formed.

8. The IoT-based remote fault diagnosis method of a press equipment according to claim 1, characterized by, The diagnosis result in the step S5 is transmitted to a remote monitoring terminal through a 5G or WiFi communication mode, the staff checks the diagnosis result through the display interface of the terminal, enters maintenance operation records and fault processing effect information through the terminal, and the entered information is fed back to the cloud diagnosis platform through a data transmission channel.

9. The stamping equipment remote fault diagnosis system based on Internet of Things, which is suitable for the stamping equipment remote fault diagnosis method based on Internet of Things in any one of claims 1-8, characterized in that, The system comprises a data acquisition module, an edge computing module, a data transmission module, a cloud diagnosis module and a remote interaction module. The data acquisition module is used for synchronously collecting equipment operation data, production process data and external environment data to form three-dimensional coupling data. The edge computing module is in communication connection with the data acquisition module and is used for data preprocessing and pseudo-fault data elimination. The data transmission module is in communication connection with the edge computing module and the cloud diagnosis module and realizes data interaction. The cloud diagnosis module is used for feature fusion, fault identification, evaluation and positioning, and the remote interaction module is used for diagnosis result display and maintenance information feedback.

10. The IoT-based remote fault diagnosis system of a press equipment according to claim 9, characterized by, The data acquisition module comprises an equipment operation sensing unit, a process parameter acquisition unit and an environment monitoring unit, the equipment operation sensing unit is composed of a vibration sensor, a temperature sensor, a sound sensor, a current sensor and an industrial camera, the process parameter acquisition unit is in communication connection with a stamping production line control system, and the environment monitoring unit is composed of a temperature and humidity sensor, a voltage monitor and a dust sensor.

Images